**1. Introduction**

The olive tree is a central plant in the history of civilizations of the Mediterranean basin. It has historically been considered a sacred tree: in Genesis, the olive branch is a symbol of peace. Even in Greek Mythology, the olive branch was considered a symbol of peace and life. Ancient Romans rewarded valiant citizens with crowns of olive branches. In many cultures in the Mediterranean area, the olive branch is a symbol of justice. In addition, in the traditional culture of many Italian regions, olive leaves, as well as olive oil, have played the role of good luck against the evil eye [1]. Beyond the symbolic properties attributed to olive leaves, they have historically had a pharmacological role in Mediterranean countries and are widely used in traditional herbal medicine to prevent and treat various diseases [2,3].

This historical and anthropological excursus outlines the notion that olive leaves have almost never had a nutritional function in the tradition of Mediterranean areas, despite the importance of olive oil consumption [4,5]. The emergence of modern medicine and culture, however, has led to a loss of interest in olive leaves, and they have gradually become a processing waste for olive oil producers, with high disposal costs and no economic value.

Nevertheless, recent scientific literature has shown that olive leaves are rich in nutrients that could be used as by-products by the food industry to enrich food products with functional properties [6,7]. An essential point to understand the real market uptake of food enriched with food by-products that are not part of the traditional diet is to estimate consumer acceptance for such products, especially for food enriched with olive leaves.

Consumers consider food a complex good that contains both quality attributes as well as nutritional ones. Consumer acceptance is shaped by different factors, such as food habits and sociodemographic characteristics [8,9], food origin [5,10,11], the information that consumers access, and their trust in food production [12,13], the role of transparency [14], the range and prices of existing products [15–19], and the perception of health benefits and safety of food products [20,21].

In this context, the new objective of food production is no longer just food security but also the satisfaction of consumer needs and preferences with new products with both functional properties [22] and quality aspects (e.g., local, organic, etc.) [18,23]. In this context, the production of functional and nutritious food, obtained from upcycled ingredients, is a recent development [6,24–27].

Several studies referring to the use of unusual products in food production (i.e., insect products, cultured meat, etc.) report how new products can cause a strong food technophobia and neophobia in consumers [28–32]. Furthermore, studies on consumer acceptance of food derived from upcycled ingredients have found a negative impact of food technophobia and neophobia on the likelihood of acceptance [25,33]. However, interestingly, studies found that consumers may accord these products a premium status if promoted as a new food category akin to organic foods [34]. In fact, according to Bhatt et al. [34], consumers perceive "value-added surplus products" as having benefits for society and the individual.

The environmental sustainability of food production has become a matter of growing interest for consumers worldwide in recent years [35–37]. Often, environmental sustainability is understood by consumers as the preference to purchase organic food products, which are perceived as being healthier than conventional foods and better for the environment [18,38–42]. More recently, consumers have no longer been satisfied with the benefits offered by organic products only [43] but have been demanding domestic and local products as well [5,22,44–46]. In this respect, some studies suggest that consumers are interested in local food not only because they associate higher quality with these products but also for environmental friendliness, preference for their cultural roots [5,46], and support for the local economy [47] and local farmers [48].

Against this background, it seems to be of interest to analyze the possible impact of some frequent food consumption habits, such as buying organic products or giving importance to food origin in consumer acceptance of food derived from upcycled ingredients. In fact, foods obtained from upcycled ingredients of olive oil productions are new to the consumers, and their acceptance may be a problem despite their health or environmental benefits.

The main objective of the present study is to assess the association between the willingness to try food obtained from upcycled ingredients and consumer preferences for organic food and food origin. In particular, by estimating ordered probit models, we test if attributes, such as superior nutritional and/or environmental properties, of food with olive leaves are appealing to consumers. By doing so, we also investigate the possible market niches for these products.

In addition, to better understand consumer responses to food attributes, we run the analysis by age groups, as many studies have shown that elder and younger generations may actually differ in food preferences [49,50].

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

To investigate the association between the willingness to try food obtained from upcycled ingredients and the preference for organic and food origin, we surveyed a sample of 852 Italian consumers. The core of our analysis was an ordered probit regression [51] of a discrete variable measuring the willingness to try food with upcycled ingredients on a set of regressors, including demographic variables, measures of consumer environmental responsibility, technophobia, and concerns for product origin.

Data were collected through a web-based survey administered in Italy between April 2018 and April 2019 with a convenience approach. Participants were reached via different social media networks, which is becoming a more popular means of reaching participants in social sciences research for both convenience and inclusion reason [52]. In fact, the use of the Internet makes it as convenient as possible for participants to take part in the survey and allows reaching a high number of participants from all Italian regions. In particular, the information was then posted on Twitter, LinkedIn, and Facebook pages. As sampling in Internet research studies is not random and could generate selection bias, to minimize this possible problem, we posted the questionnaire on pages and online groups with a general target audience (e.g., web pages of Italian radio programs).

Before answering the questions, participants were briefly informed about the research project that motivated the survey.

Respondents were given a short, four-section questionnaire. Section 1 collected the demographic information, Section 2 assessed the respondent's attitudes toward the covariates of interest (organic food and food origin), Section 3 investigated the respondent's technophobia [29,30], and Section 4 asked the respondent's willingness to try food obtained from by-products [33]. Descriptive statistics of the sample and the questionnaire are presented in Table 1.


**Table 1.** Questionnaire and descriptive statistics of variables (*n* = 852).

To lessen collection cost and maximize the response rate, we minimized the number of items in the questionnaires and drafted questions as five-point Likert scales. The final design was extremely parsimonious and included 11 questions (see Table 1).

We measured the attitude toward organic food, asking respondents to state the frequency of their purchase of organic food on a 5-point scale (from never to always). To limit the possible self-representation bias, we referred to a specific action (buying organic food) instead of asking to report the attitude directly. This approach was possible because Italian consumers, on average, are familiar with organic products, given the sharp increase in organic consumption in the last decade (in 2017, 78% of Italian family had bought organic food at least once, and 48% had bought them at least once a week [53]).

Consumer perception of food origin was blurred mostly because of the overlapping of several different concepts, such as local food, typical food, or food safety. As a consequence, the questionnaire asked to report the importance of product origin in the food purchase decision.

Food technophobia (or food technology neophobia) is defined as consumers' fear, dislike, or avoidance of novel food technology [29]. Perito et al. [33] found that it is a key driver limiting consumer acceptance of food with olive by-products. As a consequence, we included it as a control variable in our empirical investigation.

Technophobia is a complex attitude to measure. Several contributions in the literature proposed scale measures [29,54,55]. In this paper, we adopt the approach proposed by Perito et al. [33]. The measure is based on three statements:

NNNT: There is no need for new food technologies because there are so many types of foods; NTOR: The benefits associated with innovative food technologies are often overestimated; NTLQ: New food technologies reduce the natural quality of foods.

Respondents were asked to agree or disagree with the statements on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree). Cronbach α of 0.78 confirmed that the construct items were consistent. We defined the technophobia index (TFI) as the average of the scores of the three variables.

Finally, the questionnaire asked respondents to report their willingness to try food with olive leaves as upcycled ingredients on a 5-point scale from 1 (No, Absolutely) to 5 (Yes, Absolutely). Because Perito et al. [33] identified environmental and nutritional concerns as key drivers of consumer acceptance, we conditioned the answer to two situations: A) the upcycled ingredients of olive leaves has superior nutritional properties (TNUT), and B) the upcycled ingredients reduces the environmental impact of food production and consumption (TENV). In this way, we provide useful insights to novel food researchers willing to market food with upcycled ingredients and different characteristics. Our hypothesis is that products with different attributes (nutritional or environmental) may be appealing to different consumers.

To investigate the association between consumer willingness to try food enriched with upcycled ingredients, organic food, and food origin, we ran a regression of the dependent variables TNUT and TENV on a vector of demographic variables (gender, age, education, and employment status), the technophobia index (TFI), and the attitudes regarding organic food (ORGANIC) and product origin (ORIGIN). Given the discrete nature of the dependent variables, we used an ordered probit model. In fact, in our case, the dependent variables were ordinal, but not continuous in the sense that the metric used to code the variables was substantively meaningful. For instance, the 5-point scale adopted to measure the dependent variable assigned the numerals to the categories but the metric underlying response identification was not necessarily the same as the linear metric relating the numerals. In other terms, the difference between 0 and 2 on the coded responses may be quite different from the difference between 2 and 4. A widely used approach to estimating models of this type is an ordered response model. The basic assumption of such models is that there is a latent continuous metric underlying the ordinal responses observed by the analyst.

The model estimated, assuming that the values of TNUT and TENV were the observable outcome of latent variables, is the following:

$$\hat{Y}\_h = f(\text{GEND, AGE, EDILATION, EMPLOYMENT, TFI, ORGANIC, ORIGIN}) \tag{1}$$

where h = N, E; GENB is a binary variable that is equal to 1 if the respondent is female; AGE is the respondent's age; EDUCATION is a categorical variable with four entries: elementary school, middle school, high school, and college, depending on the respondent's degree; EMPLOYMENT is a five-entry categorical variable equal to the worker (if the respondent is employed or self-employed), unemployed, student, homemaker or retired, depending on the respondent's status.

Regarding the variable AGE, a further refinement is then introduced as a separate regression for each age group has been conducted to exploit the large sample size of our survey.
