**Drivers and Inhibitors in the Acceptance of Meat Alternatives: The Case of Plant and Insect-Based Proteins**

**Wim de Koning 1,2,3 , David Dean 1 , Frank Vriesekoop 2,3, \* , Luis Kluwe Aguiar 2 , Martin Anderson 2 , Philippe Mongondry 4 , Mark Oppong-Gyamfi 4 , Beatriz Urbano 5 , Cristino Alberto Gómez Luciano 6 , Bin Jiang 7 , Wendy Hao 2 , Emma Eastwick 2 , Zheng (Virgil) Jiang <sup>2</sup> and Anouk Boereboom 2,3**


Received: 5 August 2020; Accepted: 10 September 2020; Published: 14 September 2020

**Abstract:** Insects as an alternative protein source has gained traction for its advantageous environmental impact. Despite being part of many traditional food cultures, insects remain a novelty in Western cultures and a challenging concept for many. Even though plant-based protein alternatives are not facing the same barriers, product unfamiliarity and limited exposure hinder adoption, which could be detrimental to growth within the food sector. This study is aimed at evaluating plant- and insect-based proteins as alternative dietary proteins. A model indicating the drivers of consumer attitudes towards meat-alternative proteins and consumer willingness to try, buy, and pay a premium was tested. Further, 3091 responses were collected using surveys in nine countries: China, USA, France, UK, New Zealand, Netherlands, Brazil, Spain, and the Dominican Republic. Structural Equation Modelling was used to analyze the data. We found that consumer's behavioral intentions towards both plant-based and insect-based alternatives are inhibited by food neophobia but to an extent, are amplified by the perceived suitability and benefits of the protein, which in turn are driven by nutritional importance, environmental impact, healthiness, and sensory attributes for both alternatives. The expectation of the nutritional value of meat is the strongest (negative) influence on perceived suitability/benefits of plant-based protein and willingness to try, buy, and pay more for plant-based proteins, but it only has a relatively small impact on the suitability/benefits of insect-based protein and no impact on willingness to try, buy, and pay more for insect-based proteins. Overall, we conclude that consumer adoption towards meat alternatives is complex and is strengthened by the perceived suitability/benefits of the protein and general importance of perceived food healthiness and sustainability. Conversely, adoption is hindered by dietary factors and the experiential importance of meat and food neophobia.

**Keywords:** willingness to try; neophobia; structural equation model

#### **1. Introduction**

Much has been publicized about how the unsustainable ways of traditional meat production and consumption [1–3] is detrimental to both the environment and human health [4–6]. As a result, meat, particularly red meat, has attracted much criticism in recent years [7,8]. This, compounded with demographic pressures and socio-economic growth trends, has encouraged new product development and the introduction of a variety of alternatives to traditional animal proteins, thus extending the availability of meat substitutes in many markets. In order for the necessary changes to become a reality in our current food systems, we need to have a better understanding of how consumers view meat alternatives and gauge their willingness to change their purchasing and consumption habits.

Meat alternatives are not new, particularly plant-based and mycoproteins such as Quorn. Nevertheless, Quorn's global market share has not spread much outside the UK and the product has remained niche in most countries outside the UK. Conversely, pulses, which are a great source of plant proteins, have been a traditional part of staple diets in many cultures for millennia [9]. More recently, plant-based protein sources have been part of extensive new product development by the food industry bringing meat alternatives to the market that are promoted under higher sustainability credentials [10,11], thus catering to the burgeoning vegetarian and vegan segments.

The potential of insects as an alternative protein source has also gained traction because of advantages in resource usage, such as land, feed, water, and energy, and the role they can play in circular production systems [12,13]. Insects are and have been part of food cultures of large swathes of the world population. Yet, it is felt that in the more economically developed western countries, insect eating, entomophagy, and the consumption of products made with insect protein are still a novelty and a challenging concept for many consumers. Consumers' unfamiliarity and limited exposure to different food products hinders the adoption of new foods, which holds true for most foods made with alternative proteins [14,15].

#### *1.1. Theoretical Underpinning*

Western consumers tend to possess an ingrained barrier to eating insects and insect-based products, which is expressed through fear and disgust [16]. Such a behavior is typical of a food neophobic trait. Kush et al. [13] posited that consumers tended not to change their purchasing behaviors easily. The consumers' reluctance to change could be attributed to an inbuilt evolutionary-derived encoded instinct to protect humans against potential poisonous foods over familiar ones that are more beneficial to health and growth [17,18]. Thus, a predisposition to avoid unusual foods is based on instinctual neophobia [19], which has been socially constructed and filtered through the consumers' system of values [20]. This could play a major role with regards to protein consumption, where an aversion to alternative proteins could constitute a major impediment for replacing meat for another substitute because of the consumer's values, dietary habits, and preferences [21]. This is not unlike when plant-based proteins were first introduced into people's diets more widely [22].

Some of this behavior can be described as food neophobia, which is considered an expression of an aversion trait in consumers' choice behavior with regards to new foods [23]. However, the more frequent and intense the exposure to a new food product through information, education, and experimentation, the lesser the rejection by consumers. Therefore, it can be argued that food neophobia boundaries can be shifted over time. Clark and Bogdan [24] demonstrated that considerable barriers continue to confront the expansion of the market for plant-based proteins. However, their research suggested that once consumers have adopted plant-based meat alternatives, they were more likely to try new plant-based protein versions within the same product category in the future.

Schouteten et al. [25] compared meat, plant, and insect protein in the format of burgers. The overall liking of the plant and insect burger was similar, however the majority of consumers expressed disappointment for both alternatives compared to the traditional meat burger. However, when informed of the ingredients prior to tasting, the approval of the insect burger was significantly higher compared to when the information was not disclosed [25]. Gómez-Luciano et al. [26] found a greater willingness to purchase plant-based protein compared to insect-based proteins, however the responses varied between countries analyzed. Despite a reluctance to immediately adopt new foods, consumers indicated to being open to future changes, supporting a growing dietary shift to alternative dietary proteins [27]. These findings are in agreement with van der Weele et al. [28] who concluded that organizational and institutional coordination were required to enable the acceptance of meat alternatives (insect, pulses, and cultured meat), with recommendations to drive nutritional, sustainability, technological, and societal changes.

It is well understood that one of the major constraints concerning consumers' willingness to engage with sustainable food innovations is the consumers themselves [25–27]. Pliner and Hobden [29] developed a Food Neophobia Scale (FNS), which has since served to measure the consumers' willingness to consume foods that they might not be familiar with or have held a life-long aversion to. Cox and Evans [30] investigated food-related neophobia one step further and considered the possible aversion to new foods produced by novel technologies, which has been coined as Food Technology Neophobia [30]. Both the Food Neophobia scale and the Food Technology Neophobia scale have been widely validated in many different contexts [23,31,32]. However, Bäckström et al. [32] mentioned that familiarity played an important role in people's willingness to try a product that they do not recognize or have not encountered before. Consequently, unfamiliar products would face barriers to consumption as they clash with habit-bound consumer behavior [32]. Capitanio et al. [33] concluded that the aversion to consume novel foods was driven by a fear of what a food product could contain regarding ingredients and the processes used in its production. Chang et al. [34] argued that for organic foods, when too much processing had taken place, a product's perceived authenticity would be diminished, resulting in a lower purchase intention, which agrees with Eyhorn et al. [35]. Furthermore, despite a greater willingness to try a novel food product, consumers' intentions to pay more for meat alternatives is often low [36,37]. Therefore, despite the growing literature around the topic, there is still the need to investigate the drivers that influence consumers' attitudes towards meat alternatives. This study's contribution is to bring to light what consumers' attitudes would be toward willingness to buy, willingness to try, and willingness to pay a premium for meat alternatives such as plant- and insect-based products.

#### *1.2. Model Development*

The overarching aim of this study was to evaluate whether plant- and insect-based proteins could be realistic meat alternatives from the consumers' point of view. In order to test a theoretical model, attitudes towards the two types of meat substitutes were analyzed and the extent to which there were differences in consumers' attitudes and preferences between the alternatives was tested. Meat functioned as the default to which consumers could compare a widely accepted meat alternative (plant-based) and a meat alternative that could be integrated into a circular production system (insect-based) [38,39]. It also aimed at establishing a model indicating the drivers of consumers' attitudes towards meat-alternative proteins and consumers' willingness to try, buy, and pay a premium for them.

The model (Figure 1) was designed based on the literature that supported the notion that new and unfamiliar foods affected consumer behavior [20,31]. It was expected that Food Neophobia and Food Technology Neophobia would inhibit consumers' willingness to try, to buy, and pay more for meat-alternative proteins. Nine hypotheses were tested (Figure 1). The consumers' attitudes towards the importance of meat taste, texture, smell, and the nutritional importance of meat were expected to be negatively influenced by their perception of meat-alternative suitability and benefits [19,21]. Suitability and benefits were defined as a combination of sensory benefits, nutritional importance, environmental impact, and health influence that was unique to the meat substitute in question. The importance consumers placed on healthiness and the environmental impact of their food choices, in general, was likely to enhance their assessment of meat substitutes [40]. Consumers' attitudes towards the suitability of and benefits derived from a particular meat-alternative protein should also augment their willingness to adopt it [26,27]. Therefore, the proposed model should establish a better understanding of how consumers viewed meat alternatives and their willingness to change their purchasing and consumption habits.

**Figure 1.** Conceptual Model and Hypotheses.

#### **2. Method**

A sample of 3091 responses in total was obtained from surveys carried out in nine countries. The sample was composed of 571 respondents from China (CN), 539 from the USA (US), 484 from France (FR), 366 from the UK, 268 from New Zealand (NZ), 231 from the Netherlands (NL), 216 from Brazil (BR), 210 from Spain (ES), and 206 from the Dominican Republic (DR). Data collection started in February 2017 and finished in April 2018 in CN, the US, FR, the UK, BR, ES, and the DR. From February 2018 until May 2019, data were collected in NZ and the NL. The gender distribution was 59.2% females, 38.9% males, and 1.9% who preferred not to answer. The mean age of the sample was 34, with quartile ranges of 16–21, 22–28, 29–44, and 45–86. Table 1 provides a country-by-country insight into the demographics of the survey respondents.


**Table 1.** Demographics (gender and age) of the survey respondents per country.

\* NA = not applicable.

The questionnaire was initially written in English and then translated into the various respective languages by native speakers who were fluent in both English and their mother tongue to improve the accuracy of meaning and avoid misunderstandings by the various linguistic cohorts. The languages were also adjusted for variations in grammar/spelling, i.e., UK-English, US-English, and NZ-English; ES-Spanish and DR-Spanish; as well as Brazilian Portuguese. The translated versions were back-translated into English to ensure that the meaning had not deviated from the initial word concept or idea. The various collaborators and co-authors were responsible for distributing the survey at a country level (mainly through social media and existing e-mail contact lists). All data gathered were centrally collected and collated at Harper Adams University (HAU) in the UK. In most instances, the questionnaire was distributed in a digital format, however when requested, a hardcopy version was also made available. In the DR, the responses were predominantly collected using a hardcopy, catering for the relatively scant access to the Internet in that country. The research and questionnaire were approved by the Harper Adams University (HAU/UK) Research Ethics Committee (HAU-0006-201701). Furthermore, as part of the ethics declaration, each questionnaire also included a contact e-mail at HAU, so that questions arising from answering the questionnaire could be addressed.

#### *Questionnaire and Scaling*

The questionnaire included various distinct sets of questions and statements consistent with a previous study [26]. The participants gave their informed consent to partake in the survey. The first group of statements probed the respondents' attitudes towards new foods, new food technologies, health, convenience, and the environmental impact of their food choices (Table 2). More specifically,

the following scales were used in the questionnaire to measure the various constructs: Food Neophobia Scale, with 10 items, adapted from Pliner and Hobden [29] (Table 2, 08.1 through to 08.10); Food Technology Neophobia Scale, with five items, which was inspired by Cox and Evans [30] (Table 2, 09.1 through to 09.8); Healthiness of Food Choices, with three items, adapted from the "impact of the healthiness of food choices" scale [40] (Table 2, 10.1 through to 10.3); and Environmental Impact of Food Choices, with three items, adapted from the "environmental impact of food choices" scales in Roberts [41] and Verbeke [37] (Table 2, 12.3 through to 12.3). Many of the above-mentioned scales were adapted from previously described tools [26,29,30,37,40,41] in relation to assessing people's willingness to engage with new foods. In these adaptations, we made careful choices with regards to which survey items to include in our study to avoid unnecessary duplication, utilize the most appropriate items, and avoid potential survey fatigue. For instance, the original food technology neophobia scale [30] contains items covering health and environmental factors, however we found that these topics were better addressed using the survey items used elsewhere [37,40,41]. As such, we also detached those sub-topics from the original scale and addressed them separately. The second group of statements probed the respondents' perceived importance of meat in terms of its nutritional benefits and sensory experience (Table 2). More specifically, a 3-item scale measured Meat Nutritional Importance (Table 2, 13.1 through to 13.3) and a 3-item scale measured Meat Taste, Texture, and Smell Importance (Table 2, 14.1 through to 14.3). All the questions were presented in the form of statements to which the respondents expressed their opinion using a five-point Likert scale ranging from "strongly disagree" to "strongly agree" (Table 2).


**2.**ScaleLoadings,Reliabilities,andConvergentValidity.


**Table2.***Cont.*

*Foods* **2020**, *9*, 1292

The questionnaire then included descriptions of plant-based and insect-based alternatives to meat proteins. Consumers were asked about their perceptions of the suitability of or the benefits derived from plant-based and insect-based proteins. These questions consisted of six items measuring healthiness, safety, nutrition, sustainability, taste, and affordability relative to meat protein (Table 2). Finally, a consumer behavioral intention scale was used to measure aspects such as willingness to try, willingness to buy, and willingness to pay more for plant-based and insect-based proteins. The questionnaire also collected some demographic characteristics of the respondents.

#### **3. Analysis**

A two-step Structural Equation Modelling was used. The first step was related to the evaluation of the measurement model using confirmatory factor analysis. This step evaluated the measurement scales and their items, examining construct convergent and discriminant validity and reliability. The second step tested the model, assessing the significance of the hypothesized relationships between the variables and confirming that goodness-of-fit criteria were satisfied. This two-step analysis was selected due to its appropriateness in the measurement and examination of structural models and testing coefficient paths. For an excellent discussion on the ongoing development and generally accepted process for employing the type of Structural Equation Modelling used in this research, see [42].

#### *3.1. Construct Validity and Reliability*

Construct validity was evaluated using factor loadings and average variance extracted (AVE). As shown in Table 2, the result of convergent validity assessment indicated that except for the Food Neophobia scale item, "Some foods look too weird to eat," all of the standardized loadings were above the cut-off level of 0.5, as set by Anderson and Gerbing [43]. Except for the Food Neophobia and Food Tech Neophobia scales, Table 2 also shows that the AVE of all the scales was higher than the 0.5 cut-off level as suggested by Hair et al. [44]. Unfortunately, the removal of any items to those scales resulted in the lowering of Cronbach's Alpha and Composite Reliability values, so it was decided not to take remedial action.

Table 2 also shows that the scales demonstrated adequate reliability. All but one (Environmental Impact Influence) of the scales had Cronbach's Alpha values above the cut-off level of 0.7 and all the scales had composite reliability values above the suggested cut-off level of 0.7 [44].

The discriminant validity of the construct scales was acceptable using both the Fornell-Larker criterion and the Heterotrait-Monotrait (HTMT) ratio methods. Table 3 shows that the Fornell-Larcker criterion was satisfied as the shared variance between the constructs was lower than the variance captured by the construct (along the diagonal). The HTMT ratio was also satisfied as the HTMT correlation estimates between the scales were below the recommended threshold of 0.85 [45], confirming adequate discriminant validity.



**Table 3.**Scale Discriminant Validity.



#### *3.2. Structural Model*

Following Hair et al. [44], a bootstrapping method with 500 repetitions was applied to assess the significance of the indicator weights and the path coefficients. In addition, the corrected *R* <sup>2</sup> of all constructs was estimated as a diagnostic tool to evaluate the model fit. The Goodness of Fit (GoF) measure applies the geometric mean of the communality and the average *R* 2 for endogenous dependent constructs. The standard for evaluating the outcomes of the GoF analysis is small (0.02), medium (0.25), and large (0.36) [44]. In this research, a GoF value of 0.390 (see Table 4) shows that the proposed model of the relationship between consumer food attitudes and their assessment of and willingness to try and purchase plant-based and insect-based proteins is large, signifying that the model performs well.


**Table 4.** Model Goodness of Fit (GoF) Index.

Chin et al. [46] argued that an investigator should be able to employ the magnitude of *R* <sup>2</sup> and Stone-Geisser's *Q*<sup>2</sup> value as a criterion for the predictive relevance of a model for a particular construct. The results of *Q*<sup>2</sup> calculations for all the endogenous constructs were greater than zero, indicating that they have satisfactory predictive relevance [44].

Further, Table 4 depicts some results from testing the structural model, indicating that the model does a good job of explaining the variance of willingness to try, buy, and pay more for both meat substitutes. The model explains 33.1% (*R* <sup>2</sup> = 0.331) of the variance of Plant-base willingness and 31.0% (*R* <sup>2</sup> = 0.310) of the variance of Insect-based willingness. However, the model was able to explain 24.3% (*R* <sup>2</sup> = 0.243) of the variance of consumer perceptions of plant-based suitability/benefits compared with only 4.2% (*R* <sup>2</sup> = 0.042) of the insect-based protein suitability/benefits.

#### **4. Results and Discussion**

#### *4.1. Food Neophobia*

Food neophobia inhibits willingness to adopt both meat substitutes (Table 5), fully supporting hypotheses H1a/b, but food tech neophobia only inhibits willingness to adopt for plant-based substitutes, supporting H2b.

Faccio and Fovino [19] made it very clear in their review that the relationship between neophobia and technological innovation in the agrifood industry was complex and required nuance when the concept of neophobia was used outside its original context. Their contention was that a consumer's willingness to try new or unusual food was filtered through their system of norms and values and until new foods or processes become more mainstream, some resistance or avoidance is expected. Our results show that Food Neophobia and Food Technology Neophobia would inhibit consumer willingness to try, buy, and pay more for meat-alternative proteins, however the notion of neophobia by itself might not have been a sufficient indicator to gauge consumers' drivers. The possibility that for some foods, disgust could be a greater influencer than neophobia [16,18,19] should not be overlooked, however the notion of disgust itself was outside the scope of this study.


#### **Table 5.** Direct Path Coefficients.

#### *4.2. Perceived Importance of Meat*

The results also show that meat nutritional importance only inhibited willingness to adopt plant-based substitutes (support for H3b), however meat nutritional importance negatively influenced the perceived suitability/benefits of both meat substitutes (supporting H5a/b). Meat taste/texture/smell importance inhibited willingness to adopt both meat substitutes (supporting H4a/b) and negatively influenced the perceived suitability/benefits of only insect-based substitutes (supporting H6a).

The outcome of hypotheses 3 to 6, examining attitudes towards the importance of meat taste, texture, smell, and the nutritional importance of meat, was consistent with the findings of Schouteten et al. [25] and Mishyna et al. [47].

#### *4.3. Food Choice Values*

The importance of the environmental impact of food choices positively influenced the perceived suitability of both meat substitutes (supporting H7a/b) and the importance of the healthiness of food choices positively influenced the perceived plant-based meat substitutes (supporting H8b).

The importance of healthiness, environmental impact, and suitability of consumers' food choices was examined in hypotheses 7 to 9 and the results support that the food choices were clearly linked with personal values and that these determine the feasibility of a sustainable diet. This is consistent with the information about food choices influencing overall liking [24], that the role meat plays in the diet for many people is beyond its nutritional needs [48], and people rationalize meat consumption [49]. The proposed model included attitudes that were rich in moral implications linked to neophobia values, which offered a multifaceted view of how consumers viewed meat alternatives and their willingness to change their purchasing and consumption habits.

#### *4.4. Behavioral Intension*

Food preference research has found links between food ingredients and consumers' willingness to try them. As such, barriers to trying unfamiliar products is linked to the absence of familiar ingredients and the requirement of a relationship between product and territorial context will determine the adoption of innovation [33]. Similarly, customers are more willing to try novel foods when they contain familiar ingredients, although they are unlikely to pay more for novel products—for example, organic meat, moderation of meat consumption, and sustainable fish are accepted, although willingness to pay more is lower than willingness to consume [36]. Furthermore, the readiness by consumers to adopt insects as an alternative meat ingredient where traditional meat consumption showed that only consumers with a weak attachment to meat would consider trying the insect alternative [37]. In this research, consumer perceptions of the suitability and benefits of insect-based meat substitutes augmented their willingness to try, buy, and pay more for them (supporting H9a). The model was able to account for 31% of the variance of behavioral intention and perceived suitability/benefits of insect-based protein was the dominant predictor of behavioral intention, with a notable non-significant influence of food tech neophobia, meat nutritional importance, and healthiness of food.

For plant-based substitutes, the model performed largely as proposed, explaining 33% of the variance of behavioral intention. The paths suggested that meat nutritional importance and plant-based suitability/benefits are the most important predictors of willingness to try, buy, and pay more for plant-based substitutes (supporting H9b).

#### *4.5. Plant-Based vs. Insect-Based Comparisons*

The literature [25–28] suggests that for many components of the model, plant-based meat substitutes are likely to be considered more suitable and consumers are more willing to adopt them compared to insect-based substitutes. While no specific predictions were made, Table 6 shows the Paired Sample T tests for comparisons between plant-based and insects-based examples for specific items from the suitability/benefits scales and the willingness to try, buy, and pay more scales. For every pair, the plant-based responses were significantly higher than the insect-based responses, which is most likely due to the notion that plant-based meat substitutes are well established in most cultures, while insect-based meat substitutes are still a novelty with a strong stigma attached [26,27].


**Table 6.** Plant-Based vs. Insect-Based Comparisons.


\* = *p* < 0.001.

Overall, we analyzed consumer perceptions with regards to meat and two alternative dietary protein sources in nine very diverse countries: China, USA, France, UK, New Zealand, Netherlands, Brazil, Spain, and the Dominican Republic. We analyzed our data (3091 respondents) as a single global cohort, rather than providing country-by-country analyses. A country-by-country analysis would have provided more granularity in interpretation; however, it would also have created a very complex and potentially confusing discussion. Our global approach to data interpretation does provide a clear insight into consumers' perceptions regarding alternative protein sources.

#### **5. Conclusions**

The findings in this study clearly show that there are differences in consumer attitudes and these influence behavioral intentions towards plant-based and insect-based protein as meat alternatives. To gain more insight into behavioral intentions (willingness to try, buy, and pay a premium), a model was proposed and tested to evaluate the consumers' attitude drivers and determine if plant- and insect-based proteins were realistic meat alternatives. This confirms that consumer adaptation towards sustainable meat alternatives can be complex and is influenced by a diverse set of attitudinal and cognitive-based perceptions.

Our results show that consumer's behavioral intentions towards meat alternatives are inhibited by food neophobia but to a larger extent, are augmented by the perceived suitability and benefits of the protein. The perceived suitability and benefits of the protein alternatives are driven by environmental impact, healthiness, nutritional importance, and sensory attributes for both plant and insect alternatives. Food neophobia and food tech neophobia do not influence the consumer's attitude towards suitability and benefits but have a very clear influence on the behavioral intentions and tend to decrease the willingness to try, buy, and pay more for meat-alternative proteins. The model also shows that consumer attitudes about the environmental impact and to a lesser extent, the healthiness of food, lead to stronger perceived suitability and benefits of plant-based protein. Stronger importance of meat nutrition and to a lesser extent, meat taste, texture, and smell, lead to lower levels of plant-based protein suitability and perceived benefits and lower willingness to try, buy, and pay more for plant-based proteins. For insect-based protein, consumer attitudes towards the suitability and benefits are a strong predictor of willingness to try, buy, and pay more, but those attitudes do not seem to be clearly derived from importance of healthiness, environmental impact of food in general, or their attitudes towards meat. The importance of meat nutritional value is the strongest (negative) influence on perceived suitability/benefits of plant-based protein and willingness to try, buy, and pay more for plant-based proteins, but it only has a small impact on the suitability/benefits of insect-based protein and no impact on willingness to try, buy, and pay more for insect-based proteins.

This study indicates that consumer preferences are influenced by behavioral intentions but does not consider all possible underlying individual attributes such as educational status, knowledge of food and its origins, nutritional values of meat and its alternatives, or the ability to cook a meal. Neither does it consider the potential change in those behaviors with consideration to the importance of, for example, further processing of food ingredients. The contribution of this study is evident by the model created, which is a valuable tool to evaluate what needs to change in consumer attitudes

to alter their behavioral intentions. The consumer's understanding of the nutritional role of meat in their diets and the sensory aspects of meat seem to be pivotal as they influence both attitudes and behavioral intentions.

This study is based on 3091 respondents from nine countries and did not answer the cultural role of meat consumption. Further studies should focus on whether food tech neophobia is a larger driver in more technologically advanced meat alternatives such as fungal-based protein and cultured meat. Further, it is unclear what role culture plays as a driver of consumer attitudes towards meat alternatives, such as whether meat substitutes are more accepted in low meat-eating cultures compared to high meat-eating cultures.

**Author Contributions:** Conceptualization, F.V.; Methodology, F.V., W.d.K., L.K.A., M.O.-G., B.U., B.J., and Z.J.; Supervision, F.V., W.d.K., M.A., and P.M.; Investigation, F.V., L.K.A., M.O.-G., B.U., C.A.G.L., W.H., E.E., Z.J., and A.B.; Data curation, F.V.; Data analysis, D.D.; Writing, W.d.K., D.D., F.V., and L.K.A.; Project Administration, F.V. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### **Sensory and Conceptual Aspects of Ingredients of Sustainable Sources—Finnish Consumers' Opinion**

**Saara Lundén 1, \*, Anu Hopia 1 , Laura Forsman <sup>1</sup> and Mari Sandell 1,2**


Received: 7 October 2020; Accepted: 11 November 2020; Published: 15 November 2020

**Abstract:** Sustainable strategies that enable development of alternative sustainable novel ingredients for food are needed to ensure adequate resources for food in the future. Determining consumer attitudes and acceptance of novel ingredients is essential for wider usage of products including these ingredients. The purpose of the study was to reveal consumers' perspectives on novel, and partly traditional but marginally utilized, ingredients to be used in regular cooking and their sensory characteristics and nutritional and environmental aspects. Consumer attitudes were obtained with two online consumer surveys. Consumer surveys revealed the most interesting ingredients. Plant-based ingredients are preferred over raw materials of animal or insect origin and these are also perceived as more pleasant. Plants were also regarded as credible, ecological, natural, healthy and nutrient-rich. Finnish consumers are not ready to adopt insects into their diet. Neither synthetic meat nor three-dimensional printed food have potential without further knowledge or experience of consumers. Findings of this research give baseline information on consumer attitudes towards novel ingredients. Further research is needed to investigate the perceived pleasantness when the potential ingredients are tasted.

**Keywords:** consumer attitudes; ingredients; pleasantness; sensory; sustainability

#### **1. Introduction**

A crisis of resource sustainability is facing us as a population. A key driver for this is the socio-culturally defined selectivity in consumption habits within developed countries [1]. Sustainable strategies that enable increasing agricultural production and the development of alternative sustainable novel ingredients for food are needed [2]. Determining consumer attitudes and acceptance of novel ingredients of sustainable sources is essential for the commercialization and wider usage of products including these ingredients.

Environmental preservation has become one of the main concerns of consumers [3]. Consumers' interest in environmentally friendly products has grown over recent decades [3–7]. However, regarding food choices, health-related issues and food origin are more preferred reasons than environmental awareness [8]. Other major dichotomies consumers use in categorizing food ingredients are natural vs. unnatural and positive vs. negative [9] which might hinder the acceptance of many underutilized ingredients. Furthermore, consumers perceive ingredients more risky when they are not familiar with them [9,10]. Thus, though consumers might pursue more sustainable food choices, many potential new ingredients suffer from unfamiliarity and various bad images, which makes their application in new products risky for the food producers. Therefore, it is essential to gather understanding about the preliminary images consumers have about new sustainable ingredients and the differences between

these images, in order to opt for the most potentially acceptable ingredients for the development of new sustainable food products.

Insects have been allowed to be used as food in the European Union (EU) since the beginning of 2018 after the Novel Food Regulation (Regulation (EU) 2015/2283) came into force regarding edible insects [11]. Opinions and acceptance regarding insects as food in Western countries have recently gained a lot of academic interest (see e.g., [12–18]). Insects are eaten in numerous countries around the world and insect-based food was estimated to include 2000 species of edible insects [19]. However, Western consumers, such as the Europeans, are only beginning to familiarize themselves with insects as food [14,15]. Rejection of insects as food is mainly caused by disgust which is primarily not based on sensory properties of insects but on knowledge of the history and nature of a potential food [14]. This is supported by the results of Megido et al. [12] who reported high willingness to eat and cook insects as food in the near future after tasting insect preparations. Other predictors for the acceptance of edible insects are previous insect consumption, food neophobia, gender, sensation seeking and food technology neophobia [20]. Willingness to consume insects is found to be culturally relative and differing even in European subcultures [15,21]. Consumers in Northern Europe have a more positive attitude towards insect food compared to consumers in Central Europe [15].

Food wastes and by-products are another possible group of novel sustainable ingredients. A circular economy model can be implemented to the food sector by recycling its by-products and hence creating added value with fewer resources [22]. By-products have been considered as low value and discarded without further processing [23]. Recently, food by-products have been studied as a source of sustainable ingredients or bioactive compounds to be used in functional foods as they can have high nutritional value [23,24].

Three-dimensional food printing has suggested to have implications for future food development. Implications include reducing food waste using food that is usually discarded, such as fruits and vegetables having poor quality in appearance [25]. However, little research has been conducted on consumer perception of three-dimensional-printed (3D-printed) food. Manstan and McSweeney [25] reported a positive attitude towards 3D-printed food, even when compared to a conventional counterpart. Three-dimensional printed food products were believed to be healthier and less processed than conventional food products. Results by Brunner et al. [16] oppose this finding as they found Swiss participants to have a negative overall attitude towards 3D-printed food.

Wild food plants have traditionally been used around Europe, but there has been a dramatic loss of traditional knowledge and practices and the use of these plants in nutrition is very low [26]. However, usage is highly dependent on the region and culinary culture [27,28]. In Mediterranean culinary culture, wild plants are still often used as a part of diet [28]. Food made of cultivated plants and bought from the supermarket appears on the table with relatively little effort, while collecting wild species is more time consuming and season-dependent, thus making them less convenient to be used in everyday cooking [26]. There is a live tradition to use wild berries and mushrooms in Sweden and Finland as these are freely available resources for everyone thanks to legal right of access to private land [26,29]. Wild berries are also used in the food industry and restaurants. Wild plants have recently been promoted by avant-garde restaurants in Northern Europe [26].

When novel ingredients are introduced to consumers, they can potentially cause neophobia, that is fear and refusal of new food [30]. Neophobia limits individuals' readiness to try new foods and thus restricts the marketability of new ingredients [31]. It is possible that neophobia explains the common thread among all these novel foods or ingredients.

The purpose of the study was to reveal consumers' perspectives on novel food sources for discovering new potential raw materials for food products and cooking. Our hypothesis was that consumers differ in their opinions on plant-based and other ingredients. The key motivator for this research was to find out the potential barriers and drivers for these food ingredients. Our context of the research was in attitudes of consumers towards novel ingredients and willingness to try and adopt them in cooking and food products. The ingredients need to be either novel or traditional but marginally utilized. Another important aspect to consider in relation to acceptance of novel food ingredients is consumers' motivation to eat them; such aspects were measured by asking a number of questions related to sensory characteristics and to nutritional and environmental aspects. Furthermore, differences between consumer groups were investigated to achieve a more extensive understanding of the attitudes and to identify possible groups of early adopters of the ingredients.

#### **2. Materials and Methods**

#### *2.1. Preliminary Stage*

The basis of the study was the involvement of consumers in every stage of the process. In the preliminary stage the consumers were engaged in collecting ideas for novel ingredients. This was implemented in a local food fair in autumn 2017 (total number of visitors approximately 20,000). Visitors of the annual food fair were encouraged to write down their ideas on the topic "What are we going to eat from nature 2027?". The list of raw materials (around 100) collected from consumers was supplemented by authors with ideas based on literature and insights from media to increase variation and include some current raw materials. A total of 81 raw materials, presented in Figure 1, were included in the following consumer survey.

#### *2.2. Consumer Survey 1*

A consumer survey was applied to discover the interest of Finnish consumers towards novel raw materials as food ingredients. The consumer survey was distributed as an online survey with Compusense Cloud (Compusense Inc., Guelph, ON, Canada). Randomly, 30 of the 81 raw materials were presented for each consumer in randomized order. Consumers were asked to choose at least five of the presented raw materials that they would be interested in using in cooking or eat. There was no maximum limit for choices. Gender and age were collected as demographic information of the participants.

Adult volunteers participating in the survey were recruited from a consumer register administered by the University of Turku. There was no exclusion or inclusion criteria for participation. Participants replied to the survey anonymously and they were not rewarded with any incentive. The study was conducted following the ethical principles of the Declaration of Helsinki. Consumer survey 1 worked as a pretest for consumer survey 2.

#### *2.3. Consumer Survey 2*

Another consumer survey was implemented in December 2017 to further investigate the opinion of Finnish consumers on these novel ingredients. Based on the results of the first consumer survey the ingredients were chosen for the second survey. Ten raw materials were included in the survey. The list of the 81 raw materials in the first survey also included raw materials which are already regularly in use in Finnish diets. Therefore, seeds and pulse were excluded from the next survey even though they were the most interesting raw materials according to the results. The following five most interesting raw materials were included in the survey: nettle, berry bush leaves, spruce or pine shoots, leaves and stem of broccoli and cauliflower and clover. Root vegetable tops were combined with broccoli and cauliflower parts to broaden the selection of the raw materials. These represent the generally wasted parts of the vegetables commonly used in Finnish food culture. This combination excluded peels which were considered less interesting.

#### *Foods* **2020**, *9*, 1669

**Figure 1.** Raw materials listed in order of rated as most interesting by the participating consumers. Plant-based ingredients are marked with green, animal- and insect-based in orange and either or neither with blue. Darker shade indicates the raw materials chosen to consumer survey 2.

Plant-based raw materials were abundant on the original list of the 81 raw materials. These were also selected as the most interesting ones. Nevertheless, the most interesting raw materials of animal or insect origin were considered justified to be included to the survey to obtain a broader impression on consumers' opinions on the subject. The plant-based novel ingredients considered most interesting based on consumer survey 1 did not include significant protein sources. This was another argument to include raw materials of animal and insect origin. Whey protein and milk protein were found to be the most interesting raw materials of animal origin. These were excluded from the survey based on their prevalence in food manufacturing at present. Crickets, beeswax and ants were the most interesting insect-based raw materials, and they were included in consumer survey 2. Synthetic meat and eggshells were the most interesting raw materials of animal origin. Three-dimensional-printed food was included to the list of raw materials of the second consumer survey. This was argued for, though 3D-printed food is not a raw material, but is a novel method for utilizing both plant-based and animal-based raw materials, as well as for processing by-products into edible food in the future. However, 3D-printed food could be considered equal to the other ingredients because rarely is a whole meal is printed, it is usually parts of it. The 3D printing of food was not common by the time the consumer survey was executed as no 3D-printed food, dish or food ingredient were marketed for Finnish consumers.

Second consumer survey questionnaire was assembled of the questions regarding the raw materials and questions regarding the participants as background variables. A consumer survey was distributed as an online survey with Surveypal (Surveypal Inc., Tampere, Finland). Consumers were given instructions to think about the raw material as an ingredient of a food product or in cooking. Only the name of the raw material was given in the form with no further information. Consumers' willingness to try and opinion on usage and the conceptual properties of the raw materials was gauged with 7-point Likert scale. Prospects of the raw materials were investigated with statements such as "I could eat or cook made of this raw material" and "I would be interested in trying food made of this raw material". The scale was verbally anchored from both ends (1 totally disagree, 7 totally agree). The pleasantness of the sensory characteristics of the raw materials was evaluated with 7-point hedonic scale from 1 (extremely unpleasant) to 7 (extremely pleasant). Participants evaluated pleasantness of appearance, odor and flavor, and feel in fingers and mouth. Each participant evaluated three randomly presented raw materials.

Consumers' attitudes and values were collected as background variables. Attitudes towards new food was studied with the food neophobia scale (FNS) [30]. The FNS consists of 10 items with a seven-category response scale ranging from "disagree strongly" (1) to "agree strongly" (7). Half of the items are reversed, therefore scoring of these items was reversed before calculating the FNS score as a sum of all the item scores. The Finnish translation of the FNS was used with minor revisions in wording as published in a Finnish textbook [32,33]. Participants were divided into three groups based on their FNS score. The three groups were formed following the procedure by Knaapila et al. [34]. Participants with low FNS scores (10–24) were regarded as "food neophilics"—score 25–39 indicates "median group" and score 40–67 indicates "food neophobics". Gender, level of education, part of Finland where the participant lives, type of neighborhood, type of the household and diet were collected with category scales. Participants were asked to inform on whether they grow vegetables, berries or fruits themselves or pick berries, mushrooms or other ingredients from nature for food or cooking (Yes, I grow/Yes, I pick/No, I do not grow or pick). The questionnaire was completed anonymously.

Volunteer participants responding to the survey were recruited by a commercial supplier of consumer surveys. A total of 1014 participants were recruited to obtain adequate amount of replies for each raw material. To ensure variation of the background of the participants, quotas for demographic variables were generated. Usually women tend to participate in the surveys more eagerly than men. In our study, a minimum of 30% male participants was pre-established to secure adequate representation of both genders. In addition, participants from different parts of Finland were recruited from southern, western, eastern and northern parts of Finland at approximately 25% each. Highly educated participants

are often overrepresented in the sample. Therefore, a quota for a minimum of 50% lower educated participants was created. Inclusion criteria were interest to participate and responsibility of groceries of the household alone or together with others. The group of participants was not representative of the Finnish population. The commercial supplier rewarded the participants according to their normal procedure.

#### *2.4. Statistical Analysis*

Results of consumer survey 2 were statistically analyzed using SPSS (IBM SPSS Statistics, 26, IBM Corporation, Armonk, NY, USA). Comparisons between the distribution of the results were performed to analyze differences between samples and respondent groups. Independent samples *t*-test or one-way analysis of variance (ANOVA) with Tukey's or Tamhane's post hoc test were used for variables and groups with normal distribution of categories. Most of the distributions were not normal. Therefore, a Mann–Whitney U-test and Kruskal–Wallis 1-way ANOVA methods with pairwise comparison were applied. The pairwise comparison was performed and significance values were adjusted by the Bonferroni correction for multiple tests. The criterion for statistical significance in all tests was *p* < 0.05.

#### **3. Results**

Participants of both consumer surveys were volunteer Finnish consumers. Demographic information of the participants is presented in Table 1. A more detailed description of the participants is presented in Section 3.1. and Section 3.2.1.

#### *3.1. Consumer Survey 1*

Participants in consumer survey 1 were not predetermined with quotas. A total of 380 replies of volunteer participants was received. In total, 82.8% of the participants were women, 15.4% men and 1.8% other or did not want to specify gender. Consumers who participated were 18 to 81 years old and mean age was 42.9 years. Detailed information of the participants is presented in Table 1.

Plant-based ingredients were the most interesting according to consumers similarly to the type of suggestions. Only 24 of the 81 raw materials were of animal or insect origin and only the 24th of the raw materials in order of the most interesting ones was of animal origin.

Although nettle is well known in Finland [35], it is not commonly used in current cuisine. There is potential for future usage, since 57% of the respondents were interested in using nettle. It was the most interesting of the wild vegetables. Berry bush leaves were almost as preferred as nettle (56%). Black currant leaves are to some extent used in seasoning in certain traditional food and drinks, but the amounts consumed are very small. Pine and spruce shoots are traditionally used as medicine and are, for example, eaten to avoid C-vitamin deficiency. In recent years, small food companies have started to produce food products from spruce shoots, but these are not widely used. Usage of pine or spruce shoots in cooking at homes is very rare. The fourth interesting wild vegetable was clover, which is used in salads, soups or herbal drinks, but the usage is marginal [36].

The following group after wild vegetables was the wasted parts of vegetables. The most preferred raw material representing this group was leaves and stem of broccoli and cauliflower (49%). These are used to some extent together with the other parts of broccoli or cauliflower, but they compose a great amount of wasted food material especially during the domestic season when the prices are lower. Root vegetable tops (41%), fruit seeds (35%) and potato peels (32%) represent the same group of raw materials and were also quite popular.

The most preferred raw materials of animal origin were whey protein (32%) and milk protein (30%). These are already commonly used in food products by manufacturers but not generally used by individuals at home. Synthetic meat (18%) was the first raw material of animal origin after proteins mentioned earlier. In this context it was considered to be synthesized animal cells. Eggshells were the following animal-based raw material and 15% of the respondents regarded them interesting.


**Table 1.** Participants of consumer survey 1 and 2.

Insects were not preferred by the participants. At the time of the survey, insects were not allowed to be sold as food in Finland, but it was decided that legislation would change from the beginning of 2018. Therefore, a lot of news and discussion about insects in the food sector has been underway. Thus, the interest towards insects might have been higher. Crickets were the most interesting insects in the survey and 26% of the respondents were interested in using them in food products and cooking. Beeswax was interesting to 24% of the respondents. Beeswax is used as a food additive—e.g., in coating certain fruits. Ants were the third interesting of the insect-based raw materials but only 12% of the respondents choose it as an interesting one. Ant eggs and mealworms were similarly interesting (12%). Since the beginning of 2018, mealworms have been allowed to be sold as food in Finland but are not widely used.

3D-printed food was equated with raw materials since with this method parts of dishes can be produced. In the printing process either plant- or animal-based ingredients can be utilized. Otherwise wasted materials could, for example, be printed to accepted food products. However, 3D-printed food was not considered interesting by the respondents of the study. It was considered equally interesting as a synthetic meat. This result, together with the top of the list including many wild vegetables and herbs, indicates naturalness as an important factor for consumers when considering the new interesting raw

materials. The same findings indicate familiarity or tradition to be another significant element when choosing new ingredients as food which is also shown in previous research [37]. The top raw materials are traditionally used as medicine or in cooking. They are also commonly found plants on Finns' own yards. Preferring plants over animal-based raw materials might signify either the importance of ecological aspects or unfamiliarity regarding edible insects. Insects were under discussion at the time of the survey and therefore were hypothesized to be trendy. Healthiness together with tastiness are considered when choosing raw materials for food [38]. These above-mentioned factors were included in consumer survey 2 to further investigate consumers' opinion on the subject.

Results of consumer survey 1 were used as screener of the raw materials for the consumer survey 2 as described in Section 2.2.

#### *3.2. Consumer Survey 2*

#### 3.2.1. Participants of Consumer Survey 2

Detailed information of the participants in consumer survey 2 is presented in Table 1. In total, 58.3% of the participants were women, 41.5% men and 0.2% other or did not want to specify gender. Consumers who participated were 18 to 80 years old and mean age was 50.2 years. The majority (87.3%) of the participants had higher than basic education. There was an even distribution of respondents from different parts of Finland. Representatives of city life and countryside were featured. People from households of only adults formed the majority of the respondents; 22.6% of the participants had children in their household. Most (80.5%) of the respondents had a mixed diet. Special diets for different allergies, intolerances or disease or weight control were mentioned as "other". Participants were divided into three groups based on their FNS score as explained in methods (see Section 2.2): 51.3% formed the median group, 28.7% were defined as food neophobics and 20.0% as food neophilics. Picking ingredients from nature or participants growing them themselves was assumed to affect opinions regarding parts of the raw materials in question. Nettle, berry bush leaves, pine or spruce shoots and clover can be picked from nature in Finland and they are available around the country. They also grow in gardens. Furthermore, apples, berries and root vegetables among other food ingredients are grown in gardens. Growing raw materials themselves was assumed to make them more interesting and otherwise affect opinions regarding the conceptual characteristics and pleasantness. Picking berries, mushrooms or other ingredients from nature is quite common among participants; 63.7% reported picking ingredients from nature. Frequency of picking was not predefined. Growing vegetables, berries or fruits was not as common as picking ingredients. Only 28.6% reported growing food raw materials themselves. Amounts grown or area used for growing was not predefined.

#### 3.2.2. Consumers' Opinion on Willingness to Try and Conceptual Characteristics

As the 1014 participants answered the questions regarding three randomly selected raw materials, there were 226–415 responses for each. Distribution of the responses of each statement regarding the possibility to use, willingness to try and the conceptual characteristics are presented in Figure 2A–H. Raw materials are presented in order of the interest according to consumer survey 1 so that the first five are plant-based and next five animal- or insect-based to get a view of the raw material groups based on their origin. According to consumer survey 1, the origin of the raw material (animal/insect or plant) was a significant factor for the respondents, thus it is relevant to examine these groups. Significant differences in the distribution of the responses between raw materials are presented with lower-case letters. Differences are reported with significance level *p* < 0.05.

**Figure 2.** *Cont*.

**Figure 2.** *Cont*.

**Figure 2.** (**A**–**H**) Distribution (%) of the consumer opinion on possibility to use, interest to try and conceptual characteristics of the raw materials. Raw materials with significantly different distributions are marked with different lower-case letters. Significance level *p* < 0.05.

Based on the results of consumer survey 1, differences in consumers' opinions between plant-based and other ingredients were assumed. This hypothesis was not thoroughly verified by the results of consumer survey 2. Beeswax deviated from other raw materials of insect or animal origin. Beeswax is currently used in food as a coating agent for certain foods. However, it is assumed that the majority of average consumers are not aware of this. Distribution of agreement on the statement "I could eat or cook food made of this raw material" was the same with crickets, synthetic meat, eggshells, ants and 3D-printed food. The majority of the respondents at least somewhat disagreed with the statement regarding these ingredients (Figure 2A). Consumers' opinions on nettle and berry bush leaves were the opposite. The majority of consumers at least somewhat agreed that they could eat or cook food made with nettle (70%) or berry bush leaves (71%). Consumers are responsive to nettle, since 42% of the respondents totally agree they could eat or cook food from that. Finnish consumers were not ready to adopt insects into their everyday diet. Only 23% of the respondents to some extent agreed that they could eat crickets and 43% totally disagreed. Similar responses were given for ants—only 17% agreed to some extent and 47% totally disagreed. There was a distinct difference to other raw materials of insect or animal origin in disagreement with the statement. Regarding beeswax, 14% of consumers totally disagreed whereas, regarding ants, crickets, eggshells and synthetic meat, 47%, 43%, 36% and 27% totally disagreed, respectively—i.e., they would not eat the raw material in question. Finnish consumers are not ready to adopt 3D printing as a food manufacturing practice. The majority (62%) of respondents at least somewhat disagreed that they could eat 3D-printed food. They were not even willing to try 3D-printed food. Over a third (36%) of the respondents totally disagreed—i.e., they would not be willing to try 3D-printed food (Figure 2B).

Willingness to try (Figure 2B) shows similar differences for the non-plant-based raw materials, as assessed via the statement "I could eat or cook food . . . " (Figure 2A). Consumers are not willing to try crickets, ants, eggshells, synthetic meat or 3D-printed food. Slight differences in the opinions on the plant-based raw materials were discovered compared to the statement regarding whether they could eat or cook those materials. Nettle and berry bush leaves were considered as the most credible (Figure 2C). For nettle, 73% of the respondents and 69% for berry bush leaves stated at least somewhat agreed to their credibility. These raw materials are already marginally used for food, which might explain the higher credibility. Insects are not seen as credible for usage as food. The allowance of crickets to be sold as food might explain the slightly, though not significantly, higher credibility compared to ants. However, the majority of the respondents consider insects as not credible food; 54% of respondents at least somewhat disagreed that crickets are credible and 66% that ants are credible. Distribution of replies regarding beeswax and credibility is similar to "could eat" and willingness to try. Finnish consumers do not consider 3D printing as a credible technology for food preparation; 3D-printed food was regarded as least credible together with synthetic meat, eggshells and ants. Only 11% of the respondents agreed to some extent that 3D-printed food is credible.

Consumers' opinion on the nutritional value of the raw materials was investigated. Any information about the nutrient content of the raw materials was not given in the questionnaire. Distinction between plant-based and other raw materials was not conspicuous. Nettle was considered as most nutrient-rich (81% at least somewhat agree), significantly different from all others (Figure 2D). According to consumers' opinion, clover was less nutrient-rich compared to nettle and berry bush leaves and beeswax and crickets were considered as nutrient-rich as clover. Consumers did not regard 3D-printed as nutritious food. Synthetic meat and 3D-printed food were considered comparable and the least nutritious compared to the raw materials which are not produced but are derived by growing or as side streams of food preparation.

Nettle and berry bush leaves were also regarded as the most ecological raw materials (Figure 2E). The raw material representing side streams, broccoli and cauliflower stems and leaves and root vegetable tops were regarded as equally ecological compared to wild vegetables apart from nettle. Plant-based raw materials were highly regarded as ecological and more ecological than others; 79–89% of respondents at least somewhat agreed that the plant-based raw materials are ecological. Insects together with beeswax and eggshells formed the next ecological group; 53–63% of respondents at least somewhat agreed they are ecological. Synthetic meat was less regarded as ecological than the two previous groups but more than 3D-printed food. The raw materials which can be picked from the nature were also considered as natural and plant-based raw materials above others (Figure 2F). Only 5–7% of the consumers disagreed to some extent that the plant-based raw materials are natural. Insects were also regarded as natural but significantly less so than the plant-based raw materials. Over half (55–56%) of the respondents regarded insects as natural. As assumed in the wording, synthetic meat was not regarded as natural by consumers. In total, 75% of the respondents disagreed with the statement. Additionally, 3D-printed food was not regarded as natural, as only 82% of the respondents disagreed to some extent with the statement.

Based on the public discussion, it was assumed that insects could be considered trendy. Crickets were one of the first insects approved [11]. Half (51%) of the respondents at least somewhat agreed that crickets are trendy (Figure 2G). However, 23% of the respondents totally disagreed with the idea that crickets are trendy. Wild vegetables, except clover, were considered the most trendy. Eggshells, which are part of Finns' everyday cooking, were considered the least trendy but this was not significantly different from synthetic meat or 3D-printed food.

Nettle was regarded as the most nutrient-rich (Figure 2D) and was also one of the raw materials regarded as the most healthy (Figure 2H). Similarly, synthetic meat and 3D-printed food were considered the least healthy. Beeswax is used as a coating agent and has been reported not to interact with human digestion at all [39]. However, 50% of the respondents at least somewhat agreed that beeswax is healthy. Synthetic meat and 3D-printed food were considered as the least nutrient-rich and also the least healthy; only 10% and 7%, respectively, to some extent agreed that they are healthy.

#### 3.2.3. Consumers' Image of Sensory Properties

Participants evaluated the pleasantness of the raw materials without any additional information given in the question. The responses to the questionnaire were based on either a recollection of the raw material if the person had previous experience of it or an image if the respondent had no experience of the raw material. Appearances of all the plant-based raw materials were evaluated as more pleasant compared to the raw materials of insect or animal origin apart from beeswax. Additionally, 3D-printed food was seen as less pleasant than plant-based raw materials and beeswax. The appearance of clover and berry bush leaves was the most pleasant (Figure 3A). Most (82%) of the respondents regarded the appearance of clover as at least somewhat pleasant. A proportion (71%) of respondents regarded the appearance of berry bush leaves as at least somewhat pleasant. Pine or spruce shoots, nettle, broccoli and cauliflower leaves and stems and root vegetable tops were regarded as pleasant—over 50% of the respondents evaluated these as at least somewhat pleasant. The appearance of beeswax was evaluated as not pleasant nor unpleasant, but the difference to nettle or broccoli and cauliflower leaves and stems and root vegetable tops was not significant. Eggshells were evaluated as slightly unpleasant; 47% evaluated t as somewhat unpleasant. Synthetic meat and 3D-printed food were also regarded as slightly unpleasant and were not significantly different from eggshells. The 3D printing of food has, to date, been uncommon and it was assumed that most of the consumers had no experience of 3D-printed food. Nevertheless, it was assumed that consumers would have thought food can be printed as any kind of form and therefore the appearance was evaluated as pleasant. This assumption was as discovered false. The appearance of ants was perceived as at least somewhat pleasant by 79% and extremely unpleasant by 40% of the respondents and crickets by 73% and 46%, respectively.

Pleasantness of the aroma and flavor (Figure 3B) of the raw materials of different origins deviated similarly as related to the pleasantness of appearance. Beeswax was at the same level as plant-based raw materials in terms of pleasantness of aroma and flavor. Three-dimensional-printed food was evaluated as less pleasant than plant-based raw materials and beeswax. Aroma and flavor of berry bush leaves were evaluated as somewhat pleasant (50%) or as extremely pleasant (27%). Insects, synthetic meat, eggshells and 3D-printed food were the least pleasant. The proportion of responses of the unpleasant categories was significantly larger. Respondents who evaluated the aroma and flavor of insects as pleasant were a small minority; 16% indicated some degree of pleasantness to the aroma and flavor of crickets and only 9% for ants. Over a third (39%) of the respondents evaluated the aroma and flavor of crickets as extremely unpleasant and 43% did so for ants. Participants were not quite as critical about synthetic meat and eggshells. Distribution of these two raw materials was very similar. Pleasantness of aroma and flavor was not significantly different from crickets but was more pleasant compared to ants. Synthetic meat and eggshells were not regarded as having pleasant aromas and flavors. One-fourth of respondents evaluated the pleasantness of aroma and flavor as extremely unpleasant. Only 18% of respondents indicated some level of pleasantness to synthetic meat and 19% to eggshells. Finnish consumers are not familiar with 3D-printed food and opinions regarding this raw material are not as strong. One-third of the participants evaluated the aroma and flavor of 3D-printed food as not pleasant nor unpleasant. However, it was one of the most unpleasant raw materials in the study. One-fourth (26%) of respondents regarded the aroma and flavor of 3D-printed food as extremely unpleasant.

Participants evaluated how pleasant the feel of the raw material in fingers and mouth is (Figure 3C). There was no information about the preparation of the raw material, but the questionnaire regarded

raw materials in food and cooking. Thus, participants could imagine the raw material in question either as raw or prepared in some way. In relation to nettle, some of the respondents might have imagined the plant as raw and for that reason regarded it as very unpleasant. In total, 43% of the respondents regarded the feel of nettle as at least somewhat unpleasant. Berry bush leaves, clover and broccoli and cauliflower leaves and stems and root vegetable tops were the most pleasant raw materials in terms of feeling. Berry bush leaves were evaluated as pleasant by 63% of the respondents, clover by 59% of respondents and broccoli and cauliflower leaves and stems and root vegetable tops by 56% of the respondents. The pleasantness of the feel of beeswax was at the same level with nettle, pine or spruce shoots and broccoli and cauliflower leaves and stems and root vegetable tops. The pleasantness of the feel of raw materials of insect or animal origin together with 3D-printed food were evaluated as less pleasant compared to others. The feel of synthetic meat, eggshells and 3D-printed was equally pleasant. The majority of the respondents evaluated these raw materials as unpleasant; 58% regarded 3D-printed food as unpleasant, 60% did so for synthetic meat and 66% for eggshells. According to Finnish consumers, ants feel the most unpleasant of the investigated raw materials together with crickets. Ants feel unpleasant according to 77% of respondents and crickets do to 81% of respondents.

**Figure 3.** (**A**–**C**) Distribution (%) of the consumer opinion on pleasantness of appearance, aroma and flavor and feel in fingers and mouth. Raw materials with significantly different distributions are marked with different lower-case letters. Significance level *p* < 0.05.

#### 3.2.4. Differences between Opinions' of Consumer Groups

Differences in responses between consumer groups were examined. Consumer groups with different demographic backgrounds were compared. Gender, education level, part of Finland living in, type of neighborhood living in, type of household and diet were used for grouping. Furthermore, differences between groups formed by food neophobia scores were investigated. It was assumed that picking and growing food ingredients oneself could affect opinions regarding investigated raw materials. Therefore, this background information of the respondents was also used to form consumer groups for comparison. The number of the respondents who specified gender as other or did not want to specify gender was small and this group was not compared as a group of gender. The group of other diets was small and heterogenic including diets from different reasons (i.e., weight control and allergies), thus respondents who indicated diet as other were not compared. Number of lacto-ovo-vegetarian and vegans among respondents was low, therefore these groups were not included in the comparison of diets. Results of comparison of respondent groups are presented in Appendix A Tables A1–A11. There were only few significant differences between consumers living in different parts of Finland or representing different types of households. Therefore, results of these groups are not presented in tables, but are explained in writing.

Women were more interested in trying nettle as a food and they were also more willing to eat or cook food with it. They also regarded nettle as more credible, nutrient-rich, ecological, trendy and healthy. There were also differences between age groups regarding attitude towards nettle. Respondents of age 50–64 more strongly, compared to 18–34 and 35–49-year-old groups, agreed they could eat nettle. Younger adults (18–34 years old) considered nettle as less nutrient-rich, healthy and ecological compared to the 50–64-year-old group. They also evaluated the appearance of nettle as less pleasant compared to the 50–64-year-old group. Nettle was most credible to the 50–64-year-old group. Participants with a higher education level were more willing to eat or cook food nettle. Respondents living in rural areas evaluated the aroma and flavor of nettle as more pleasant compared to others. Respondents having plant-oriented mixed diet were more willing to try and eat nettle and evaluated it as more pleasant regarding sensory properties compared to respondents with regular mixed diet. They regarded nettle as more credible, nutrient-rich, trendy and healthy. A food-related closer connection to nature, i.e., growing food oneself, using raw materials or picking them from nature, has an impact on opinions regarding this type of raw material. Respondents who pick food ingredients were more willing to try and eat nettle and consider it more pleasant compared to those who do not pick it. They also regarded nettle as more credible, nutrient-rich and healthy. Food neophobics were less willing to try and eat

nettle and consider it less credible, nutrient-rich, ecological, natural, trendy and healthy. Furthermore, food neophobics evaluated nettle as less pleasant regarding appearance, aroma and flavor, compared to food neophilics and the median group, and feel in fingers and mouth less pleasant compared to food neophilics. Comparison of different consumer groups' opinion on nettle are presented in Appendix A Table A1.

Women were more willing to try and eat berry bush leaves and consider them more credible, nutrient-rich, ecological and trendy (Appendix A Table A2). Moreover, female respondents evaluated the appearance, aroma, flavor and feel of berry bush leaves as more pleasant. The youngest group was significantly different from the 50–64-year-old group regarding whether they could eat, or were interested in trying berry bush leaves, considering whether they are credible or trendy and the pleasantness of feel, whereas both younger age groups were different from the 50–64-year-old group regarding nutrient richness and pleasantness of aroma and flavor. Respondents living in the western part of Finland were not as willing to try berry bush leaves in food as respondents from other parts of Finland (Kruskal–Wallis H = 16.469, *p* = 0.001 with mean ranks of 107.99 for West, 143.50 for South, 159.01 for East and 148.92 for North). Respondents from rural areas evaluated the aroma and flavor and feel of berry bush leaves as more pleasant compared to respondents from the center of large cities. The only difference between respondents from different types of household was in naturalness; respondents from adult households regarded berry bush leaves as more natural compared to single households (Kruskal–Wallis H = 7.633, *p* = 0.022 with mean ranks of 121.70 for single households, 149.16 for adult households and 145.50 for families with children). Respondents who grow food ingredients themselves considered berry bush leaves as more ecological and natural. Furthermore, consumers who pick food ingredients from nature were more willing to try and could eat berry bush leaves, considering them more credible, ecological, natural and trendy, and evaluating them as more pleasant. Food neophilics evaluated the sensory characteristics of berry bush leaves as the most pleasant and healthy. All the FNS groups were different in relation to the statement "I could eat or cook food made of . . . ", "I consider this credible", "This raw material is natural". Food neophilics most strongly agreed and neophobics least strongly agreed with these statements. Food neophobics were less interested to try berry bush leaves as food and consider them less ecological compared to the other FNS groups. Food neophilics regarded berry bush leaves as more nutrient-rich and trendy compared to food neophobics.

Female respondents were more willing to try pine or spruce shoots and evaluate them as more pleasant in appearance (Appendix A Table A3). Furthermore, they regarded this raw material as more credible, nutrient-rich, natural, trendy, healthy and pleasant. The 50–64-year-old respondents, compared to 35–49-year-old respondents, were more willing to try pine or spruce shoots and also more strongly agree with the idea of eating or cooking food made of it. They also considered pine or spruce shoots as more nutrient-rich and ecological compared to others. Younger adults do not consider pine or spruce shoots as healthy as 50–64-year-old people. Representatives of adult households regarded pine or spruce shoots as more ecological compared to representatives of single households (Kruskal–Wallis H = 9.313, *p* = 0.010 with mean ranks of 98.97 for single, 128.43 for adult household and 108.32 for families with children). Consumers having plant-oriented mixed diet regarded pine or spruce shoots as more nutrient-rich and natural. Respondents who grow or pick ingredients for food themselves had more positive attitude towards pine or spruce shoots. Food neophilics were more willing to try and eat pine or spruce shoots, and consider them more credible, nutrient-rich, natural, trendy, healthy and pleasant compared to other FNS groups. Food neophilics and the median group regarded pine or spruce shoots as equally ecological but more than food neophobics. Food neophobics regarded pine or spruce shoots as less ecological than others. All FNS groups were different from each other in terms of whether they could eat or cook, willingness to try, credibility, trendiness and healthiness.

Women were more willing to try the leftover parts of vegetables as food raw material over men and also regarded them as more credible, nutrient-rich, ecological, natural and trendy (Appendix A Table A4). Furthermore, pleasantness of aroma, flavor and feel were evaluated higher among women. The age group of 50–64 years old regarded the leftover parts of vegetables as more credible and nutrient-rich compared to the age group of 35–49 years old. Moreover, they evaluated the leftover parts as more pleasant regarding sensory properties and were more willing to try than the 35–49-year-old group. Respondents with a higher education evaluated the leftover parts of vegetables as looking and feeling more pleasant. Respondents living in the center of a larger city were more willing to try the leftover parts of vegetables and could eat and cook food made of them compared to respondents in rural areas. Furthermore, they regarded this raw material as more credible. Consumers from rural areas did not regard the leftover parts of vegetables as pleasant as others. Representatives of families with children considered this raw material as more ecological (Kruskal–Wallis H = 7.824, *p* = 0.020 with mean ranks of 103.78 for single households, 123.56 for adult households and 136.53 for families with children) and natural (Kruskal–Wallis H = 13.190, *p* = 0.001 with mean ranks of 100.88 for single households, 122.25 for adult households and 144.73 for families with children) compared to representatives of single households. Consumers from adult households evaluated the aroma and flavor (Kruskal–Wallis H = 9.991, *p* = 0.007 with mean ranks of 104.54 for single household, 133.95 for adult households and 109.70 for families with children) and feel (Kruskal–Wallis H = 7.137, *p* = 0.028 with mean ranks of 105.87 for single households, 131.49 for adult households and 113.46 for families with children) of this raw material as more pleasant compared to consumers living in single households. Consumers having a plant-oriented mixed diet were more willing to try and eat the leftover parts of vegetables. They considered this raw material as more credible, nutrient-rich and pleasant in aroma, flavor and feel. Respondents who grow or pick food ingredients from nature themselves evaluated the leftover parts of vegetables as more pleasant and nutrient-rich. There was a significant difference between pickers and non-pickers in willingness to try and eat, credibility, trendiness and healthiness. Leftover parts of vegetables were most pleasant and trendy to food neophilics. FNS groups differed similarly in willingness to try and credibility. Food neophobics had significantly lower agreement to statements "I could eat . . . ", "This is nutrient-rich", "This is natural" and "This is healthy".

There was a significant difference in opinions on clover between consumer groups based on gender in all the studied variables apart from nutrient-rich, trendy and healthy variables (Appendix A Table A5). The youngest group was less willing to eat or cook food made of clover compared to 50–64-year-old consumers. Furthermore, the youngest group regarded it as least pleasant in sensory properties. The 35–49-year-old group regarded clover as the most nutrient-rich and natural. Respondents with higher education level indicated higher willingness to try and eat clover compared to basic education. Moreover, highly educated participants regarded clover as more ecological and natural. Clover is commonly growing in gardens in Finland. It is assumed that this plant is well-known by the consumers living in town houses. This might be one reason why consumers living in the center of a smaller city or municipality were more willing to try and eat clover in food compared to the consumers living in housing estates. There was a parallel difference between these two groups in credible, ecological, trendy and healthy variables. Consumers having plant-oriented mixed diets were more willing to try and eat clover and consider the raw material more pleasant. There was a difference between diets in all the investigated variables. Respondents picking or growing ingredients regarded clover as more nutrient-rich and more pleasant in appearance. The median group was more willing to try and eat clover compared to food neophobics and food neophilics more than median group and food neophobics. Furthermore, there was similar difference between FNS groups in ecological, natural and trendy variables. Food neophobics regarded clover as less credible, nutrient-rich and pleasant in appearance compared to others. Food neophilics evaluated the aroma, flavor and feel of clover as more pleasant compared to other FNS groups.

There was no difference between genders in willingness to try or whether they could eat crickets (Appendix A Table A6). However, male respondents evaluated the appearance and feel of crickets as more pleasant compared to females. The youngest adults (age 18–34) were more willing to try and consider crickets as a possible part of their diet compared to 50–64-year-old participants. Respondents with a higher level of education regarded crickets as more credible and ecological compared to respondents with a basic education. Representatives with a higher or intermediate level of education regarded crickets as more nutrient-rich, natural, trendy and healthy. Crickets were evaluated as looking, smelling and tasting more pleasant by respondents with higher education levels. Consumers living in the center of large cities in Finland were more willing to try and adopt crickets in their diet compared to consumers living in a rural area. Furthermore, they considered them as more credible and nutrient-rich. Representatives of families with children could more likely eat crickets compared to single household representatives (Kruskal–Wallis H = 6.280, *p* = 0.043 with mean ranks of 107.86 for single households, 116.48 for adult households and 136.41 for families with children). Participants with plant-oriented mixed diets regarded crickets as more credible, nutrient-rich, ecological, trendy and healthy. They also evaluated the appearance of crickets as more pleasant. Consumers who pick food ingredients from nature regarded crickets as more nutrient-rich and healthy. Food neophilics were more willing to try and eat crickets compared to the median group and median group more than food neophobics. A similar difference was also in all the other variables except trendy where the median group was not different from food neophobics.

Beeswax was regarded as more ecological and natural by women (Appendix A Table A7). The 50–64-year-old respondents evaluated beeswax as more appealing in all the sensory properties compared to the 35–49-year-old group. Pleasantness of feel of beeswax was evaluated higher also by the oldest age group (65–80 y) compared to 35–49-year-old participants. Representatives of families with children regarded beeswax as more natural compared to single (Kruskal–Wallis H = 7.295, *p* = 0.026 with mean ranks of 105.56 for single, 122.23 for adult household and 132.89 for families with children). Respondents having plant-oriented diet were more willing to try and eat beeswax. Moreover, they regarded it as more credible, nutrient-rich, natural, trendy and healthy. Pleasantness of aroma and flavor was higher according to respondents having plant-oriented diet. Respondents who grow or pick food ingredients themselves are more willing to try and eat beeswax. Furthermore, they consider beeswax as more nutrient-rich, trendy and healthy as well as more pleasant aroma and flavor. Moreover, respondents who pick food ingredients from nature regarded beeswax as more credible, ecological, natural and evaluated it as looking and feeling more appealing compared to those who do not pick. Food neophobics were not as willing to try and eat beeswax as median group and neophilics. They regarded beeswax less credible and natural as well as aroma, flavor and feel of beeswax as less pleasant compared to other food neophobia groups. Food neophobics evaluated beeswax as less nutrient-rich, ecological, trendy, healthy and appearance of beeswax less pleasant compared to food neophilics.

Youngest (18–34 y) consumers were more interested in trying synthetic meat and more willing to adopt it as a part of their diet compared to the 35–49-year-old group (Appendix A Table A8). Furthermore, the youngest consumers regarded it as more credible and the aroma and flavor as more pleasant compared to the 35–49-year-old group. Respondents with a higher education level regarded synthetic meat as more trendy compared to respondents with a basic education. Consumers living in southern part of Finland agree more with the statement "I could eat . . . " than consumers from the north (Kruskal–Wallis H = 10.251, *p* = 0.017 with mean ranks of 130.55 for South, 111.08 for West, 129.76 for East and 98.57 for North Finland). Respondents who stated their place of residence as the center of a large city are more willing to try synthetic meat and regarded it as more credible compared to respondents living in a housing estate or rural area. Consumers from large cities regarded synthetic meat as more ecological compared to representatives of rural areas. Respondents having plant-oriented diets regarded synthetic meat as more credible and natural. Food neophobics are less interested in trying synthetic meat compared to others. Furthermore, they regarded synthetic meat as less ecological, trendy, healthy and pleasant compared to others. Food neophilics regarded synthetic meat as more nutrient-rich compared to others and more credible compared to food neophobics.

Consumers from East Finland could more potentially eat eggshells (Appendix A Table A9) compared to consumers from the north and regarded eggshells as trendier (Kruskal–Wallis H = 9.564, *p* = 0.023 with mean ranks of 111.26 for South, 104.72 for West, 138.31 for East and 105.20 for North

Finland) and feeling more pleasant (Kruskal–Wallis H = 15.828, *p* = 0.001 with mean ranks of 121.79 for South, 100.74 for West, 140.84 for East and 98.68 for North Finland) compared to consumers from the western and northern parts of Finland. Consumers who pick food ingredients from nature regarded eggshells as more natural, healthy and feeling more pleasant and they could more potentially eat eggshells as a part of their diet. Food neophilics are more willing to try and eat eggshells compared to others. Furthermore, they regarded eggshells as more nutrient-rich, ecological, trendy and healthy and looking, smelling and tasting less pleasant compared to other FNS groups. Food neophobics evaluated the pleasantness of feel of eggshells lower compared to food neophilics. Food neophobics regarded eggshells as less credible compared to others. Opinions on naturalness of eggshells were different between all the food neophobia groups.

Men were more willing to try and eat ants compared to women (Appendix A Table A10). Men evaluated the appearance, aroma, flavor and feeling of ants as more pleasant. Furthermore, male respondents regarded ants as more credible, nutrient-rich, ecological, natural and healthy. Participants with higher education levels could more potentially eat or cook food made of ants, evaluated ants as more pleasant and regarded them as more nutrient-rich and trendy compared to participants with a basic education. Respondents with a basic education regarded ants as less credible, ecological, natural and healthy compared to other participants. Respondents living in South or East Finland consider ants trendier compared to respondents from the western part of Finland (*p* = 0.015, mean for West 3.25, East 3.55, South 3.64 and North 3.74). There were no significant differences between consumer groups based on age, place of residence, diet or growing or picking ingredients by oneself. Opinions of food neophobia groups regarding ants were significantly different in all the investigated variables. Food neophobics were less willing to try and eat ants and evaluated them as less pleasant compared to others. Food neophilics regarded ants as more credible, ecological, natural, trendy and healthy compared to others. Food neophilichs considered ants as the most nutrient-rich and food neophobics as the least nutrient-rich.

Men regarded 3D-printed food as more nutrient-rich, ecological and natural (Appendix A Table A11). Furthermore, men evaluated the appearance, aroma and flavor of 3D-printed food as more pleasant. Unlike assumed, the oldest (65–80 y) consumers regarded 3D-printed food as more natural compared to the 35–49-year-old group. Respondents living in adult households regarded 3D-printed food as feeling more pleasant compared to respondents living with children (*p* = 0.018, mean for families with children 2.53, single households 2.76 and adult households 3.11). Contrary to other raw materials in the research opinions of food neophobia groups, these results were more similar. However, food neophilics were more willing to try and eat 3D-printed food and evaluated the appearance as more pleasant compared to food neophobics.

#### **4. Discussion**

In this study we explored the attitudes of Finnish consumers towards possible ingredients for future food. Differences in opinions between female and male respondents were noteworthy. Females were more open to plant-based raw materials and also regarded the conceptual characteristic higher. In accordance with previous research, men were more interested in trying ants and perceived crickets as more pleasant compared to women [40].

The group of 50–64-year-old respondents was more open to plant-based raw materials compared to the youngest group. Differences between education levels were not as comprehensive. A higher level of education indicated more openness to nettle, clover, leftover parts of vegetables and insects. Growing food ingredients by oneself or picking ingredients from nature for food and cooking indicates a close relation to nature and close relation to food ingredients and their origin. This might be a reason for more open attitude towards the raw materials that can be found from nature—i.e., wild food plants and insects. All the raw materials are novel or presumably quite unfamiliar to most of the Finnish consumers since they are not widely used at present. Therefore, it was assumed that food neophobia would contribute to attitudes towards investigated raw materials. This was comprehensively correct for all the investigated raw materials. Furthermore, food neophobia also affects the conceptual characteristics, not only willingness to try, but also the potentiality to use in food and cooking or pleasantness.

Based on the findings of our study, Finnish consumers are open to using nettle and berry bush leaves as a part of their diet. These ingredients were also regarded as the most ecological, natural, trendy, healthy and nutrient-rich by the respondents. This might be explained by the tradition of the use of these ingredients, though the use at present is marginal [26]. Pleasantness or willingness to try and using wild greens have not to our knowledge been studied and these results give valuable insights on their usage as novel ingredients in the future. Reception of clover as food ingredient is not as positive as nettle and berry bush leaves. Nevertheless, 58% of consumers show some degree of interest to try this ingredient. Similarly, as a grass protein, clover might have potential as a protein source for novel foods [41,42]. Leftover parts of vegetables are also perceived as potential novel ingredients by Finnish consumers. Women are more willing to try leftover parts of vegetables. Previous studies have shown women express higher motivation towards avoiding and reducing food waste [43,44], which explains the gender effect.

Finnish consumers are very cautious about synthetic meat, as only 25% of the participants indicated some degree of willingness to try synthetic meat. This is in contrast with previous research on synthetic meat. Almost the same proportion of Belgian participants indicated strong interest in trying cultured meat [45]. Weinrich et al. [46] reported German consumers to be unenthusiastic to try cultured meat while 57% of the participants indicated interest to try. More than half (54%) of the Italian respondents were willing to try cultured meat [47]. However, Belgian, German and Italian respondents received, at the least, basic information about cultured meat before indicating their interest [45–47].

Similarly, Finnish consumers do not express high interest towards 3D-printed food, since 36% of the consumers totally disagree with the statement "I would be interested in trying food made of 3D-printed food". This is in contrast with the results of Manstan and McSweeney [25], where consumers showed higher interest towards 3D-printed food over conventional. Akin to Finns, Swiss consumers have negative attitude towards 3D-printed food [16]. However, well-designed communication has been shown to have the potential to positively shape consumers' attitudes towards 3D-printed food [16]. Finnish consumers did not consider 3D-printed food as healthy, whereas Canadians perceived 3D-printed food as healthier compared to conventional counterparts [25].

Crickets and ants were representatives of insects in this study. Finns are not willing to adopt insects into their diet. Almost half of the respondents disagreed with the idea trying either crickets or ants. This finding is in line with previous research of the attitudes of Western citizens towards insects [17,40,48,49].

The origins of the novel ingredients included in the study varied very much. Some ingredients, e.g., wild food plants, are traditionally used and thus might be more familiar to some consumers whereas other ingredients, such as insects or 3D-printed food, might be very unfamiliar to most. Information of the familiarity or prior knowledge and experience of the ingredients would have given valuable insights for the interpretation of the results, since food exposure and familiarity is shown to affect liking and consumption of food products [37]. Previous research indicated that information changes the attitude towards unfamiliar food [50].

Information of the place of residence was given but not any further specific information about the place where the participants live. This might have also given background to the familiarity of the ingredients of natural sources. Clover, nettle and berry bush leaves commonly grow in gardens and even people living in housing estates or in the city center of a smaller city might have them in their own backyard. This information might have supplemented the information about the familiarity of the ingredients.

Our results were obtained from participants' image of potential novel food ingredients. The image is a pre-existing factor that food producers need to understand when considering which previously unfamiliar or lesser used ingredients to incorporate into new food products or meals. Furthermore, consumers' beliefs and perceptions about certain conceptual characteristics of ingredients, such as

naturalness or nutrient richness, might affect how they are accepted in differing product categories or brand positionings. Further research is needed to investigate the perceived pleasantness of the ingredients which could potentially be used in future food.

#### **5. Conclusions**

Our results showed that consumers differed in their opinions about possible ingredients of edible products and meal. Based on our research, females, 50–64 years old, and neophilic respondents were more open to plant-based materials than others. Study participants were cautious about synthetic meat, 3D-printed food and insects in their diet. The number of consumers with basic education was not equal to other levels, which was a limitation of the study. However, the educational qualifications are high among adults in Finland. Variation of the ingredients led to difficulty of presenting comparable pictures of the ingredients. Although appearance is a significant determinant in the opinion-formation process, we decided not to give any additional information apart from the name of the ingredient. However, participants had the opportunity to search information while completing the questionnaire since the questionnaire was completed online and this was not reported. In general, plant-based ingredients were more agreed to by the consumers regarding conceptual characteristics than the other ingredients. From this point of view, they may have potential for future food ingredients.

**Author Contributions:** Conceptualization, S.L., L.F., A.H. and M.S.; methodology, S.L. and M.S.; software, S.L.; formal analysis, S.L.; investigation, S.L. and M.S.; resources, M.S.; data curation, S.L.; writing—original draft preparation, S.L.; writing—review and editing, L.F., A.H. and M.S.; visualization, S.L.; supervision, M.S.; project administration, M.S.; funding acquisition, M.S., A.H. and L.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Technology Agency of Finland (TEKES), grant number 172/31/2017, the Academy of Finland, grant number 309408 and Turku University Foundation, grant number 4–1507.

**Acknowledgments:** All the consumers are thanked for their participation. Taina Kivelä, Saila Mattila, Merja Nylander, Johanna Paasio, Sari Puputti and Juha Sola are acknowledged for their help during the food fair.

**Conflicts of Interest:** The authors declare no conflict of interest.


### **Appendix A**

**Table A1.** Significant differences between participant groups in distribution of responses regarding nettle. Means (ANOVA and *t*-test) or mean ranks (Mann–Whitney U) are presented. Significantly different respondent groups are marked with different letters. Significance level *p*<0.05.





*Foods***2020**, *9*, 1669














*Foods***2020**, *9*, 1669



*Foods***2020**, *9*, 1669






*Foods***2020**, *9*, 1669



### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Is It All about the Price? An Analysis of the Purchase Intention for Organic Food in a Discount Setting by Means of Structural Equation Modeling**

### **Felix Katt and Oliver Meixner \***

Institute of Marketing & Innovation, Department of Economics and Social Sciences, University of Natural Resources and Life Sciences, A-1180 Vienna, Austria; felix.m.katt@gmail.com

**\*** Correspondence: oliver.meixner@boku.ac.at; Tel.: +43-1-47654-73515; Fax: +43-1-47654-73509

Received: 13 March 2020; Accepted: 4 April 2020; Published: 8 April 2020

**Abstract:** In recent years, discount grocery retailers have expanded their global reach and added to their traditional no-frills offerings to also cater to hedonic consumer needs. In addition to a larger product assortment and a more pleasant shopping experience, they now sell organic food, which commands a price premium compared to non-organic alternatives. To understand organic food in a discount setting, this study sets out to examine the factors that influence discount grocery shoppers' purchase intention for organic food. To study this relationship, this paper tests several factors in a structural equation model, finding a positive relationship between hedonic shopping values, environmental concern, as well as health consciousness and the purchase intention for organic food. In our model, based on a US consumer survey (*n* = 394), price consciousness exhibited a direct and negative relationship with the purchase intention for organic food. Furthermore, this study found that that the impact of environmental concern, health consciousness, and hedonic shopping value is greater on the purchase intention of organic food than that of price consciousness, even in this discount setting. This study concludes by discussing these results and attempting to outline potential areas for future research, as well as managerial implications.

**Keywords:** organic food; discount supermarket; purchase intention; structural equation model (SEM); grocery retailing

#### **1. Introduction**

The success story of discount grocery retailing has been widely studied in the last two decades: from studies on their business model and internationalization strategies [1,2] to consumer price attribution [3], loyalty [4], and shopping value [5]. Historically known for low prices, traditionally undercutting other supermarkets and hypermarkets by 15 to 30 percent [1], limited product assortments, and efficient operations [6–8], the industry is now undergoing substantial changes. Over recent decades, the two leading companies in this industry, Aldi and Lidl, have shaped and cultivated the image of the "hard discounter" with little convenience, products piled in boxes, low price, private label products, and limited investment in technology [9]. With these characteristics, the concept has proven widely successful. From conquering Europe in the 1990s [1,6] to the current expansion overseas [10], both Aldi and Lidl have written a global success story. Through their business models, they have been able to make swift gains of market share in mature markets dominated by supermarkets with a strong brand and service orientation, for example, the United States [1]. Both companies have managed to appeal to what is now a very broad range of consumers, from the initial lower-income bargain hunters to today's "hybrid consumers" [11], with higher incomes and selective spending preferences. Nevertheless, in their European core markets, Aldi's and Lidl's success has begun to

slow down considerably. Therefore, discount grocery retailers have adjusted their business model in recent years—notably, by offering a broader and deeper product assortment, by introducing new (more premium) private label (PL) tiers, and by a steady increase in national brands (NBs) available to shoppers [2,12]. In addition to changes in their product assortment, many discount grocery retailers have moved toward a greener image, from investments in sustainability [13] to offering an increased number of organic food product choices [14,15].

It is thus critical for practitioners and scholars alike to understand what this organic food offering means for discount grocery retailers. The question is whether they can appeal to their core customers with this organic offering and whether they can win over environmentally-minded consumers preferring organic food, who might otherwise not frequent their stores. As consumer demand for organic food has increased in the past two decades, researchers have extensively studied the purchase behavior associated with organic food (see [16] for a systematic review). This behavior, however, has very rarely been studied in a discount retailing context, but rather, from a price-sensitivity [17,18] or frugality [19] perspective. In a study that examined organic food in a discount retail setting, Gottschalk and Leistner [20] found that discount grocery shoppers tend to be more price-sensitive when buying organic food, whereas repeat buyers are more strongly influenced by other product characteristics. They also found that the availability of organic food generally triggers a purchase, hinting at "supply [creating] its own demand" [20]. To add to these findings, this study aims to evaluate the drivers that influence discount grocery shoppers' purchase intention for organic food, building a framework and testing it using a structural equation modeling (SEM) approach. Up until now, no comparable studies focusing on purchasing drivers for organic food within a discount grocery setting and using the SEM approach seem to have been published. Therefore, we aim to contribute to the literature by focusing on purchasing drivers for organic food within a discount grocery setting and using an SEM approach to help better understand customers' purchase intention for organic food in discount grocery stores.

This study is structured into four distinct sections. It commences by outlining the conceptual model developed for this study, as well as the hypotheses to be tested. Next, the methodology employed is detailed, and the experimental design discussed. Thereafter, the results are presented, and the study is concluded with a discussion of the said results, potential future research avenues, managerial implications, and limitations.

#### **2. Conceptual Model**

There are numerous context-dependent factors that can influence consumers' purchase intention and behavior in a shopping situation. To understand how the discount grocery retail setting might affect consumers' purchase intention for organic food items, we draw on commonly cited drivers of purchase behavior traditionally associated with discount grocery retail (price consciousness), common drivers of organic food purchase intention (environmental concern and health consciousness), and an emerging factor in discount grocery retailing—hedonic shopping value. These form the basis of our conceptual model. Such an attitudinal approach to determining the drivers for purchase intention is grounded in the theory of reasoned action (TRA) developed by Fishbein and Ajzen [21], as this theory can be effective in explaining psychological and cognitive antecedents to decision-making [22] as well as attitudes serving as an important predictor of behavioral intention [23]; it is therefore frequently used in the organic food context [24,25]. These four individual drivers, as well as their hypothesized relationship with the purchase intention for organic food for this study's model, are discussed in detail below. Figure 1 summarizes the conceptual framework for this study.

**Figure 1.** Conceptual framework. Note: H<sup>1</sup> to H<sup>6</sup> denote the stated hypotheses.

#### *2.1. Hedonic Shopping Value*

Discount grocery retailers, traditionally known for low prices and a no-frills service offering, have moved from only offering utilitarian shopping value to its consumers to also catering to hedonic needs [5] by offering a more stimulating in-store atmosphere and a broader product assortment. Nevertheless, the question of providing utility and value still forms the traditional basis of discount grocery retail. Utilitarian shopping value can be described as a task-oriented way of shopping, focusing on efficient outcomes without many emotions involved [26]. In the context of discount grocery retailing, this means giving consumers the impression of buying good value at low prices in a shopping environment that is not overly complicated (e.g., by providing limited in-store stimuli). In contrast to this, hedonic shopping value results from fun, enjoyment, and entertainment—that is, "hedonically rewarding shopping experiences are not akin to a negative sense of work" [26]. As organic food is often a more premium (i.e., more expensive) alternative to non-organic options, shoppers may not directly derive value in a cost or efficiency sense. However, with regard to organic food, consumer value perception may be based on a variety of factors other than price, i.e., they may derive hedonic shopping value. Generally speaking, discount grocery shoppers perceive the products they are able to buy to be of good value [27], and in this discount context, the said value perception generally holds a positive effect on purchase intention [10]. Therefore, we aim to understand what effect hedonic shopping value might have on organic food purchase intention in a discount grocery setting as a premium food category like organic could be considered a departure from the low-cost, no-frills commercial strategy. We propose that

**H1.** *Hedonic shopping value has a direct and positive e*ff*ect on the purchase intention for organic food in a discount grocery setting*.

#### *2.2. Price Consciousness and Purchase Intention for Organic Food*

Ever since their inception, discount grocery retailers such as Aldi and Lidl have attracted customers with low prices; their main customer base was, and probably still is, price-conscious consumers [6]. While, in general, grocery consumers tend to base their purchase decisions not only on price, but on a variety of attributes, such as quality [28], price consciousness can play a substantial role in purchase decisions [29], especially for the majority of shoppers who frequent discount grocery retailers [30]. As organic food items are generally more expensive than their non-organic alternatives [31], a high price can not only lead to a decreased purchase intention for organic food [32] but can also potentially act as a barrier to consumption all together [33]. Given the price sensitivity of discount grocery shoppers, we, therefore, hypothesize the following relationship:

**H2.** *Price consciousness has a direct and negative e*ff*ect on the purchase intention for organic food in a discount grocery setting*.

#### *2.3. Environmental Concern and Purchase Intention for Organic Food*

Environmental concern, defined by Schultz and colleagues [34] as a concern about environmental challenges caused by human behavior, is an oft-cited driver for organic food purchase behavior with a positive influence on purchase intention [35], actual purchase behavior [36], and the willingness to pay a price premium [37,38]. This may be explained by the finding that organic food consumers also engage in eco-friendly behavior, such as food waste reduction [39] and recycling [40]. As environmental concern evolved to be a more mainstream topic of public discourse [41], discount grocery retailers started to shift toward a greener image. Given the effects of consumer environmental concerns, we would like to understand its impact on the purchase intention for organic food in a discount grocery retailing setting. Thus, we propose that

**H3.** *Environmental concern has a direct and positive e*ff*ect on the purchase intention for organic food, even in a discount grocery setting*.

#### *2.4. Health Consciousness and Purchase Intention for Organic Food*

Health consciousness can be an important factor in consumer food purchase decisions [42], especially with regard to organic food items. Several studies have not only found health consciousness to be a driver for organic food purchases [43,44], but also a strong influence on consumers' willingness to pay an organic premium [44,45], perhaps explained by consumers perceiving organic food alternatives to deliver greater health benefits [46]. Analogous to environmental concern, we aim to understand the role of health consciousness in the context of discount shopping behavior for organic food. Therefore, we hypothesize the following relationship:

**H4.** *Health consciousness has a direct and positive e*ff*ect on the purchase intention for organic food, even in a discount grocery setting*.

#### *2.5. Other Interactions*

In addition to these main effects, we aim to understand the interaction between the proposed main drivers in a grocery discount setting, i.e., hedonic shopping value and price consciousness, and the interaction between the established main drivers for the purchase intention for organic food. Given that discount grocery retailers main customer base was and probably still is price-conscious consumers [6], we suspect that this diminishes the hedonic shopping value that may be derived for consumers in such a discount setting even in light of the aforementioned move of discount grocery retailers to also cater to hedonic needs [5]. Furthermore, we expect that the other proposed drivers for discount grocery shoppers' purchase intention for organic food exhibit a positive relationship given their importance in organic food consumer behavior in general [16], especially given the findings of Tsakiridou et al. [47], as well as De Magistris and Gracia [48], who uncovered a link between attitudes towards one's health and the environment in organic food consumers. Nevertheless, this might be contrasted by Gschwandtner [31], who found that only health but not environmental friendliness leads to a higher organic food willingness to pay.

We thus propose that

**H5.** *Hedonic shopping value is negatively correlated with price consciousness in a discount grocery setting*.

**H6.** *Environmental concern is positively correlated with health consciousness in a discount grocery setting*.

Altogether, the conceptual framework of this study leads to Figure 1, visualizing the hypothesized relationships between the independent and dependent variables.

#### **3. Methodology**

#### *3.1. Data Collection*

For this study, we employed a questionnaire survey to collect data to analyze the developed conceptual framework. The research setting was the United States, a country with a diverse grocery retailer landscape. Data was collected through Amazon Mechanical Turk (MTurk), a frequently used

crowdsourcing platform for human tasks such as surveys. MTurk allows researchers to anonymously recruit study participants based on pre-selected criteria such as country of residence and is frequently used in consumer survey research in the organic food sector [49]. To ensure reliable and valid results, we followed previous researchers' guidelines [50,51] for recruiting respondents through MTurk and employed measures such as restricting survey-takers to respondents with high MTurk approval ratings. The shopping behavior at discount supermarkets (in this study, i.e., Aldi and Lidl) was self-reported at the end of the survey, leading to an eligible initial sample of 411 participants. This self-reported shopping behavior was elicited by asking respondents to select their most frequented grocery retailers to ensure that discount grocery shoppers were adequately captured. Of this initial sample of 411, 17 respondents were excluded due to failing an attention check, for straight-line answer patterns, or for not completing the majority of the survey, resulting in a final sample of 394 (response rate = 95.9%). The sample is skewed toward the more educated share of the US population, but is in line with the median household income [52] and mean age (38.2 years) [52]. Table 1 provides an overview of the demographics of the sample.


**Table 1.** Sample descriptive statistics.

#### *3.2. Measurement Instruments and Analysis*

To employ valid measurement instruments, the scales used in this study were adapted from previous studies. The scales used were measured on a 7-point Likert scale, with a score of one denoting "strongly disagree" and a score of seven denoting "strongly agree". The statistical analysis was conducted using the software solutions SPSS (Statistical Package for Social Sciences, version 26) and AMOS (Analysis of Moment Structures, version 26, Mount Pleasant, SC, USA). First, an exploratory factor analysis (EFA) was conducted to test the factorial structure of the selected items from the questionnaire to develop reliable multi-item scales. The environmental concern (Cronbach's α = 0.736) scale, adapted from Wei, Ang, and Jancenelle [49] and De Magistris and Gracia [48], was operationalized with four statements, as was price consciousness (Cronbach's α = 0.801), the items of which were adapted from Gil and Soler [53]. Health consciousness (Cronbach's α = 0.902), which was adapted from Michaelidou and Hassan [54], was operationalized with six statements, as was hedonic shopping value (Cronbach's α = 0.955), using the items developed by Babin et al. [26]. Lastly, the purchase intention for organic food was adapted from Yadav and Pathak [55] and operationalized with five statements.

The details of the measurement instruments, as well as their sources, are shown in Table 2. Following the EFA, we conducted a confirmatory factor analysis (CFA), validating the measurement model, as well as testing, fitting, and modifying the structural model. The developed hypotheses were tested by way of standardized regression coefficients (β) and *p*-values (*p*). The results are discussed in detail in the next section.


\* reverse-coded.

#### **4. Results**

Generally, the respondents answered highly for price consciousness (mean = 5.757, SD = 0.912) questions, as well as showing a positive attitude toward their health (mean = 5.581, SD = 0.955) and toward the environment (mean = 4.806, SD = 1.235). The purchase intention for organic food was also slightly positive (mean = 4.838, SD = 1.523). The participants' responses for the hedonic shopping value factor (mean = 4.230, SD = 1.672) were close to neutral. Figure 2 visualizes the responses received, showing a homogenous response pattern for all constructs except for environmental concern, which may be explained by employing two reverse-coded items, which tend to act as cognitive "speed bumps" for respondents, behaving differently from the normal (positively) coded items [56]. The two reverse-coded items were used to reduce social desirability bias for the resulting environmental concern construct, which exhibits good validity (Cronbach's α = 0.736).

**Figure 2.** Overview of responses (*n* = 394). Notes: *y*-axis: number of responses; *x*-axis: 7-point scale (1 = strongly disagree, 2 = disagree, 3 = somewhat disagree, 4 = neither agree nor disagree, 5 = somewhat agree, 6 = agree, 7 = strongly agree); \* reverse-coded items.

#### *4.1. Measurement Model*

The CFA model was operationalized after deleting one environmental concern item (EC2, "I can think of many things I'd rather do than work toward improving the environment"), which improved the average variance extracted (AVE) for the environmental concern construct above the 0.500 threshold proposed by Hair, Black, Babin, and Anderson [57]. The resulting measurement model exhibited good fit indices: χ <sup>2</sup> = 712.034, degrees of freedom (df) = 242, Tucker-Lewis Index (TLI) = 0.924, comparative fit index (CFI) = 0.933, root mean square error of approximation (RMSEA) = 0.070, and standardized root mean square residual (SRMR) = 0.054. The regression coefficients corresponding to all the measurement items were significant (*p* < 0.001). This is in line with the recommended cutoff values for SEM [57,58].

Next, reliability and validity were assessed. Regarding convergent validity, all factors exhibited good composite reliability (CR), with values ranging from 0.772 to 0.956, and higher than the AVE, which for all factors was above the 0.500 threshold. Regarding discriminant validity, all AVEs are greater than the maximum shared variance (MSV), and the square root of the AVEs is greater than the inter-construct correlations. Table 3 provides an overview of reliability and validity measures.


**Table 3.** Overview of reliability and validity measures.

To account for method bias [56], we conducted an additional CFA for the constructs and an added common latent factor. The results show that the latent factor accounts for 24% of the total variance, below the typical method variance found by Williams, Cote, and Buckley [59]. This model also exhibited good fit indices (χ <sup>2</sup> = 707.425 df = 241, TLI = 0.924, CFI = 0.934, RMSEA = 0.070, SRMR = 0.060).

#### *4.2. Structural Model*

The results of the structural model exhibited good fit indices (χ2 = 712.034, df = 242, TLI = 0.924, CFI = 0.933, RMSEA = 0.070, SRMR = 0.054). The first hypothesized relationship was the direct and positive relationship of hedonic shopping value with the purchase intention for organic food. H<sup>1</sup> of our model is supported (β = 0.226, *p* < 0.001). In addition to this relationship, the hypothesized (H3) direct and positive relationship of environmental concern with the purchase intention for organic food was supported (β = 0.128, *p* = 0.016). Similarly, we found a direct and positive relationship between health consciousness and the purchase intention for organic food (β = 0.459, *p* < 0.001), supporting H4. Next, we examined the relationship between price consciousness and the purchase intention for organic food, finding a negative and significant relationship (β = −0.130, *p* = 0.022), thus supporting H2. Lastly, we examined the hypothesized interactions between the exogenous variables. While we did not find empirical support for H5, the hypothesized negative correlation between hedonic shopping value and price consciousness (*r* = 0.014, *p* = 0.809), we found a significant positive correlation between environmental concern and health consciousness (*r* = 0.245, *p* < 0.001), supporting H6. Overall, our predictors managed to explain a sizeable portion of the purchase intention for organic food (*R* <sup>2</sup> = 0.321). Figure 3 and Table 4 summarize the results for the structural model.

**Figure 3.** Results for structural model. Notes: β = standardized regression coefficient; *R* <sup>2</sup> = coefficient of determination; \*\*\* *p* < 0.01; \*\* *p* < 0.05; n.s. = not significant.


Model fit: χ <sup>2</sup> = 712.034, df = 242, TLI = 0.924, CFI = 0.933, RMSEA = 0.070, SRMR = 0.054. Explanatory power of model: purchase intention for organic food: *R* <sup>2</sup> = 0.321.

#### **5. Discussion**

In our study, we set out to understand the purchase intention for organic food in a discount grocery setting. Our findings support most of the hypothesized effects in our proposed model. The effect of health consciousness on the purchase intention is by far the strongest and is generally in line with organic food purchasing behavior found in the existing literature [41,43,60]. Similarly, our findings are in line with the literature for environmental concern and its positive relationship with the purchase intention for organic food [35,36]; this seems to also hold true in this discount grocery shopping setting. The positive effect of health consciousness (β = 0.462) that we found in this discount setting is far stronger than the effect of environmental concern (β = 0.130) on the purchase intention for organic food, perhaps hinting at a stronger self-orientation (toward one's own health) of discount shoppers rather than an altruistic focus (toward the environment). The hypothesized negative effect of price consciousness on the purchase intention was also supported. The effect—especially if viewed in conjunction with the positive relationship we found between hedonic shopping value and the purchase intention—is not overly surprising, as it hints at an underlying notion that organic food is indeed a premium product, and discount grocery shoppers act accordingly, which may also be reflected in the lack of a significant relationship between price consciousness and hedonic shopping value. Overall, the positive effect of hedonic shopping value on purchase intention supports other studies that demonstrate that utilitarian shoppers may move toward hedonic shopping behavior [5]. Additionally, the negative effect of price consciousness is in line with previous studies finding that price can act as a barrier in organic purchase situations [32,61] and holds true in the discount grocery setting of our study. In the said discount grocery setting, this leads us to conclude that price consciousness seems to still be a highly important factor in purchase decisions for shoppers—even for premium products such as organic food [47]. But when comparing the magnitude of the regression weights in our structural model, we find that health consciousness and hedonic shopping value have a far stronger impact on the purchase intention for organic food in this discount setting. This leads us to conclude that while it is an important factor, it is not all about the price.

#### *5.1. Managerial Implications*

The findings of this study hold several implications for discount grocery retailers, which lead to three courses of action. (1) As the relationship between price consciousness and purchase intention was found to be negative, discount grocery retailers might consider looking into the comparative case, that is, further cultivating the image that even premium products, such as organic food, can be bought at a lower price at their supermarkets. Additionally, however, as environmental concern and health consciousness are significant drivers of the purchase intention for organic food, discount grocery retailers might also benefit from specifically promoting these factors, even in this discount setting. (2) Our findings suggest that the greener image that discount grocery retailers have been seeking to portray may be positively received by their customers, as we found a positive relationship between environmental concern and the purchase intention for organic food in the discount setting. The implication for discount grocery retailers can be that these image investments are paying off in a low-price environment and may thus be worth maintaining. (3) And finally, our findings support the notion that addressing the hedonic side of consumers does indeed have merit in the discount setting. Nevertheless, traditional players like Aldi and Lidl should be careful to strike the right balance between a more upscale image and the traditional no-frills approach, especially as other competitors enter this new void in the German home markets by strictly focusing on the traditional no-frills (i.e., hard discount) approach, from which Aldi and Lidl are moving away.

#### *5.2. Contribution and Future Research Areas*

This paper contributes to the existing literature in a number of ways. This study looks at the purchase intention for organic food in a discount setting, linking it to price consciousness. Additionally, in this discount grocery setting, it tests the established drivers of organic purchase intention, environmental concern, and health consciousness. Furthermore, this study adds to the literature [5] by looking at hedonic value in a discount setting. Still, as a result of our findings, we would envision six potential avenues for future research: (1) We would urge researchers to conduct willingness to pay (WTP) studies in a discount grocery context to understand the organic premium consumers are willing to pay in a discount setting. (2) Similarly, and perhaps more broadly, we could envision contingent valuation studies attempting to analyze individual organic food attributes and their corresponding prices, especially in a comparative setting between discount and non-discount (i.e., full-range) grocery retailers. (3) Future studies could attempt to further understand price consciousness in the discount setting by studying its effect on different products and product categories. (4) Future research might also attempt to examine the aspect of hedonic shopping value in light of new entrants in the discount retailing market, who cultivate a more traditional, hard discount offering. (5) Additionally, we could envision other researchers expanding our model by introducing other factors related to the purchase intention for organic food. (6) Lastly, future studies could attempt to replicate the findings of this study using different countries, an experimental design in-store, or actual purchase data.

#### *5.3. Limitations*

In conclusion, some limitations of this study should be mentioned. Firstly, the choice of constructs may be a limitation. While environmental concern and health consciousness are often employed in organic food research [16], price consciousness is traditionally associated with discount grocery retailing [6], and while hedonic shopping value has played a role in the recent discount shift [5], additional and/or other constructs may be of interest in studying organic food purchases in a discount grocery setting. This, however, may be a topic for future research, as outlined in the previous section. Secondly, it should be noted that discount grocery shoppers are not necessarily distinctly different from other grocery shoppers: they may merely exhibit certain traits in a more or less pronounced manner. Lastly, for this study, we recruited US respondents, and we, therefore, acknowledge the possibility that our findings may not hold true for individuals with different cultural backgrounds for multiple reasons, such as scaling biases in the survey items.

#### **6. Conclusions**

In our study, we set out to shed light on the purchase intention for organic food in a discount grocery retail setting. In a structural equation model based on a US consumer survey (*n* = 394), we found a positive relationship between hedonic shopping value, environmental concern, as well as health consciousness and the purchase intention for organic food. Furthermore, our results show that price consciousness exhibits a direct and negative relationship with purchase intention. Additionally, we found that that the impact of environmental concern, health consciousness, and hedonic shopping value is greater on the purchase intention of organic food than that of price consciousness—even in this discount setting.

**Author Contributions:** Conceptualization, F.K.; Data curation, F.K.; Formal analysis, F.K. and O.M.; Investigation, F.K.; Methodology, F.K. and O.M.; Project administration, O.M.; Resources, F.K. and O.M.; Software, F.K. and O.M.; Supervision, O.M.; Validation, O.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** We would like to thank the anonymous reviewers for their valuable comments and the respondents who participated in our survey.

**Conflicts of Interest:** The authors declare that there are no conflicts of interest regarding the publication of this article.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Role of Sensory Appeal, Nutritional Quality, Safety, and Health Determinants on Convenience Food Choice in an Academic Environment**

**Hena Imtiyaz 1 , Peeyush Soni 2, \* and Vimolwan Yukongdi 3**


**Abstract:** The present research aims to investigate the extent to which sensory appeal, nutritional quality, safety, and health determinants influence purchase intention, consumption, and satisfaction of consumers towards convenience food. The non-probability purposive sampling approach was adopted for the recruitment of consumers. A pre-tested questionnaire was used to collect data from 501 consumers. Descriptive statistics, confirmatory factor analysis, and structural equation modeling were adopted to analyze the data. Factor loading, Cronbach's alpha, composite reliability, average variance extracted, and correlations estimate of constructs revealed good internal consistency and reliability of scale items as well as convergent and discriminant validity of the constructs. The path analysis of structural model demonstrated positive relationship between sensory appeal, nutritional quality, safety attributes, healthiness, and purchase intention of convenience food. Further, the path analysis of structural model revealed that purchase intention with consumption as well as consumption with satisfaction were positively associated for convenience food. Sensory appeal was the key determinant influencing purchase intention, consumption, and satisfaction of consumers towards convenience food. The good taste, pleasant appearance, nice smell, and appealing texture within sensory appeal were the most important factors influencing purchase intention, consumption, and satisfaction of consumers towards convenience food. Further, the consumers in emerging economies such as India focus more on sensory appeal in convenience food choice.

**Keywords:** convenience food; determinants; purchase intention; consumption; confirmatory factor analysis; structural equation modeling

### **1. Introduction**

Busy and hectic lifestyles, increase in working population and urbanization, increase in per capita and disposable incomes, diminishing trend of cooking skills and motivation, the rapid expansion of convenience food retail chains, significant improvements in food processing and packaging technologies, and significant change in food-related lifestyles have increased the demand and consumption of convenience food in both developed and emerging economies [1,2]. The global, United States, and European convenience food market is anticipated to rise at the compounded annual growth rate (CAGR) of 4.49%, 4.2%, and 4.5% respectively during 2020 to 2025. The convenience food market in the Asia Pacific region is expected to grow at a CAGR of 8.79% during 2020–2025. The Indian convenience food market generated a revenue of USD 261 million in 2017. It is anticipated to grow at a CAGR of 16.24% during 2019–2024 and reach a revenue of USD 931 million in 2024 [1]. The key market players of convenience food in India are Nestle, ITC, MTR, Capital Foods, CG Food, Haldiram, Bambino, GITS, Kohinoor, Kitchens of India, Maiyas, and Vshodaya [2].

**Citation:** Imtiyaz, H.; Soni, P.; Yukongdi, V. Role of Sensory Appeal, Nutritional Quality, Safety, and Health Determinants on Convenience Food Choice in an Academic Environment. *Foods* **2021**, *10*, 345. https://doi.org/10.3390/foods10020345

Academic Editor: Antti Knaapila Received: 4 January 2021 Accepted: 3 February 2021 Published: 5 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Sensory characteristics such as taste, appearance, freshness, texture, color, and smell are essential motivating factors, driving consumers towards shopping and consumption of convenience food products. Due to advances in food processing and packaging technology, the sensory appeal of convenience food products has been considerably improved in recent years. The sensory appeals undoubtedly are believed to influence consumers' perception, purchase intention, consumption, and satisfaction towards convenience food products significantly [3–10]. Prescott et al. [11] revealed that convenience, sensory appeal, quality, safety, price, and health are the important determinants influencing consumer shopping and consumption of convenience food. However, the magnitude and importance of each determinant may vary across Japan, Taiwan, Malaysia, and New Zealand. Wang et al. [12] revealed that sensory attributes, particularly taste, were the most important motivating factors positively associated with consumers purchase intention of traditional and Western convenience food in mainland China. The quality of convenience food products also drives consumers towards its purchase and consumption. Therefore, it is directly linked to the consumers' perception, purchase decision, and consumption behavior [13–18]. The food quality certification from authorized agencies and brands provide the details of production process, ingredients, nutritional facts, shelf life, cooking instruction, place of production, sensory appeal, quality, safety, and environmental issues which are the primary concerns of consumers while purchasing convenience food [19]. Ojha et al. [20] revealed that highpressure processing (HPP), pulse UV light, and irradiation technologies should be adopted to enhance sensory appeal, quality, and safety of convenience food products.

Food safety, another important determinant, influences the shopping and consumption of convenience food products. The consumers usually expect that the government food regulatory authorities, food processing industries, and marketing agencies take responsibility for the safety of convenience food products. Food safety is one of the most influential factors in terms of shopping and consumption of convenience food products [21–23]. The primary concerns of consumers about food safety are chemical, microbiological, and technological issues as well as the place of origin/place of products [24]. Vital demographic characteristics such as age, gender, education, marital status, and employment status considerably influence the food safety knowledge and practices [25,26]. Misra et al. [27] revealed that application of novel food processing technologies reduced processing time and energy consumption as well as assured high food safety of convenience food products. Health is one of the prime concerns of consumers while purchasing and consuming convenience food products. It is generally believed that regular and excessive use of convenience food causes obesity and other health-related problems [28,29]. Health is a multidimensional construct that influences the purchase intention and consumption of convenience food [4,30,31]. Hoek et al. [32] stated that the government regulatory authorities, responsible for the formulation of food laws and regulations, should prioritize health and health-related attributes of convenience food.

Socio-demographic trends in emerging economies have recently been indicating a major shift. These include more educated and entrepreneur youth population residing in megacities with increased proportion of monthly income on food, lack of time to spend on cooking, multiple income family, and above all the dynamic lifestyle. The confluence of these driving vectors leads to a forthcoming sprawl of convenience food. Several studies in the recent past have been carried out to seek the effect of various factors on convenience food choice, most of which focused on markets in developed and industrialized countries [4,7–9,23,26,31,33]. Due to diversity in tradition, culture, food habits, social structure, religious beliefs, and ethical values, the consumers in emerging economies might not respond to such factors in the same way as reported in aforementioned studies. Hence, it is important as well as timely to administer such research. Considering the impressive market growth and economic importance of convenience food in emerging economies such as India, the main goal of the study is to "examine the role of sensory appeal, nutritional quality, safety, and health determinants on purchase intention, consumption and satisfaction of consumers towards convenience food" in an academic environment.

#### **2. Theoretical Background and Development of Hypotheses**

#### *2.1. Sensory Appeal*

Sensory appeal plays a significant and important role on perception, purchase decision, consumption, and satisfaction of consumers towards convenience food [8,9,12]. Sensory characteristics such as taste [6,8,9]; flavor [8,34]; appearance [5,6]; freshness [5]; texture [6]; smell [6,8]; and overall liking [8] are important motivating factors driving consumers towards shopping and consumption of convenience food. Due to advances in food processing and packaging technologies, the sensory attributes have been improved considerably in recent years to motivate consumers towards convenience food choice [20]. Studies carried out in the past revealed that taste within sensory attributes was the key factor influencing consumer perception, purchase intention, and consumption of convenience food [8,9]. Considering the aforementioned research findings, the present study proposed the following hypothesis:

#### **Hypothesis 1.** *Sensory appeal is positively related to purchase intention of convenience food*.

#### *2.2. Nutritional Quality*

Nutritional quality is another most important determinant, which motivates and drives consumers towards convenience food choice as well as being directly linked with perception, purchase decision, and consumption. Nutritional quality attributes such as nutritional value [15], natural ingredients [16], protein content [16], fiber content [16,17], vitamin content [17], mineral content [16], and nutritional quality certification [12] are the important factors, which drive consumers towards purchase intention and consumption of convenience food. Mascarello et al. [33] revealed that consumer's positive perception towards quality attributes considerably influenced the purchase intention of convenience food. Based on the aforementioned research findings, the following hypothesis is proposed:

**Hypothesis 2.** *Nutritional quality attribute is positively related to purchase intention of convenience food*.

#### *2.3. Safety*

Food safety is another important determinant that influences the purchase decision and consumption of convenience food. Most developed countries have stringent food safety regulation to safeguard the consumers. However, in developing countries like India, food safety regulation enforcement is still in the development stage. The safety attributes such as additives [21,35], pesticides [21,23], hormones [21], color [23], artificial ingredients [23], and safety certification [16] contribute significantly in purchase intention, consumption, and satisfaction of consumers towards convenience food. Based on the aforementioned research findings, the following hypothesis is proposed:

**Hypothesis 3.** *Safety attribute is positively related to purchase intention of convenience food*.

#### *2.4. Health*

Health is the prime concern of consumers while purchasing and consuming convenience food. Health is a multidimensional construct that embodies overall wellbeing of consumers regarding physical, mental, and social aspects [30]. Health-related issues such as calories [36], fat [36], salt [37], sugar [37], and balanced diet [38,39] play important roles in influencing consumers for purchase intention, consumption, and satisfaction towards convenience food. Hoek et al. [32] stated that government regulatory authorities responsible for the formulation of laws and regulations should prioritize health and health attributes of convenience food. Based on the aforementioned research findings, the following hypothesis is proposed:

**Hypothesis 4.** *Healthiness is positively related to purchase intention of convenience food*.

#### *2.5. Purchase Intention, Consumption, and Satisfaction*

The purchase intention of consumers towards convenience is a complex process, and it is governed by a wide range of determinants. However, the importance of each determinant, which drives consumers towards purchase intention of convenience food, depends on foodrelated attitude and behavior. The perceived value of products, which is directly associated with convenience, sensory appeal, nutritional quality, safety, health, and price, has a positive influence on consumers' purchase intention for convenience food [40,41]. Apart from social, cultural, and economic determinants, convenience food consumption is also influenced by convenience, sensory appeal, nutritional quality attributes, safety attributes, healthiness, and price [30,42]. Consumer satisfaction is a strategic focus of consumeroriented food industries and marketing agencies to retain and maintain the consumers for repeated purchase and consumption of their convenience food. Convenience [43], sensory appeal [9], nutritional quality [33], safety [23], healthiness [31], price [44], and physical wellbeing [45] are important attributes of convenience food, which lead to consumer satisfaction and loyalty. In light of the aforementioned research findings, the following hypotheses are proposed.

**Hypothesis 5.** *Purchase intention is positively related to consumption of convenience food*.

**Hypothesis 6.** *Consumption of convenience food is positively related to consumer satisfaction*.

The conceptual model for the current study is based on aforementioned research findings to assess the role of sensory appeal, nutritional quality, safety, and health determinants influencing purchase intention, consumption, and satisfaction of consumer towards convenience food (Figure 1).

**Figure 1.** Conceptual model.

#### **3. Materials and Methods**

#### *3.1. Development, Pre-Testing, and Structure of Questionnaire*

A comprehensive literature review provided guidelines to develop a questionnaire to assess the role of sensory appeal, nutritional quality, safety, and health determinants on purchase intention, consumption, and satisfaction of consumers towards convenience food. Pre-testing is an important step to ensure the accuracy and reliability of the questionnaire [46]. The questionnaire was pre-tested at Sam Higginbottom University of Agriculture, Technology and Sciences, Allahabad, India to develop and optimize the questionnaire. The questionnaires were pre-tested with 30 participants comprising students, in-service professionals, and food processing and nutrition experts to identify and remove potential problems and ensure its comprehensibility. After completing the questionnaire, each participant was asked to give his/her feedback regarding clarity, comprehension, and

potential problems to examine the role of sensory, appeal, nutritional quality, safety, and health determinants on purchase intention, consumption, and satisfaction of consumers towards convenience food. The suggestions made by the participants were included in the final questionnaire to ensure accuracy and precision in data collection [12,38,47,48].

The questionnaire was divided into eight sections. The first section was designed to collect general information of consumers such as socio-demographic characteristics, food habits, food preferences, frequency of eating convenience food, health concerns, etc. The second section of the questionnaire was designed to gather data regarding the sensory attributes (appearance, smell, texture, taste) influencing purchase intention of consumers of convenience food. The third, fourth, and fifth sections of the questionnaire were framed to collect data regarding nutritional quality (nutritive value, mineral, vitamin, natural ingredients, fiber, food quality certificate), safety (hormones, insecticides, pesticides, additives, food safety certificate), and health attributes (calories, fat, salt, sugar, balanced diet) influencing purchase intention of convenience food. The sixth, seventh, and eighth sections of the questionnaire were designed to collect data for purchase intention, consumption, and satisfaction of consumers towards convenience food (Appendix A).

#### *3.2. Participants*

The non-probability purposive sampling method was adopted for the recruitment of the participants because researchers were targeting a specific group of participants, i.e., university/college students, teaching and non-teaching staff, and professionals from corporate sector as they are the major consumers of convenience food consumption [47,49]. A total number of 550 participants were selected from four major cities of Northern India. The total population of four cities is approximately 8.25 million. The sample size of 550 participants taken in this study was higher than 400 as recommended over the population of 0.250 million with a confidence level of 95% and 5% margin of error [47,50]. A total number of 49 questionnaires were dropped due to incomplete information. The final sample size was 501, which resulted in a response rate of 91.09%.

#### *3.3. Data Collection*

The structured and pre-tested questionnaires were distributed to 550 participants in four universities, eight colleges, and twelve corporate offices in January 2019. The participants were requested to gather at the conference/meeting rooms provided by the universities, colleges, and corporate sectors. The participants were informed one day in advance regarding time and venue to achieve desired number of participants as well as to avoid inconvenience. A group of 25 participants were invited to complete the questionnaire. The researcher distributed the questionnaire to the participants and briefed them about purpose, objectives, and importance of the study. The influence of aforementioned determinants on purchase intention and consumption of convenience food were determined on five-point Likert scale (strongly disagree = 1, disagree = 2, don't know = 3, agree = 4, strongly agree = 5). The participants were directed to choose one from 1 to 5 for each question [44,47,51].

#### *3.4. Data Analysis*

The statistical software SPSS version 24 was used to determine mean, standard deviation, skewness, and kurtosis. Further, SPSS was employed to determine Cronbach's alpha to assess internal consistency and reliability of the scale items of questionnaire [47,52,53]. The AMOS software version 23 was used to perform confirmatory factor analysis (CFA) and structural equation modeling (SEM). The CFA was carried out to estimate factor loading, composite reliability, average variance extracted, and model fit indices. The composite reliability of the constructs of the questionnaire was determined to examine the reliability of scale items [38,47,48,52]. The factor loading and average variance extracted were determined to assess the convergent validity of the constructs of measurement model [12,38,39,48,52]. The correlations amongst the construct and square root of average

variance extracted were used to examine the discriminant validity of constructs [54]. The statistical indices such as comparative fit index (CFI), Tucker–Lewis index (TLI), goodness of fit index (GFI), root mean square error of approximation (RMSEA), and standardized root mean-square residual (SRMR) were determined to examine the fit of measurement model [39,47,48,55]. The structural model was constructed to examine the association between sensory appeal, nutritional quality, safety, health, and purchase intention as well as purchase intention with consumption and consumption with satisfaction of consumers towards convenience food. The CFI, TLI, GFI, RMSEA, SRMR, and χ <sup>2</sup>/df (Chi square/degree of freedom) were determined to assess the fit of the structural model [47,48,53]. The standardized estimate (path coefficient), standard error, *t*-value, and *p*-value were determined to test the hypotheses [31,47,48].

#### **4. Results**

#### *4.1. Descriptive Statistics*

Table 1 demonstrates the socio-demographic characteristics of participants. The participants were students and teaching and non-teaching staff from universities/colleges and professionals from corporate sectors. The participants comprised of 41.3% males and 58.7% females with age ranging from 18–65 years (average age = 30.37). The participants consisted of 48.9% single and 51.1% married in which 34.1% and 65.9% were unemployed and employed, respectively. The participants' education level ranged from high school to doctoral, i.e., high school (0.40%), senior secondary school (7.0%), diploma (1.4%), undergraduate (33.9%), master (34.5%), and doctoral (22.8%). The annual family income of the participants ranged from USD 700 to USD 40,000.


**Table 1.** Socio-demographic profile of participants.

*Note:* Total sample size = 501; 1 USD = INR 72. \* 10 = high school; \*\* 10 + 2 = senior secondary school.

The mean participants' score for sensory appeal was higher, followed by safety, nutritional quality, and health determinants influencing purchase intention, consumption, and satisfaction of consumers towards convenience food. The mean participants score of

the items revealed that the "good taste" within sensory appeal construct; "food quality certification" within nutritional quality construct; "food safety certification" within safety construct; and "balanced diet" within health construct were the most important factors in relation to purchase intention, consumption, and satisfaction of consumers towards convenience food (Appendix A; Table 2). The skewness for different items of sensory appeal, nutritional quality, safety, health, purchase intention, consumption, and satisfaction were within the threshold value of −1 to 1. The kurtosis for different items of sensory appeal, nutritional quality, safety, health, purchase intention, consumption, and satisfaction fall within the acceptable range of −2 to 2 (Table 2). The skewness and kurtosis values indicated that participants' score/data recorded for different items of sensory appeal, nutritional quality, safety, health, purchase intention, consumption, and satisfaction were normally distributed [4,53].


**Table 2.** Mean participants' score, factor loading, Cronbach's alpha(α), composite reliability (CR), and average variance extracted (AVE) of product determinants influencing purchase intention, consumption, and satisfaction of consumers for convenience food.


**Table 2.** *Cont.*

\*\*\* Significant at *p* ≤ 0.01; skewness: −1.067 to 0.322; kurtosis: −1.163 to 1.865. Note: See Appendix A for detailed description of the items. Measurement model fit indices: CFI = 0. 911; TLI = 0. 903; GFI = 0. 901; RMSEA = 0.072; SRMR = 0. 074.

#### *4.2. Measurement Model*

The factor loading of all items of sensory appeal, nutritional quality, safety, health, purchase intention, consumption, and satisfaction for convenience food were significant (*p* ≤ 0.01). The factor loadings for different items of sensory appeal, nutritional quality, safety, health, purchase intention, consumption, and satisfaction constructs ranged from 0.608 to 0.963, which were higher than the threshold value of 0.50 [39,48,52,55], hence all items were included for the interpretation of the factors influencing purchase intention, consumption, and satisfaction of consumers towards convenience food [39,48,55]. Cronbach's alpha for sensory appeal, nutritional quality, safety, health, purchase intention, consumption, and satisfaction constructs ranged from 0.740 to 0.897, which exceeded the threshold value of 0.70 [38,47,52]. Composite reliability for sensory appeal, nutritional quality, safety, health, purchase intention, consumption, and satisfaction constructs ranged from 0.852 to 0.979 that exceeded recommended minimum cut off value of 0.70 [47,48,52]. Cronbach's alpha and composite reliability values obtained for different constructs revealed good internal consistency and reliability of scale items of questionnaire [3,48,54,55]. The average variance extracted for sensory appeal, nutritional quality, safety, health, purchase intention, consumption, and satisfaction constructs ranged from 0.521 to 0.864, which were higher than the minimum acceptable cut off value of 0.50 [39,48,54]. The factor loading and average variance extracted values obtained for different constructs and items for each construct confirmed the convergent validity of the constructs of measurement model [39,54,55]. The square root of average variance extracted estimates (diagonal values) were higher than the correlation estimates amongst constructs (Table 3), which confirmed the discriminant validity of constructs [47,48,54].

The comparative fit index (CFI), Tucker–Lewis index (TLI), goodness of fit index (GFI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR) were used to examine the fit of measurement model relating sensory appeal, nutritional quality, safety, and health aspects with purchase intention, consumption, and satisfaction towards convenience food. The CFI was 0.911 (≥0.90); TLI was 0.903 (≥0.90); GFI was 0.901 (≥0.90); RMSEA was 0.072 (≤0.08), and SRMR was 0.074 (≤0.08), which were within the acceptable range (Table 2). The CFI, TLI, GFI, RMSEA, and SRMR revealed that measurement model fit well with data [47,53,55,56].


**Table 3.** Discriminant validity of the constructs.

#### *4.3. Structural Model*

The structural model was constructed to examine the association between sensory appeal, nutritional quality attributes, safety attributes, healthiness, and purchase intention as well as purchase intention with consumption and consumption with satisfaction of consumers towards convenience food. The CFI was 0.913 (≥0.90), TLI was 0.906 (≥0.90), GFI was 0.903 (≥0.90), RMSEA was 0.073 (≤0.08), SRMR was 0.075 (≤0.08), and χ <sup>2</sup>/df was 3.9 (<5.0), which were within the recommended acceptable range (Figure 2). The CFI, TLI, GFI, RMSEA, SRMR, and χ <sup>2</sup>/df values demonstrated a good fit of the structural model [39,47,53,56].

The results of the structural model presented in Figure 2 and Table 4 demonstrate the extent of the relationship among sensory appeal, nutritional quality attributes, safety attributes, healthiness, and purchase intention, as well as purchase intention with consumption and consumption with satisfaction for convenience food. Hypothesis 1 (H1), which proposed positive relationship between sensory appeal and purchase intention of convenience food was accepted, because standardized estimate (ß) of the path of structural model was significant (Hypothesis 1: ß = 0.788, S.E. = 0.053, *t*-value = 5.448, *p* ≤ 0.01). Hypothesis 2 that postulated positive relationship between nutritional quality attributes and purchase intention of convenience food was accepted because standardized estimate (ß) of the path of structural model was significant (Hypothesis 2: ß = 0.639, S.E. = 0.056, *t*-value = 6.094, *p* ≤ 0.01). Hypothesis 3, which postulated positive relationship between safety attributes and purchase intention of convenience food, was accepted as the standardized estimate (ß) of the path of structural model was significant (Hypothesis 3: ß = 0.511, S.E. = 0.032, *t*-value = 16.063, *p* ≤ 0.01). Hypothesis 4 that proposed positive relationship between healthiness and purchase intention of convenience food was accepted, because the standardized estimate (ß) of the path of structural model was significant (Hypothesis 4: ß = 0.491, S.E. = 0.031, *t*-value = 15.594, *p* ≤ 0.01). Hypothesis 5, which postulated positive relationship between purchase intention and consumption of convenience food was accepted because standardized estimate (ß) of the path of structural model was significant (Hypothesis 5: ß = 0.998, S.E. = 0.016, *t*-value = 61.962, *p* ≤ 0.01). Further, Hypothesis 6 that proposed positive relationship between consumption and satisfaction towards convenience food was also accepted (Table 4) as standardized estimate (ß) of the path of structural model was statistically significant (Hypothesis 6: ß = 0.728, S.E. = 0.022, *t*-value = 32.516, *p* ≤ 0.01).

≤

≤

≤

≤

≤

χ **Figure 2.** Structural equation modeling to assess the role of product determinants on purchase intention, consumption, and satisfaction of consumers towards convenience food. Structural model fit indices: CFI: 0.913; TLI: 0.906; GFI: 0.903; RMSEA: 0.073; SRMR: 0.075; χ <sup>2</sup>/df = 3.91.

**Table 4.** Structural model results to examine the association of between product determinants and purchase intention, consumption, and satisfaction for convenience food.


\*\*\* Significant at *p* ≤ 0.01.

#### **5. Discussion**

The sensory appeal plays a significant role in driving consumers towards shopping and consumption of convenience food. The mean participants' score of the sensory appeal construct and the standardized estimate of the path of structural model revealed that sensory appeal was the most important determinant influencing purchase intention, consumption, and satisfaction of consumers towards convenience food (Tables 2 and 4). Further, the mean participants' score of the items indicated that taste was the key factor influencing purchase intention, consumption, and satisfaction of consumers towards convenience food as compared to appearance, smell, and texture. Previous studies carried out under a wide range of social, cultural, and economic conditions also predicted sensory appeal as the most important determinant influencing shopping and consumption of convenience food [4–7,16,57]. Previous findings revealed that convenience, sensory appeal, nutritional quality, price, and health are important determinants influencing convenience food choice; however, the magnitude and importance of each determinant varied significantly across the social, cultural, economic and food related lifestyle [3,11,58].

In recent years, consumers have been more concerned about the quality and safety of convenience food. The development of novel and advanced food processing technologies such as high-pressure processing (HPP), pulse UV light, and irradiation has improved the quality of convenience food significantly [20,27]. The standardized estimate of the path of structural model indicated that nutritional quality attributes positively influenced purchase intention, consumption, and satisfaction of consumers towards convenience food (Table 4). The mean participants' score indicated that food quality certification from a food regulatory agency was the most important among the factors under food nutritional quality construct, which drives consumers towards purchase intention and consumption of convenience food (Table 2). The findings of the previous studies also indicated that consumer perception towards quality attributes significantly influence purchase intention and consumption of convenience food [14,33,57]. Petrescu et al. [59] revealed that Belgian and Romanian consumers assign high values to quality attributes and often use taste, appearance, and freshness as an indicator to assess the quality of convenience food. The present findings also indicated that taste, appearance, and smells were key factors influencing convenience food choice, but these factors were considered under sensory appeal of convenience food (Table 2).

Food safety is another important aspect of convenience food that is directly associated with public health, food security, environmental protection, and sustainable development. The analysis of the structural model demonstrated that the safety attribute was positively associated with purchase intention, consumption, and satisfaction of consumers towards convenience food (Table 4). The mean participants' score revealed that food safety certification was the most important factor under the food safety construct which drives consumers towards purchase and consumption of convenience food (Table 2). Previous studies also reported that food safety is an influential factor, which drives consumers towards convenience food choice [21–23,26]. The novel food processing technologies, i.e., HPP, pulse UV light, and irradiation, could be utilized by food processing industries in the production process to improve food quality and safety standards of convenience food [20,27].

The health benefit greatly influenced consumers towards convenience food choice. Due to technological development in processing, preservation, storage, and marketing, the sensory appeal, nutritional quality, safety, and health attributes of convenience food have been improved significantly in recent years [20,27]. The results of the structural model demonstrated the positive association between healthiness and purchase intention of convenience food (Table 4). The mean participants' score of the health construct as well as different items within the health construct revealed that consumers are satisfied with the healthiness of convenience food. The overall results of the present study showed that the convenience food products are perceived as healthy and their consumption does not pose any threat to health (Table 2). This is due to the fact that the consumption of convenience food is not excessive, therefore the consumers have not reported any diverse

effect of consuming convenience food on health. In contrast, consumers in developed and industrialized countries believe that consumption of convenience food has implication on health, diet quality, obesity, and chronic disease risk [32]. Vita et al. [57] revealed that high salt content, high fat content, and presence of nitrites had a negative impact on purchase intention of processed ham, but good taste, pleasant color, and juiciness diminishes the effect of aforementioned unhealthy compounds, which strongly support the findings of the present study in which consumers assign high values to sensory attributes as compared with quality, safety, and health attributes of convenience food.

In recent years, convenience food has spread into the lifestyle of consumers in emerging economies such as India due to time scarcity, competitive environment, and significant changes in food-related lifestyle [1,2]. Food production, processing, distribution, consumption, and waste disposal contribute largely to emission of greenhouse gasses, resources depletion, global warming, and environmental degradation [60–62]. Hence, environmentally sustainable food production, distribution, and consumption is important for sustainable development. Environmentally sustainable food consumption is the foremost important step to minimize the use of natural resources and emissions of greenhouse gasses, toxic waste, and pollutants, which in-turn enhance sustainable development and quality of life [63–65]. Convenience food involves production and transportation of raw materials, pre-processing manufacturing, packaging, distribution, consumption, and waste disposal which can be optimized in order to minimize the environmental degradation [66]. Food consumption behavior of consumers is a complex process and is strongly associated with lifestyle and socio-cultural environment. The consumers may express environmental concern, but during buying process normally ignore purchasing environmentally sustainable food products due to convenience, time pressure, availability, and price [67–69]. Previous studies carried out in developed and industrialized countries revealed that consumers should be encouraged to purchase environmentally sustainable convenience foods such as organic and minimally processed food to minimize negative effect on human health and environment [70]. Schmidt Rivera et al. [66] revealed that the environmental impact of ready to eat food was higher than equivalent home-cooked food. Further, consumers should be educated and encouraged to curtail ready-to-eat convenience food and consume more home-cooked food. In order to promote sustainable food consumption, consumers should be encouraged to purchase and consume plant-based foods because animal-based foods are more resource intensive and less environment friendly [71–73]. Environmental sustainability has become a severe concern, especially in developed and industrialized countries but in India, where the present study has been conducted, the consumers' concern towards environmentally sustainable food production and consumption is insignificant. Sharma and Jha [74], in their study conducted in India, revealed that consumers' individualism was negatively associated with sustainable food consumption. Government food regulatory bodies, non-government organizations, social and environmental activists, and policy makers should encourage and promote environment sustainable production, processing, distribution, and consumption of convenience foods in emerging economies such as India.

Though the present study provides in-depth knowledge and information regarding the role of sensory appeal, nutritional quality, safety, and health determinants on convenience food choice in emerging economies like India, the present study has some limitations. Due to time constraints, the study was carried in four cities in India which limits the generalization of the findings. Hence, future research should be carried out in different cities and countries in order to obtain more generalized and representative results. The present study concentrates on specific groups of consumers which also limits the applicability of the results. Therefore, future research should include wide range of consumers to improve overall applicability of the results. School children constitute an important consumer segment for convenience food. Hence, it is recommended to carry out similar studies for school children across cities and countries to provide them safe and healthy convenience food. Since environmental sustainability is a matter of grave concern, it is recommended to incorporate the environmentally sustainable aspects of convenience food in future studies in emerging economies, especially in India. For instance, packaging size of the convenience food would determine the amount of waste (packaging material after use) to be disposed of by a city. Similarly, it will influence the city's environmental footprint in terms of waste collection and disposal. The findings of such studies would definitely enhance the knowledge and understanding about consumers' purchase and consumption behavior towards convenience food in emerging markets.

#### **6. Conclusions**

The confirmatory factor analysis results indicated satisfactory and acceptable value of reliability of scale items and validity of the constructs of questionnaire. The model fit indices revealed that measurement and structural model relating sensory appeal, nutritional quality, safety attributes, and healthiness with purchase intention, consumption, and satisfaction of consumers towards convenience food were fitted well with data. Sensory appeal, quality attributes, safety attributes, and healthiness have a positive relationship with purchase intention, consumption, and satisfaction of consumers towards convenience food. Sensory appeal such as good taste, pleasant appearance, nice smells, and pleasant texture play the most important role in motivating and driving consumers towards purchase intention and consumption of convenience food. The overall result reveals that consumers give more importance to sensory appeal as compared with quality, safety, and health attributes during the purchase and consumption of convenience food in emerging economies such as India.

The conceptual framework and findings provide some theoretical and practical contributions. First, to the best of the authors' knowledge, the present comprehensive study expands previous research by adding consumer satisfaction to a conceptual model relating sensory appeal, nutritional quality attributes, safety attributes, and healthiness with purchase intention, consumption, and satisfaction for convenience food. Second, the empirical evidence reveals that consumers in emerging economies assign high values to sensory appeal in convenience food choice, compared to quality, safety and health attributes, which shall add new information to literature. Third, the food processing industries should ensure that convenience foods are free from hormones, insecticides, pesticides, non-permissible additives, non-permissible colors, and artificial ingredients during production, processing, transportation, and marketing of convenience food to minimize health risk. Fourth, food processing industries should ensure the recommended level of calories, salt, sugar, and fat content in convenience food to provide a healthy and balanced diet to consumers. Finally, government food regulatory agencies should have strict food laws and regulations for mandatory food quality and safety certification to enhance consumers trust on convenience food.

**Author Contributions:** Conceptualization, H.I., P.S., and V.Y.; methodology, H.I.; formal analysis, H.I.; investigation, H.I.; resources, H.I.; data curation, H.I.; writing—original draft preparation, H.I.; writing—review and editing, P.S. and V.Y.; supervision, P.S. and V.Y. All authors have read and agreed to the published version of the manuscript.

**Funding:** No external funding was received for the research project. The first author received a fellowship to pursue doctoral degree. The research project was the part of the doctoral degree program.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Doctoral Progress Review Committee of the Asian Institute of Technology.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Acknowledgments:** The first author is grateful to the School of Environment, Resources and Development, Asian Institute of Technology, Bangkok, Thailand for providing fellowship to pursue the doctoral degree in Agribusiness Management.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**

Description of the questionnaire.



#### **References**


### *Article* **Assessing Consumer Preferences for Suboptimal Food: Application of a Choice Experiment in Citrus Fruit Retail**

**Wen-Shin Huang 1 , Hung-Yu Kuo 2 , Shi-Yuan Tung <sup>2</sup> and Han-Shen Chen 2,3, \***


**Abstract:** Amid the trend of sustainable development, reducing food waste is a global concern and campaigns to reduce food waste have been launched. For example, the term "food sharing" has originated from Germany and promotes sharing food instead of wasting. "The Guerilla Kitchen", which originated from Netherlands, is an organization that also promotes avoiding wasting food. Consequently, more and more people are paying attention on this issue and we think it is necessary to understand people's acceptance of suboptimal food, as discarded suboptimal food represents a significant proportion of food waste. Additionally, at least one-third of the food globally produced each year is classified as suboptimal and cannot be sold in the market because of a poor appearance, damaged packaging, or near expiration date, thus presenting challenges for environmental, social, and economic sustainability. Previous studies on suboptimal food have focused more on appearances and packaging dates and less on investigating traceable agricultural and price discounts, which is where food classified as suboptimal entails a discount. Moreover, citrus product attributes such as appearance, size, freshness indicators, traceable agricultural products, and price discounts were determined in terms of consumer preference through pre-measurement here, then using a choice experiment method to clarify which attributes consumers care about most (*N* = 485 respondents). Conditional logistic regression and a random parameter logit model (RPL) are employed to examine the various properties of a marginal willingness to pay (WTP). RPL was also used to deduce the respondents' choices based on differences in appearance and freshness indicator. The results showed that consumers place greater emphasis on the freshness indicators (harvesting/packaging date labels) and appearance of suboptimal citrus fruits but do not focus on the size. Consumers are willing to purchase citrus fruit with a flawed appearance, although the price needs to be reduced from the original price. Although suboptimal food does not reduce health, people may still not buy it and this result in food wastage. As a result, it is essential to increase awareness regarding suboptimal foods and reduce food waste to support sustainable development.

**Keywords:** food choice motivations; food waste; willingness to pay; consumer behavior

#### **1. Introduction**

According to United Nations statistics, nearly one-third of the total international output of food is wasted every year, and the annual cost of food waste disposal is as high as 940 billion US dollars [1]. Kretschmer et al. (2013) [2] highlighted that according to the data of the US Food and Agriculture Organization in 2013, about three-quarters of food is wasted at production sites, households, and restaurants, and household food waste has increased over time. About 15.9% of food waste in the USA comes from the consumer, and USA households produced 27 million tons of food waste in 2015 [3]. Additionally, on the basis of food waste, the Environmental Protection Administration of the Executive

**Citation:** Huang, W.-S.; Kuo, H.-Y.; Tung, S.-Y.; Chen, H.-S. Assessing Consumer Preferences for Suboptimal Food: Application of a Choice Experiment in Citrus Fruit Retail. *Foods* **2021**, *10*, 15. https://dx.doi.org/ 10.3390/foods10010015

Received: 4 December 2020 Accepted: 22 December 2020 Published: 23 December 2020

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/ licenses/by/4.0/).

Yuan (2018) [4] estimated that about 3.68 million tons of food materials are consumed in Taiwan every year, and an average of 158 kg is wasted per person, of which nearly 50% is discarded by individuals and families. These numbers were higher than those of other Asia-Pacific countries, thereby indicating that Taiwanese people waste too much food [4]. Stuart (2009) [5] and Bilska et al. (2016) [6] suggested that the proportion of waste at the supply chain is often higher than that at the consumer side and that the sources of food waste in the supply chain include mislabeling outer packaging, poor product appearances, being near the expiration date, and outer packaging damage. Lebersorger and Schneider (2014) [7] pointed out that a significant proportion of food is wasted at the retail stage that is disposed while in good condition, and said food is only discarded based on an expiration date that has been passed (e.g., in Austria, more than a quarter of discarded food and products are suboptimal products). Lombart and Louis (2014) [8] argued that selling suboptimal products can lead to a positive effect on the image of the store as a being responsible stakeholder, with the potential to influence consumer's retail preferences or loyalty. Plazzotta et al. (2017) [9] pointed out that fruit and vegetable waste is mainly generated before reaching consumers due to programmed overproduction and the unfulfillment of retailer quality standards. Therefore, retailers or companies may not purchase products that cannot reach their standards. Aschemann-Witzel et al. (2017) [10] stated that more than one million people in the world experience chronic hunger every year and some regions even face food crises, resulting in a serious global food imbalance. Aschemann-Witzel et al. (2015) [11] stated that food waste has an impact on environmental, social, and economic sustainability. Thus, food waste has been regarded as a moral issue due to global inequality in food access and growing food security issues [11–13], and it is indeed necessary to solve food waste challenges to achieve sustainable development in food supply chains [14].

The Homemaker's United Foundation (2016) [15] highlighted that suboptimal food, which is unmarketable because it does not meet traditional aesthetic standards, mainly includes vegetables, fruits, and meat. These items usually do not procure good sales due to their poor appearance and can only be used as feed, fertilizers, canned juices, donations for food banks, or even thrown away. Suboptimal food is divided into three categories on the basis of its characteristics, namely, appearance standards (e.g., weight, shape, and size are required to meet the ideal standards) [16], the marked expiration date (e.g., food approaching or exceeding its expiration date) [17], and the packaging (e.g., food packaging exhibiting visual damage, such as a dented can or a torn wrapper) [17]. Additionally, it is necessary to confirm that these categories/aspects do not pose any safety risks and that the food is still appropriate for human consumption [11,18,19].

Dion et al. (1972) [20] stated that when selecting foods, most consumers select foods with a perfect appearance and shape, undamaged packaging, and a long shelf life, thus resulting in suboptimal food that cannot be sold. Göbel et al. (2015) [18] suggested that the main reason for the waste of vegetables and fruits is the influence of retailers on product quality standards and specifications. White et al. (2016) [17] revealed that consumers may choose less suboptimal food because of factors such as an abnormal shape, damaged packaging, or marked date of expiration. Helmert et al. (2017) [21] also revealed that only a few consumers will choose suboptimal food when the quality or safety of suboptimal food is similar to that of optimal food. Symmank et al. (2018) [22], with the help of an example, stated that the appearance of bananas will affect German consumer purchase intentions and that they attach importance to the shelf life of bananas. Aschemann-Witzel et al. (2018) [23] highlighted that printing requests/instructions such as "no food waste" on food packaging can increase the possibility of Uruguayan consumers choosing suboptimal food. Halloran et al. (2014) [19] highlighted that food waste could be reduced through communication and improved food packaging and labeling.

Additionally, consumers will consider the price when choosing products. Helmert et al. (2017) [18] pointed out in their research that price badges can influence the attention of European consumers, cognitive processing, and purchase intentions regarding suboptimal

food. Many retailers in Europe offer products close to the shelf life at lower prices to attract consumers [11]. According to Grewal et al. (1998) [24], retailer price discounts can affect consumer purchase intentions. Verghese et al. (2013) [25] also pointed out that the precondition for consumers to buy suboptimal food is a price discount. The study of de Hooge et al. (2017) [26] stated that consumers in Denmark, Germany, Norway, Sweden, and the Netherlands are willing to buy suboptimal food with price discounts. Aschemann-Witzel (2018) [27] pointed out that when the price of suboptimal food is reduced, consumers may have more incentives to buy them. De Pelsmacker et al. (2005) [28] believed that when discussing consumer purchase behaviors, they cannot be judged solely by their attitude and preference toward the product but must be analyzed in terms of their purchase intention or willingness to pay (WTP). The so-called WTP is the amount that consumers are willing to pay for a product that they think is most appropriate [29]. Tsiros and Heilman (2005) [30] revealed that consumer WTP for a product will decrease with the shortening of its shelf life. The research results of Nandi et al. (2016) [31] also pointed out that Indian consumers are willing to pay a higher WTP for fruits and vegetables grown in an ecofriendly way. Previous studies on consumer WTP for suboptimal food show that price discounts will attract purchase decisions [10,25,32]. Since there is presently no definite range for price discounts for suboptimal food in Taiwan, this study includes price discounts as an attribute variable to explore the prices that Taiwanese consumers are willing to pay for suboptimal food.

The contingent valuation method (CVM) is often used to evaluate consumer WTP for non-market goods. The CVM asks respondents about their WTP for a certain good in a hypothetical market through a questionnaire survey [33,34]; however, because the CVM has some limitations in terms of application, a strategic bias may be caused because of overestimation or underestimation. For example, respondents deliberately conceal their real preference for non-market goods for their own interests [35]. Therefore, the choice experiment method (CEM) has gradually become an important evaluation tool for measuring the value of non-market goods [36]. The CEM was first proposed by Louviere and Hensher (1982) [37] and Louviere and Woodworth (1983) [38]. Its theoretical framework is derived from random utility theory [39]. CEM has been widely used in non-market value evaluation in recent years. For example, Tait et al. (2016) [40] used a CEM approach to explore the WTP for mutton products with an environmental label certification for consumers in Britain, China, and India. Scarborough et al. (2015) [41] used a CEM to discuss British consumer preferences for products with traffic light labels in supermarkets. Ortega et al. (2015) [42] used a CEM to study Chinese retailer preferences for food quality and safety attributes. Grebitus et al. (2015) [43] used a CEM to explore the roles of human values and generalized trust on stated preferences when food was labeled with environmental footprints. Meyerding et al. (2019) [44] used a CEM to explore German consumer preferences for product attributes (e.g., place of origin and production method) and WTP for local fresh tomatoes and their processed products (e.g., ketchup). Thøgersen et al. (2019) [45] used a CEM to explore German, French, Danish, Chinese, and Thai customer preferences for product attributes (e.g., the country of origin, organic badge, and price) for organic foods produced in the corresponding countries.

In summary, previous studies on suboptimal food have mainly focused on the appearance, date, and packaging, as well as consumer preferences for different types of suboptimal food. Taiwan is known by many as the "kingdom of fruit." According to the Yearly Report of Taiwan's Agriculture 2018 Supply and Demand for Food by the Council of Agriculture of the Executive Yuan, citrus represents the largest fruit production category in Taiwan (524,087 metric tons). Therefore, citrus is used as the main product for investigation in this study. In addition, the appearance of citrus fruit is limited by fungal decay [46] and peel defects [47] during store retailing, affecting the willingness to pay. The contribution of this study lies in dividing the attributes of suboptimal food in terms of the appearance, size, freshness indicator, traceable agricultural products, and price discount and using a CEM to deduce the overall preferences of consumers and the consumer WTP for various

product attributes for suboptimal food. This is achieved via conditional logit (CL) and random parameter logit model (RPL) analyses. It is expected that the research results can be used as a reference for retailers as well as other sales channels and enhance the public's awareness of suboptimal food. When consumers purchase citrus, the appearance, size, freshness indicator, and price discount are the primary factors which impact the purchase intention. On the other hand, a traceable agricultural seal is generally used for agricultural products in Taiwan, which is why this attribute is discussed in terms of suboptimal fruit here.

#### **2. Materials and Methods**

*2.1. Survey Design*

This study explores the product attributes of suboptimal citrus fruit, including the appearance (complete and defective), size (large, medium, and small), freshness indicator (labeled and unlabeled), traceable agricultural products (with and without certification), and price discount. The variety of the citrus fruit is *Citrus poonensis,* which is the most common in Taiwan. The details are shown in Table 1.


**Table 1.** Attributes and levels of suboptimal citrus fruit. NT: New Taiwan dollars.

To develop an easy method for respondents to fill out questionnaires, this study adopted an orthogonal design which was submitted through Statistical Product and Service Solutions. Ninety-six (23 × 31 × 41) sets were selected, and after factoring out the redundancies, each choice set contained two random number substitutes and one status quo which included flawless appearance, medium size, no freshness indicator, uncertified traceability, and a discount price of 40 New Taiwan (NT) dollars. Each survey included three choice sets extracted from among them, with a total of 15 possible questionnaire versions.

In this study, we used judgmental sampling to survey questionnaire answers face-toface by paper and pencil in supermarkets, mass merchandisers and traditional market. First, the study conducted a pre-test questionnaire, with the aim of understanding consumers' overall consumption preferences and WTP for suboptimal food. The questionnaires were issued from 1 March 2020 to 31 March 2020 to consumers who had purchased suboptimal food in the past six months. During the first stage, 150 questionnaires were issued, out of which 121 were valid, and the effective questionnaire recovery rate was 80.67%. The official survey was divided into three parts. Part one included the degree of importance that consumers place on the suboptimality of five cases of citrus fruit, asking consumers

to rank the issue in importance from 1 to 5 according to their own personal beliefs. Part two presumes that the consumers are going to purchase citrus products with various attributes (i.e., overall superior fruit) and provides plans for citrus products based on their appearance, size, freshness indicator, and food traceability seal, then assigning a discount to be displayed to help consumers choose the one they like best based on their own personal preferences toward the attributes of suboptimal food (as shown in Figure 1).


**Figure 1.** Example questionnaire choice set.

Part three is the demographic information of the survey respondents, which includes their gender, age, education, marital status, monthly income, and whether they would consider buying products with a traceability seal or scanning QR codes that can allow pertinent information about the product to be read.

#### *2.2. Choice Experiment Method*

A CEM approach is used to establish a hypothetical market to investigate consumer preferences for non-market goods. As the CEM has the ability to evaluate multiple attributes and levels, different alternatives are combined for the important characteristics

related to non-market goods or services. Through the choice sets with different situational assumptions, respondents could select appropriate alternatives according to their preferences to avoid errors in the evaluation. In regard to the empirical model, the conditional logit (CL) regression model is used to estimate consumer average preferences for multiple attributes for fresh food, as well as their marginal willingness to pay (MWTP) for the attribute levels [48]. Secondly, the random parameter logit model (RPL) can be used to explore the preference heterogeneity and the WTP for different characteristic attributes for respondents with different socioeconomic backgrounds [49]. Bazzani et al. (2017) [50] used the CEM to explore Italian consumer preferences for various product attributes (e.g., product sources and production methods) and WTP for local and organic foods. Kallas et al. (2019) [51] discussed Spanish consumer purchase intentions and WTP for innovative patties containing black pork products enriched with porcini mushrooms as a natural source of dietary fiber or blueberries as a natural antioxidant source. Ceschi et al. (2017) [52] investigated Italian consumer preferences for product attributes for apples, for example, being organic, their color (bicolored, green, and red), origin, and import country, also evaluating their WTP.

First, this study adopted the CEM to construct a preference and utility model for suboptimal citrus fruit, then applying the CL and RPL to estimate the utility function for preferences, then finally exploring the MWTP attributes in terms of the demographic information of the respondents. In the binomial model below, "j" is the utility function that was arrived at for the hypothetical respondents "i" through a product substitution, as in Formula (1).

$$\mathbf{U\_{i\bar{j}} = V\_{i\bar{j}} + \varepsilon\_{i\bar{j}}} \tag{1}$$

where Uij represents the attribute of the *i*-th respondent facing the *j*-th option, Vij represents the observable part of the utility function, and εij represents the residual item, i.e., the unobservable random utility.

The hypothesis assumes Vij to be the linear form of the substitute's observable attributes Xij; thus, it is possible to take the consumer utility i contained in the k item attributes of substitute j and assume the price variable of that set of attributes to be P<sup>j</sup> , therefore the consumer utility function i can be expressed as Formula (2):

$$\mathbf{U}\_{\mathbf{i}\mathbf{j}} = \mathbf{V}\_{\mathbf{i}\mathbf{j}} + \boldsymbol{\varepsilon}\_{\mathbf{i}\mathbf{j}} = \sum\_{\mathbf{k}=1}^{\mathbf{k}} \alpha\_{\mathbf{k}} \mathbf{X}\_{\mathbf{j}\mathbf{k}} + \beta \mathbf{P}\_{\mathbf{j}} + \boldsymbol{\varepsilon}\_{\mathbf{i}\mathbf{j}} \tag{2}$$

where Uij is the utility that consumer i derives from product j; Xjk is the *k*-th attribute of the product j, and P<sup>j</sup> is the price of product j. In additional, α<sup>k</sup> and β are the parameters to be estimated.

In order to explore the reasons behind respondent preferences for certain parameters for suboptimal citrus fruit, Formula (2) was expanded into Formula (3), which adds in the random utility function of the respondent demographic information:

$$\mathbf{U\_{i\bar{j}}} = \sum\_{\mathbf{k}=1}^{\mathbf{k}} \alpha\_{\mathbf{k}} \mathbf{X\_{\bar{j}k}} + \sum\_{\mathbf{k}}^{\mathbf{K}} \sum\_{\mathbf{q}}^{\mathbf{Q}} \gamma\_{\mathbf{k}\mathbf{q}} \mathbf{X\_{\bar{j}k}} \mathbf{Z\_{\bar{i}\mathbf{q}}} \mathbf{\beta P\_{\bar{j}}} + \varepsilon\_{\mathbf{i}\bar{j}} \tag{3}$$

In the formula, Uij is the utility that the *i*-th respondent derives from the product j and Xjk is the k-th attribute of the product j. Ziq is the qth demographic information of the respondent i. α<sup>k</sup> is the attribute variable coefficient. γkq is the overlap coefficient of the attribute variable and demographic information.

By separating the price variables in Formula (3), it can be seen clearly and the price that consumers are willing to pay may be more easily analyzed by Formula (4):

$$\mathbf{U}\_{\rm ij} = \mathbf{V}\_{\rm ij}(\mathbf{X}\_{\rm ij}\mathbf{S}\_{\rm i}) + \varepsilon\_{\rm ij} \tag{4}$$

where Vij is the utility coefficient of observable variable Xij and respondent characteristic Si , which represents the respondent's preference, and εij is the residual item.

To measure the WTP for the product attributes, we took the total differential of Formula (2), treated the utility as a constant, and assumed that dUij = 0, which gave Formula (5):

$$\mathbf{d}\mathbf{U}\_{\mathbf{ij}} = \sum\_{\mathbf{k}=1}^{K} \alpha\_{\mathbf{k}} \mathbf{d}\mathbf{X}\_{\mathbf{jk}} + \beta \mathbf{d}\mathbf{P}\_{\mathbf{j}} = \mathbf{0} \tag{5}$$

When other attribute variables remain constant (dXj1 = dXj2 = · · · = dXjk−<sup>1</sup> = 0), finding the consumer WTP for Xjk attribute(s) of product j can be carried out as per Formula (6):

$$\mathbf{d}\mathbf{U}\_{\mathbf{i}\mathbf{j}} = \sum\_{\mathbf{k}=1}^{\mathbf{K}} \alpha\_{\mathbf{k}} \mathbf{d}\mathbf{X}\_{\mathbf{j}\mathbf{k}} + \beta \mathbf{d}\mathbf{P}\_{\mathbf{j}} = \mathbf{0} \tag{6}$$

#### **3. Results and Discussion**

#### *3.1. Sample Size and Composition*

To understand the attributes of the respondent consumption preferences for suboptimal citrus fruit, this study utilized judgmental sampling to target consumers who have purchased suboptimal food. A total of 670 questionnaires were issued. After factoring out invalid questionnaires, a total of 485 valid questionnaires were obtained, representing a 72.4% questionnaire recovery rate. The screening rule for invalid questionnaires was when respondents did not ever purchase suboptimal food. In such cases, the questionnaires were classified as invalid questionnaires. The largest respondent proportion was males (51.1%). Age was primarily concentrated in the 41–50 age range (27.8%), followed by 51–60 years of age (25.8%), and 31–40 years of age (22.7%), showing that the middle-aged demographic is more likely to buy suboptimal food than other consumer groups. In terms of education level, the proportion with a tertiary education was the highest (48.5%). In terms of marital status, the majority of respondents were married (57.5%). The average monthly incomes of individuals were primarily in the NT\$ 40,001–60,000 range (35.1%), followed by the NT\$ 20,001–40,000 range (28.0%). More than half (57.7%) of consumers had purchased traceability-certified products, but most did not scan the QR code to read the pertinent information. The respondent purchasing locations for suboptimal food (detailed below) were mainly supermarkets (73.8%), followed by mass merchandisers (47.4%), which are companies that affordably sell large quantities of goods that appeal to a wide variety of consumers, and traditional markets (42.1%). The most purchased foods (detailed below) were vegetables (57.7%), followed by fruits (53.0%), and whole grains and tubers (51.1%).

#### *3.2. Emphasis on Attributes*

The results of the research show that the degree of emphasis placed on the various attributes was highest for freshness indicators (3.86 points), followed by appearance (3.22 points), traceability certification (2.66 points), price discounts (2.64 points), and size (2.62 points). The results of the weight comparison analysis show that consumers place greater emphasis on the freshness indicators and appearance of suboptimal citrus fruits but do not focus on the size, which is different from the results of the study by de Hooge et al. (2017) [26], which found that consumers in Germany, the Netherlands, and Norway place great emphasis on price discounts for suboptimal food. It is presumed that the aforementioned countries rely on imports due to climatic and environmental factors that affect the types and quantities of fruits and vegetables that may be locally grown, therefore emphasizing pricing changes.

#### *3.3. Preferred Suboptimal Products by Consumers*

This study analyzed the 11 most preferred attribute sets included in the suboptimal citrus fruits as selected by the respondents. The results show that the most preferred set of attributes includes a flawed appearance and moderate size with a freshness indicator,

traceability certification, and a discounted price of NT\$ 25 (accounting for 11.48% of respondents). The second most common set was the set with a flawless appearance, large size, freshness indicator, no traceability certification, and a discounted price of NT\$ 25 (accounting for 10.65%). The least preferred attribute set by consumers was a flawed appearance with a discounted price of NT\$ 25 (2.27%) and a flawed appearance with a discounted price of NT\$ 35 (0.96%). It is presumed that a possible reason for this is that consumers themselves are not willing to buy citrus fruits with a flawed appearance while paying little attention to discounts and will only buy them in the hopes that, flawed appearance notwithstanding, there are other certifications that can guarantee the product.

#### *3.4. Conditional Logit Analysis Results*

Based on random utility function given by Formula (1), a utility model for suboptimal citrus fruits was established to understand consumer preferences for suboptimal citrus fruits, as in Formula (7):

$$\mathbf{U}\_{\overline{\mathbf{i}}\overline{\mathbf{j}}} = a\mathbf{1}\mathbf{E}\mathbf{D}\overline{\mathbf{i}}\mathbf{j} + a\mathbf{2}\mathbf{S}\mathbf{Z}\mathbf{1}\mathbf{i}\mathbf{j} + a\mathbf{3}\mathbf{S}\mathbf{Z}\mathbf{2}\mathbf{i}\mathbf{j} + a\mathbf{4}\mathbf{F}\mathbf{R}\mathbf{i}\mathbf{j} + a\mathbf{5}\mathbf{T}\mathbf{A}\mathbf{P}\mathbf{i}\mathbf{j} + \mathbf{3}\mathbf{F}\mathbf{U}\mathbf{N}\mathbf{D}\mathbf{i}\mathbf{j} + \varepsilon\mathbf{i}\mathbf{j} \tag{7}$$

In the formula, i = 1, 2, 3, . . . , 485, which is the total sampling of 485, and j = 1, 2, 3, . . . , 12, which represents the 12 choice sets for the suboptimal fruit attributes.

A coefficient for the attribute variable was estimated for Formula (7) through NLOGIC 4.0 Conditional Logit, and then the coefficient value was substituted into Formula (7) to find the discounted prices that were willing to be paid for each attribute. The empirical estimation results are summarized in Table 2.

**Table 2.** Conditional logit empirical estimation results. WTP: Willingness to pay.


\*\*\*, \*\*, and, \* are significant at 1%, 5%, and 10%, respectively.

In terms of the levels for suboptimal citrus fruit attributes, appearance, freshness indicators, and traceability certifications were all at the 1% significance level, while only a small size (SZ2) was at a 10% significance level, meaning that willingness to consume is affected by whether the appearance is flawed, size is too small, and whether the product has freshness indicators and traceability certifications. Maintaining the status quo (ASC) was both positive and significant at a 5% significance level, indicating that consumers would prefer to maintain the status quo.

Second, a coefficient was estimated through the utility function from Formula (1) and was substituted into the theoretical model (Formula (6)) to calculate the WTP. The prices that were willing to be paid for the attributes were in the following order: appearance (NT\$18), small size (NT\$36), freshness indicator (NT\$68), and traceability certification (NT\$63). The results of the analysis revealed that, of the four suboptimal citrus attribute levels, the price discount for appearance is the highest, which means if sellers want consumers to buy suboptimal food, the price needs to be reduced from the original NT\$40 to NT\$18. Additionally, based on the results of the study, consumers are willing to pay more for citrus fruits with freshness indicators and traceability certifications, which indicates that consumers prefer these two product attributes. Furthermore, because consumers care about the freshness indicators the most, no matter how much the discounted price is, they will not

purchase fruit without a freshness indicator. This is because there are no other guarantees such as freshness indicators or traceability certifications on the suboptimal products, and consumers may believe that these fruits could endanger their health. According to a study from de Hooge et al. (2018) [53], fruits, vegetables, and foods with dented packaging should not be regarded as inferior products. Although they are visibly different from the best products, visual flaws are considered a sign of authenticity. Tsalis (2020) [54] believes that retailers in most countries and regions only sell suboptimal food as a cheap product that does not legitimately generate purchase intentions. Wang et al. (2018) [55] pointed out that product certification labels can eliminate the uncertainty that consumers face when buying products. Thus, WTP for products with certification labels will increase accordingly.

#### *3.5. Random Parameter Logit Analysis Results*

Since the CL assumes that the parameters in the respondents are fixed, the average preferences of the respondents were evaluated, while the RPL is based on the attribute parameters of the respondents taking the form of a normal distribution, where the differences in preferences for the suboptimal citrus attributes can be evaluated. The results of using the CL and RPL to evaluate suboptimal citrus attributes were quite dissimilar. The RPL presented respondent preferences for appearance, larger or smaller sizes, freshness indicators, and traceability certifications, while the CL presented respondent preferences for appearance, small size, freshness indicators, and traceability certifications, except for large sizes. Additionally, the RPL also reflects the heterogeneous distribution of respondent preferences for various attribute parameters. As shown in Table 3, keeping the status quo was found at a significance level of 1%, and appearance and freshness indicators are both significant and indicate heterogeneity in respondent preferences between appearance and freshness indicators. This means that consumers cared more about appearance and freshness indicators for suboptimal products than other attributes.


**Table 3.** Conditional logit (CL) and random parameter logit model (RPL) empirical estimation results.

\*\*\*, \*\*, and \* are significant at 1%, 5%, and 10%, respectively.

The coefficient value estimated through the utility function (Formula (1)) was substituted into the theoretical model (Formula (6)) to calculate the respondent WTP. The prices for each attribute were in the following ascending order: appearance (NT\$19), larger size (NT\$36), smaller size (NT\$33), freshness indicators (NT\$64), and traceability certifications (NT\$59). The analysis results show that, of the five attribute levels for suboptimal citrus fruit, a price discount for appearance ranks the highest, which means that consumers are willing to purchase citrus fruit with a flawed appearance, although the price needs to be reduced from the original price of NT\$40 to NT\$19. The freshness indicator and traceability certification results show that consumers are willing to pay more for citrus fruit with these two product attributes. Freshness indicators represent the highest price increase, changing from the original price of NT\$40 to NT\$64. Jaeger et al. (2018) [56] pointed out

that the product appearance, aroma, expiration date, and overall sensory evaluation are the purchase intention determinants. Hingston and Noseworthy (2020) [57] pointed out that consumer aversion to agricultural products with an abnormal appearance depends on their personal experience with these foods. The factors which impact consumer purchase intentions for food are inferences about taste, texture, and safety. Van Boxstael et al. (2014) [58] pointed out that most consumers have different opinions on shelf life labels and expiration dates for different food types.

*3.6. Exploration of Respondent Demographic Information in Suboptimal Food WTP Heterogeneity*

The results of the RPL analysis show that there were random parameters for the appearance and freshness indicators. Therefore, this study compared WTP with the respondent demographic information based on the two aforementioned attributes. The results of this analysis are shown in Table 4.


**Table 4.** Heterogeneity of respondent demographic information in terms of WTP for suboptimal food.

\*\*and \* are significant at 5% and 10%, respectively; ASC: keep the status quo; ED: appearance; FR: freshness indicator. TAP: traceability.

There was a significant difference in terms of the average monthly income of individuals who were willing to pay for appearance. Of them, respondents with an average monthly income between NT\$40,001 and NT\$60,000 were willing to pay a lower price, indicating that those with an average monthly income of the middle class are less willing

to buy products with a flawed appearance. Of these respondents, those aged between 21–30 years with a tertiary education or above and those whose monthly income was between NT\$20,001 and NT\$40,000 were willing to pay a higher price. This shows that young people and those with higher education and a monthly income from NT\$20,001 to NT\$40,000 more greatly emphasize product freshness and are therefore willing to pay more for products that display freshness indicators. This result is consistent with the study by Tsakiridou et al. (2011) [59] that identified the consumers who are willing to pay higher prices for fruits with food safety labels.

#### **4. Conclusions**

#### *4.1. Concluding Remarks*

The study results show that, of the suboptimal citrus fruit certification attributes, the most important is the freshness indicator, followed by appearance, traceability certifications, price discounts, and finally size. Based on the results found here, the suboptimal citrus attribute set most preferred by respondents was the following: appearance flaws, moderate size, freshness indicators, traceability certifications, and a discounted price of NT\$25. The second-most preferred attribute set was the following: perfect appearance, larger size, freshness indicators, without traceability certifications, and a discounted price of NT\$25. The least preferred certification plan was the one with only appearance flaws and a discounted price of NT\$35, along with the one with only appearance flaws and a discounted price of NT\$25. This was presumably due to consumers not being willing to buy suboptimal citrus fruits with a flawed appearance and paying less attention to price discounts, although, a flawed appearance notwithstanding, there are other certifications that can provide a quality guarantee for the products. Thus, retailers should upgrade their food preservation systems to keep products fresh. In terms of primary producers, they can used suboptimal foods in food processing such as the production of fruit jams, canned vegetables, candied fruits, etc.

#### *4.2. Recommendations*

#### 4.2.1. Managerial Implications

This study has analyzed the importance of each fruit attribute based on respondent preferences and has found that consumers are not overly focused on size. The reason for this is presumed to be due to the fact that citrus products purchased by consumers in the market are classified through a screening mechanism before being circulated and sold in the market. The screened products are mainly medium-sized and above (25−30 cm), with smaller sizes being rejected and removed before reaching consumers. Therefore, it is recommended that relevant government agencies provide publicity and explanations for promoting suboptimal food being processed to change its form, such as making it into juice or canned food, thus greatly increasing its value. Governments should relax any regulations on the minimum sizes of fruits as consumers will still buy smaller fruits. Mass media promotion can promote consumers to buy and eat fruit, which can not only reduce losses for growers, but also reduce food waste. Besides emphasizing freshness indicators, appearance is another important attribute. The reason for this is presumably due to the inability of consumers to accurately determine the quality of fruit. Therefore, the appearance attribute is the second priority. If a product has a freshness indicator, this represents a guarantee for both the product and the consumer.

The results of the empirical analysis here show that consumers who prefer suboptimal citrus fruit with a freshness indicator and traceability certification are willing to pay more for the purchase. Therefore, it is recommended that the government not only stipulates that packaged foods need to show an expiration date, but also advocates for the popularization of freshness indicators for bulk foods or the addition of packaging. For example, using the Kanban software to indicate harvest dates and making the label certification process more transparent, i.e., label certifications and date indications can be displayed on product packaging. Meanwhile, more food-related knowledge should be spread to

enhance public awareness. Under the assumption that a product is guaranteed, consumers can make discerning purchases that are not just for gifting or personal use based on their product recognition. Products with appearance flaws have the highest price discounts. Currently, there is no clear range of discounts in the Taiwanese market. Fresh ingredients in supermarkets and hypermarkets on site will be sold at discounts (20−40% off) based on their expiration dates. It is recommended to plan to establish a discount system based on the characteristics of the food category or the expiration date and provide references for retailers or other sales channels.

#### 4.2.2. Research Limitations and Future Research Direction

There were a number of limitations in the research process here. If it is possible to expand the scope of future research, then the research framework may be perfected. This study makes the following recommendations in connection with the conclusions and limitations of this research. Only five attributes for suboptimal food (appearance, size, freshness indicator, traceability certificates, and price discounts) were set up for this study, but there are more suboptimal food-related attributes that can be added. For example, the reuse value of suboptimal citrus fruits, damaged packaging when there is packaging, etc., can be used to better understand consumer willingness in relation to price increases or discounts and preferences for different products and attributes; however, expiry dates might make foods more appealing to consumers, but they also represent a restriction for sellers, since they have to waste more unsold food, meaning that this problem has also not been solved here.

This study only explored the consumer dimension, and the results only reflect the current consumer preferences and WTP for suboptimal food. Follow-up research can be aimed at exploring the seller dimension and understanding the opinions of different respondents toward the various attributes of suboptimal food and comparing their differences.

Future research can use a latent class model to test whether there is heterogeneity in respondent preferences for suboptimal foods.

Additionally, this kind of research can apply to the subject of waste to meat and animal products, as these foods are generally the most resource-intensive foods to produce. Therefore, reducing the wastage of these foods represents significant benefits.

**Author Contributions:** The four coauthors contributed together to the completion of this article. W.-S.H., the first author, drafted the manuscript; H.-Y.K. analyzed the data; S.-Y.T. contributed to the manuscript and literature review; H.-S.C. acted as corresponding author throughout the revision and submission process. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable for studies not involving humans or animals.

**Informed Consent Statement:** Not applicable for studies not involving humans.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the privacy and ethical.

**Acknowledgments:** We would like to express our heartfelt thanks to all the experts who have taken the time to review this article and provide valuable comments.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

