Examining the Role of Food Technology Neophobia in Shaping Consumer Attitudes and Intentions to Purchase Genetically Modified Foods

Abstract

In recent years, significant changes in food consumption habits have emerged due to various factors, including climate change, population growth, urbanization, and the depletion of natural resources. These changes pose a threat to the stability of global food systems and raise serious concerns about food security. Although this process encourages innovative and sustainable food consumption, it also makes individuals more skeptical and concerned about new foods. In this context, understanding consumer intentions regarding behaviors such as purchasing genetically modified (GM) foods is critical for predicting consumer responses and promoting responsible consumption patterns within the scope of sustainability. This study examined the effects of food technology neophobia and perceived information on attitudes and purchase intentions toward genetically modified (GM) foods. Survey data were collected from 324 participants across Turkey and analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM). The findings revealed that food technology neophobia reduces perceived benefits and increases perceived risks, whereas perceived information enhances perceived benefits and lowers perceived risks. Additionally, attitudes were found to influence the intention to purchase GM foods significantly. Global issues, such as climate change and the depletion of natural resources, highlight the importance of innovations in food technology for sustainable food production. Understanding consumer concerns and perceived knowledge levels regarding genetically modified (GM) foods is critical to ensuring that these products are accepted at the societal level in an informed and conscious way. This study contributes to the promotion of sustainable food technologies and responsible consumer behavior, in line with the objectives of Sustainable Development Goal 12 (Responsible Consumption and Production).

1. Introduction

The Sustainable Development Goals (SDGs) of the United Nations (UN) offer a globally recognized framework for addressing pressing social, economic, and environmental challenges. Among these goals, sustainable food production and the development of resilient food systems are particularly pertinent to Goals 2 (Zero Hunger), 12 (Responsible Consumption and Production), and 13 (Climate Action). Achieving sustainable patterns of production and consumption is the focus of SDG 12, which is crucial for enhancing resource efficiency, promoting sustainable lifestyles in all spheres of society, and decoupling economic growth from environmental degradation [1]. The concept of sustainability has evolved from a purely environmental approach to a multidimensional structure that also encompasses economic and social dimensions. However, achieving balance among these dimensions is a complex process. Approaches that are detached from reality and purely ideological fall short of achieving sustainable development goals. Therefore, a pragmatic approach should be adopted in sustainability practices; concrete data and applications must support theoretical frameworks. Effective and lasting solutions are only possible through action-oriented processes [2]. According to the Food and Agriculture Organization of the United Nations (FAO) [3], a key strategy aligned with these objectives is the advancement and implementation of sustainable food technologies, including genetically modified (GM) foods, which are promoted to enhance food security, reduce environmental impact, and support sustainable agricultural practices. One potential way to improve food security, reduce environmental impacts, and promote sustainable farming methods is through GM foods. However, consumer behavior and public acceptance play significant roles in the success of such inventions. New advancements in food technologies are considered key components in tackling worldwide challenges and shifting towards sustainable food production and consumption practices. Creating innovative and sustainable strategies in this context to address global challenges, such as reducing greenhouse gas emissions, minimizing food waste, and tackling food insecurity, is essential. High market failures characterize innovations in the food industry, as new food technologies can succeed in the market if accepted by consumers [4]. Consequently, these technologies are expected to play a role in establishing sustainable food systems that meet consumer needs and generate positive long-term environmental and economic outcomes. New food technologies can offer benefits, including enhanced food safety, improved nutritional content, and reduced environmental impact.

Conversely, new food technologies have driven innovation in the food sector. However, not all technologies are readily embraced and understood by consumers, and they often face resistance [5]. This is because there is a perception of the risks and uncertainty surrounding the long-term health and environmental impact of these technologies. There is a lack of trust in the regulatory oversight of these technologies, as well as cultural and ethical concerns. Therefore, it is necessary to investigate in greater depth the factors that influence the trust of all stakeholders in the food sector. GM foods have sparked widespread debate in recent years due to their potential for advancing sustainable food production and the potential health risks associated with their consumption [6]. Since their introduction, the positive and negative aspects of GM foods have been debated. The advantages of genetic modification technology include reducing costs through its application in food production, facilitating agricultural productivity, and increasing the vitamin content of food [7]. Therefore, despite the contribution of the GM food industry to environmental protection and sustainable development, many consumers remain cautious about GM foods and express concerns about the potential environmental and public health threats that these foods may pose.

With rapid urbanization and changing dietary habits, Turkey is facing increasing pressure on its food systems due to population growth and climate-related issues. Despite its strong agricultural potential, public confidence in GM foods is low due to limited awareness and strict regulations. This study is one of the few empirical investigations that simultaneously examines the effects of food technology neophobia and perceived knowledge on attitudes toward and purchase intentions for genetically modified (GM) foods in the context of Turkey, an emerging economy. While existing literature has examined chiefly these variables independently, this study fills an important gap by analyzing the aforementioned psychological determinants within a theoretically grounded and comprehensive model. One of the most original contributions of the research is that the model it employs is directly linked to the Sustainable Development Goals (SDGs). In particular, within the scope of SDG 12 (Responsible Consumption and Production), it reveals how consumers’ levels of knowledge and fear of innovation shape their approach to sustainable food technologies, especially those related to GM foods. In this respect, the study contributes to consumer education, risk communication, and policy development strategies that will facilitate the adoption of sustainable agricultural practices. In line with the research’s methodological structure, the relevant theoretical framework was first systematically presented, and research hypotheses were developed. Subsequently, the data collected through the survey method were comprehensively analyzed, and the research hypotheses developed were tested using the Partial Least Squares Structural Equation Modelling (PLS-SEM) method. The empirical findings obtained were evaluated in the context of theoretical and practical implications.

2. Literature Review and Research Hypotheses

2.1. Food Technology Neophobia

The ever-growing demand for food and the unavoidable impact of climate change necessitate increased flexibility and innovation in crop resilience and production systems [8]. According to the European Commission’s definition [9], novel foods are edible foods that were not traditionally consumed within the European Union before 1997 and are newly developed, innovative, and produced using new technologies or production processes. New food technologies can be crucial in addressing global challenges and transitioning to sustainable food production and consumption patterns. However, a new food technology can be successful if consumers accept it. Although new food technologies and novel foods have the potential to enhance food quality, availability, safety, and sustainability, consumers may reject them for reasons such as food technology neophobia [10]. The food industry’s innovations often require greater acceptance from the market, partly because of neophobia, in which specific individuals reject new or unfamiliar foods. Considering that resources are limited worldwide, it is widely acknowledged that individual consumption behavior is unsustainable. The global food system accounts for approximately 30% of annual greenhouse gas emissions, making it necessary to take measures for sustainable food consumption [11]. As competition for food sustainability grows, producers must innovate and create new foods that gain market acceptance. Food neophobia refers to the reluctance to consume new or unfamiliar food. Food technology neophobia refers to the fear of unfamiliar or new technologies in GM food. Neophobic individuals tend to have unfavorable perceptions and lower expectations regarding food tastes [5]. Some consumers prefer “natural”, i.e., less processed, products because they are concerned about the risks of new food technologies [12]. Consumers characterized by high food neophobia tend to resist trying or consuming “unfamiliar” foods, including GM foods. Hence, it is hypothesized that consumers’ food neophobia concerning the utilization of gene technology in food production influences their perceptions of the benefits and risks associated with GM foods [13]. The research model is presented in Figure 1 below, based on the hypotheses derived from the findings in the literature.

The literature on consumer acceptance of food focuses on the role of food technology neophobia in consumer decision-making. Researchers such as Evans et al. [14] and Tuorila and Hartmann [10] have linked food technology neophobia to mistrust in science, concerns about sustainability, and a lack of knowledge about food production. Martin et al. [12] showed that food technology neophobia is a crucial factor in determining consumers’ perceptions of risk and benefit, and it strongly influences consumer acceptance in decision-making processes related to food technologies. In this context, the following hypotheses were formulated:

H1: 

Consumers’ food technology neophobia negatively affects their perceived benefits.

H2: 

Consumers’ food technology neophobia has a positive effect on their perceived risk.

2.2. Perceived Knowledge

When consumers are highly interested in a specific product category, their knowledge of that category tends to increase. Consumer knowledge often leads to a greater likelihood of seeking new information as part of the decision-making process within the product category, further enhancing consumer knowledge [15]. Theoretically, consumers’ perceived knowledge includes familiarity and product knowledge. Familiarity represents accumulated consumer experiences that can be positive and may lead to strong beliefs about GM foods. Conversely, negative experiences can result in the rejection of GM foods.

Most people lack detailed information about genetic technology. This lack of information can increase an individual’s risk perception and reduce their level of acceptance. However, the situation is not as straightforward as more information leads to favorable attitudes. As knowledge levels increase, consumers tend to ask more critical questions about GM foods, reinforcing existing and possibly negative attitudes. Second, consumers are more likely to act on their existing attitudes rather than change them with the increased level of information provided [13]. Consumers’ understanding of GM technology is also linked to their perception of the advantages and risks associated with GM food. It is essential in correcting biased perceptions and intentions regarding GM foods [16,17]. Consumers who perceive themselves as more knowledgeable are more likely to recognize the scientific basis and potential advantages of GM foods, leading to a lower risk perception and greater acceptance. We measured perceived knowledge using a validated multi-item scale that captures the extent to which individuals believe they are informed about GM technologies. In this context, it is assumed that consumers’ perceived level of biotechnology knowledge affects GM food’s perceived benefits and risks. Accordingly, the following hypotheses are proposed:

H3: 

Consumers’ level of knowledge about GM foods positively affects their perceived benefits of GM foods.

H4: 

Consumers’ knowledge of GM foods negatively affects their perceived risk of GM foods.

2.3. Perceived Benefit and Risk

Perceived benefits refer to the positive outcomes of consumer behavior in response to an actual or perceived threat [18]. Trust in an organization or an individual critically influences perceived benefits [8]. While the industry requires innovative technologies for food development, substantial evidence suggests that many consumers are apprehensive about new foods and technologies due to perceived risks and a lack of perceived health benefits [17,19]. According to Ali et al. [8], organizational credibility, honesty, and benevolence play crucial roles in shaping consumer perceptions of the acceptance of GM foods. It is theorized that perceived risks negatively impact attitudes, whereas perceived benefits have a positive effect [13]. If GM foods provide substantial benefits, these benefits outweigh the perceived risks, leading to a positive attitude towards GM foods. Otherwise, consumer acceptance may remain very low if perceived risks outweigh the benefits [20]. When information about GM food technology is insufficient, consumers cannot objectively evaluate potential risks and benefits. In this context, the following hypotheses were developed:

H5: 

Consumers’ perceived benefits of GM foods positively affect their attitudes towards GM foods.

H6: 

Consumers’ perceived risks of GM foods negatively affect their attitudes towards GM foods.

2.4. Attitude Towards GM Foods

Consumer attitudes towards GM foods affect their overall skepticism. According to Chen and Li’s [21] theoretical framework, attitudes towards GM foods are contingent on perceived risks and benefits. As people become more informed about GM foods, they perceive that the benefits outweigh the risks. However, consumers often need to be aware of whether they consume GM foods. Guo et al. [22] reported that, as people’s perceived potential risks increase, they are less willing to purchase GM products. We propose that positive attitudes about genetically modified foods, influenced by perceived risks and advantages, are potent predictors of purchase intentions based on the Theory of Planned Behavior and relevant consumer behavior research. Therefore, this study utilizes these attitudinal mechanisms to investigate how upstream factors, such as knowledge and neophobia, indirectly influence consumer decisions. Thus, we propose the following hypothesis:

H7: 

Attitudes towards GM foods positively affect the intention to consume GM foods.

The research model, formed in line with the research hypotheses, is presented below:

3. Materials and Methods

A questionnaire was used to collect data. The first part of the questionnaire consisted of scales measuring the dependent and independent variables of the research. The second part included questions on the demographic characteristics of the participants. The study scales were obtained from those whose validity and reliability were tested by reviewing relevant literature. The 9-item scale used by Ali et al. [8] was utilized to measure fear of new food technology. The perceived risk scale consists of three statements, and the perceived benefit scale consists of four. The scales used by Rabbani et al. [23] measure perceived risk and perceived benefit. The Perceived Knowledge Scale, consisting of four statements, was adopted by Bredahl et al. [20]. The intention to consume GM food scale consisted of three statements adapted from Kim et al. [24]. Chen [13] used three statements to measure attitudes toward GM foods.

The questionnaire was initially prepared in English and translated into Turkish through a double back-translation process to ensure linguistic and conceptual equivalence. Two independent bilingual researchers translated the items from English to Turkish, and a different pair retranslated the Turkish version back into English. Discrepancies were discussed and resolved collaboratively. Following translation, a pilot test with 30 respondents from the target population was conducted to ensure clarity and reliability. Minor wording adjustments were made based on the pilot feedback. In this study, a 5-point Likert scale was used for all survey items. This choice balances simplicity and sensitivity, providing a manageable format for respondents while capturing sufficient variability in attitudes and perceptions. It also aligns with established research practices in the field, e.g., [13,20]. The 5-point scale was particularly suitable for the target population in Turkey, facilitating comprehension across diverse educational backgrounds. The Ethics Committee of ÇağUniversity approved the study questionnaire.

The population of this study consists of consumers residing in Turkey who are 18 years of age or older and have completed at least a secondary school education. These criteria were established to ensure that participants have the cognitive and educational capacity to understand questions related to GM food technologies and to evaluate the survey items in an informed manner.

Given practical constraints such as limited resources and accessibility during data collection, a convenience sampling method was employed. The surveys were administered to individuals whom the authors could reach through their connections, allowing for the collection of data quickly and easily. While convenience sampling limits generalizability, it was deemed appropriate for exploratory consumer perception research, particularly in studies that utilize structural equation modeling (PLS-SEM), where model complexity is prioritized [25]. Data were collected between March and April 2024 using online methods. The survey was distributed through social media platforms (Instagram, Facebook), messaging apps (WhatsApp), and email networks. A total of 332 questionnaires were collected via Google Forms. After excluding nine responses due to incomplete or inconsistent answers, 324 valid responses were retained for analysis.

Since a structural equation model was used in this study, the requirements of this method regarding sample size were also examined. Hair et al. [26] argue that the most commonly used method for estimating the minimum sample size in PLS-SEM is the ‘10-times rule’ method. In PLS-SEM, the most commonly used method for estimating the minimum sample size is the ‘10-fold rule’ method [26]. Hoe [27] stated that a sample size of over 200 is sufficient for structural equation analysis. Among the variations of this method, the most common one is based on the rule that the sample size should be at least ten times the maximum number of internal or external model links pointing to any latent variable in the model [28]. In this study, there are a total of 31 internal and external model connections to latent variables. In this context, it can be said that the sample size is sufficient (324 > 310).

The Partial Least Squares Structural Equation Model (PLS-SEM), which has been frequently used in marketing in recent years, was used to analyze the data. This study used PLS-SEM due to the large number of variables in the model. Chin [25] stated that PLS-SEM is appropriate for such complex structures. SmartPLS 4.0 was used as the data analysis software. Before testing the research hypotheses, the measurement model was analyzed to evaluate the reliability and construct validity of the scale constructs. Internal consistency, reliability, convergent validity, and discriminant validity were examined to assess the validity and reliability of the constructs.

Table 1. Convergent validity.

Constructs	Code	Loadings (>0.6)	VIF	Cronbach’s Alpha (CA) (>0.7)	CR (>0.7)	AVE (>0.5)
Food Technology Neophobia	FTN1	0.831	2.248	0.907	0.906	0.583
	FTN2	0.824	2.547
	FTN3	0.654	2.049
	FTN4	0.837	2.556
	FTN6	0.823	2.641
	FTN7	0.700	2.345
	FTN8	0.645	1.910
Perceived Knowledge	PK1	0.664	1.700	0.845	0.848	0.590
	PK2	0.763	2.060
	PK3	0.965	2.445
	PK4	0.637	1.858
Perceived Benefits	PB1	0.602	2.061	0.870	0.870	0.632
	PB2	0.720	2.338
	PB3	0.944	2.671
	PB4	0.868	2.801
Perceived Risks	PR1	0.963	2.863	0.920	0.919	0.792
	PR2	0.851	3.701
	PR3	0.852	3.862
Attitude Towards GM Foods	AGM1	0.926	4.929	0.941	0.941	0.842
	AGM2	0.914	4.637
	AGM3	0.913	3.795
Intention to Consume GM Foods	IC1	0.840	2.548	0.890	0.891	0.732
	IC2	0.907	2.927
	IC3	0.817	2.471

shows the results of the measurement models.

4. Results

The model fit was evaluated using several commonly accepted indices in PLS-SEM. The Standardized Root Mean Square Residual (SRMR) was 0.041, below the recommended threshold of 0.08 [29], indicating a good model fit. The Normed Fit Index (NFI) was 0.876, approaching the conventional cut-off of 0.90 [30] and suggesting an acceptable model fit. Model fit was also assessed using the d_ULS and d_G indices. For the estimated model, the d_ULS value was 0.384, well below the 95% confidence interval upper bound (HI95 = 0.588), indicating a strong model fit. The d_G value was 0.262, also below the HI95 threshold (0.293), further supporting the adequacy of the model fit. These findings confirm that the hypothesized model demonstrates a good overall fit with the data, supported by multiple fit indices, including SRMR, NFI, d_ULS, and d_G.

In this study, reliability analysis was performed first, and the Composite Reliability (CR) and Cronbach’s Alpha (CA) coefficients were calculated. Table 1 shows that the CA and CR coefficients exceed the recommended threshold of 0.70 [26]. This indicates that the scales had adequate reliability. Hair et al. [26] suggest that the factor-loading values for each statement should be examined to evaluate the appropriateness of the measurement model before examining construct validity. In this context, the loadings of all items were calculated as a result of resampling (bootstrapping) applied to the indicator loadings, using a sample size of 1000. In PLS-SEM, each exogenous indicator factor loading in the reflective structure should have a correlation of 0.708 and above to explain at least a 50% variance. However, Hair et al. [26] stated that if the AVE (≥0.50) and CR (≥0.70) coefficients are above the recommended threshold value, statements with item loadings greater than 0.4 can be accepted. According to Chin et al. [25], item loadings above 0.6 indicate that the structure explains a significant portion of the indicator variance. In this context, because the loadings of the FTN5 and FTN9 statements on the fear of food technology scale were below 0.6, they reduced the AVE value below the recommended threshold value. Therefore, they were removed from the scale.

The next step in evaluating model measurement was to assess the validity of each construct. Convergent and discriminant validity were tested to test construct validity. The CR and AVE values were calculated for all indicators to evaluate convergent validity. The recommended threshold value for CR was 0.70, and the recommended threshold value for AVE was 0.50 [31]. Table 1 shows that all CR and AVE values for the factors are above the recommended threshold value. This demonstrates that the convergent validity of the scales was ensured.

Additionally, values of five or more for the Variance Inflation Factor (VIF) indicate a potential issue with multicollinearity. All the VIF values presented in Table 1 were below 5. Thus, there is no evidence of multicollinearity among the variables in the dataset.

We then focus on assessing the discriminant validity of the constructs within the measurement model. We used the criteria outlined by Fornell and Larcker [32] and the HTMT coefficients and cross-loading values, as suggested by Henseler et al. [33]. According to Fornell and Larcker [32], the square root of the AVE values of constructs should exceed their correlation coefficients. Upon reviewing

Table 2. Discriminant validity.

Fornell–Larcker Criterion
	FTN	PK	PB	PR	AGM	IC
FTN	0.763
PK	−0.254	0.768
PB	−0.414	0.441	0.795
PR	0.640	−0.333	−0.653	0.890
AGM	−0.420	0.574	0.728	−0.578	0.918
IC	−0.278	0.386	0.594	−0.468	0.726	0.856
The Heterotrait–Monotrait Ratio (HTMT)
FTN
PK	0.247
PB	0.408	0.440
PR	0.637	0.323	0.650
AGM	0.418	0.572	0.723	0.576
IC	0.275	0.380	0.593	0.467	0.726

The bold values are the square roots of the AVE. FTN, food technology neophobia; PK, perceived knowledge; PB, perceived benefits; PR, perceived risks; AGM, attitude towards GM foods; IC, intention to consume GM foods.

, we observed that the square root of each construct’s AVE value exceeded the correlation coefficients of the other constructs, confirming their discriminant validity.

In addition, the Heterotrait–Monotrait Ratio (HTMT) coefficients proposed by Henseler et al. [33] were used to test discriminant validity. According to this criterion, the average correlation of the statements of all variables in the study was expressed as the average ratio of the correlations of the statements of the same variable. In this sense, the HTMT coefficients should be below 0.85, depending on the theoretical relationships between the constructs. When the HTMT coefficients were analyzed, all values were below 0.85 (Table 2).

4.1. Common Method Bias

In the social sciences, standard method bias leads to measurement errors and undermines the validity of results. Standard method bias is a problem that artificially increases or decreases the relationships between constructs, depending on similar responses, when a dataset related to the scales is collected from a single source. Podsakoff et al. [34] suggested that this problem should be addressed in future survey studies. To determine whether a standard method variance problem existed, a Harman single-factor test and a one-factor model test were conducted in the confirmatory factor analysis. In the Harman single-factor test, which is applied by performing factor analysis on all observations simultaneously to avoid the standard method variance problem, more than one dimension should emerge in the factor analysis, or the variance explained by the first dimension should not account for the majority of the total variance. The factor analysis reveals the existence of a multi-factorial (four-dimensional) structure (eigenvalues > 1), with the first dimension accounting for 39.67% of the explained variance. The results of the one-factor model test, using confirmatory factor analysis, also supported the findings of the Harman one-factor test. When confirmatory factor analysis was performed on all observations under a single latent variable, the model fit values were very low [χ² = 4778.56, df = 252, χ²/df = 19.002; CFI = 0.81; IFI = 0.81; NFI = 0.80; GFI = 0.45; RMSEA = 0.236]. These results showed no bias in the data collected from the same source, which may have affected the study results, and there was no standard method variance problem [34].

4.2. Structural Model Assessment and Hypothesis Testing

A total of 324 participants participated in this study. When the demographic characteristics of the participants were analyzed, 37.3% were male, 62.7% were female, 48.4% were associate and undergraduate graduates, and 26.9% were postgraduate graduates. It was determined that 47.2% were between the ages of 21 and 30, 25.6% were between 31 and 40, 16.7% were 54, and 7.1% were 51 and older.

The model fit was assessed before evaluating the hypothesis tests. To evaluate the structural model, the Variance Inflation Factor (VIF), R², the effect size coefficient (f²), and the predictive power coefficient (Q²) were considered. To evaluate the structural model, the collinearity VIF values of the variables in the research model were first examined to determine whether a linear relationship existed between the variables. If the VIF values are less than five, the model does not have multicollinearity problems. In line with the information provided in

Table 3. Demographic characteristics of the sample.

Variables	(n = 324)		Variables	(n = 324)
Gender	n	%	Age	n	%
Male	121	37.3	Under 20	11	3.4
Female	203	62.7	21–30	153	47.2
Education	n	%	31–40	83	25.6
High school and before	80	24.7	41–50	54	16.7
Bachelor’s degree	157	48.4	51 and above	23	7.1
Postgraduate degree	87	26.9

, the VIF values of the variables in the research model were below the threshold value of 5. In this study, the R² determination coefficients and f² effect sizes of the variables were analyzed in the next step to evaluate the structural model. The R² value was examined by analyzing the structural model to reflect the predictive power of the model. The appropriate value should be higher than 0.1. Chin [25] stated that the explanatory power of 0.67 and above is high, between 0.33 and 0.67 is medium, and between 0.19 and 0.33 is weak. In this context, the exogenous dimensions explained the perceived benefits (R² = 0.287), perceived risks (R² = 0.437), attitudes towards GM foods (R² = 0.546), and intentions to consume GM foods (R² = 0.525). Because the value alone was insufficient to determine the predictive power of the structural model, PLS-Predict analysis was performed. The Q² values for the PLS-Predict analysis, performed to determine the out-of-sample predictive power, are presented in

Table 4. Structural model assessment.

Endogenous Variables	PLS Evaluation Criteria
	R2	Q2	VIF	Effect Size
Food Technology Neophobia	-	-	1.069	-
Perceived Knowledge	-	-	1.069	-
Perceived Benefits	0.287	0.221	1.444	Medium
Perceived Risks	0.437	0.361	1.744	Large
Attitude Towards GM Foods	0.546	0.280	1.000	Medium
Intention to Consume GM Foods	0.525	0.117	-	Medium

Small: Q² effect size < 0.15; Medium: Q² effect size < 0.35; Large: Q² effect size > 0.35.

. When Table 4 was analyzed, it was observed that the Q² values ranged from 0.117 to 0.361. Q² values above zero indicate that the model has predictive power.

The effect size (f²) was also analyzed to test the fit of the structural model. The effect size coefficient f² represents the proportion of exogenous variables in the model that explains the rate of variation in the endogenous variables. Cohen [35] suggested that f² values between 0.02 and 0.15 have a low effect, those between 0.15 and 0.35 have a medium effect, and those above 0.35 have a high effect. According to the f² statistics, as can be seen in Table 4, there is a low effect in H₁ (f² = 0.137), H₄ (f² = 0.055), and H₆ (f² = 0.040), whereas there is a moderate effect in H₃ (f² = 0.170). There is a high-level effect in hypotheses H₂ (f² = 0.590), H5 (f² = 0.476), and H₇ (f² = 1.114). Accordingly, each variable has a negligible effect on the other variables. If any variable is removed from the model, the model will be affected within the framework of the degree of the effect.

After the structural model was evaluated, Partial Least Squares Structural Equation Modeling (PLS-SEM) was used to assess the significance of the relationships between variables. After testing the measurement model in the PLS analysis, path analysis was used to test the linear relationships between the variables. To evaluate the significance of the PLS path coefficients, t-values were calculated by resampling 1000 subsamples from the sample (bootstrapping) and 95% bias-corrected confidence intervals. Bootstrapping with 1000 resamples was performed to estimate standard errors and confidence intervals for path coefficients. We applied bias-corrected and accelerated (BCa) confidence intervals, providing more accurate and robust estimates than percentile intervals. Using 1000 resamples balances computational efficiency with accuracy, and has been widely applied in PLS-SEM research to ensure stable and reliable parameter estimates [26,33].

Table 5. PLS-SEM analysis evaluation (bootstrapping = 1000).

Relations	Original Sample	t-Value	95% Bias Confidence Intervals		f2	Results
			2.50%	97.50%
H1: FTN → PB	−0.323	4.569 *	−0.452	−0.184	0.137	Supported
H2: FTN → PR	0.594	10.694 *	0.473	0.690	0.590	Supported
H3: PK → PB	0.359	5.162 *	0.219	0.484	0.170	Supported
H4: PK → PR	−0.182	3.314 *	−0.296	−0.078	0.055	Supported
H5: PB → AGM	0.612	7.772 *	0.440	0.748	0.476	Supported
H6: PR → AGM	−0.178	2.278 **	−0.357	−0.040	0.040	Supported
H7: AGM → IC	0.726	16.851 *	0.633	−0.804	1.114	Supported

p * < 0.01; p ** < 0.05. FTN: Food Technology Neophobia; PK: Perceived Knowledge; PB: Perceived Benefits; PR: Perceived Risks; AGM: Attitude Towards GM Foods; IC: Intention to Consume GM Foods.

presents the results of the hypothesis test. According to the results of the analyses, food technology neophobia negatively affected perceived benefit (β = −0.323, t = 4.569, p < 0.01), whereas it positively affected perceived risk (β = 0.594, t = 10.694, p < 0.01). Perceived knowledge positively affects perceived benefit (β = 0.359, t = 5.162, p < 0.01) and negatively affects perceived risk (β = −0.182, t = 3.314, p < 0.01). Perceived benefits positively affected attitudes toward GM foods (β = 0.612, t = 7.772, p < 0.01). Perceived risk negatively affected attitudes toward GM foods (β = −0.178, t = 2.278, p < 0.05). Attitudes towards GM food positively affected the intention to consume GM food (β = 0.726, t = 16.851, p < 0.01). Accordingly, Hypotheses H1, H2, H3, H4, H5, H6, and H7 are supported.

4.3. Out-of-Sample PLS-SEM Model Prediction and Comparison Analysis

It is insufficient to rely solely on in-sample evaluation criteria in structural model evaluations [36]. Therefore, an assessment based on the advanced out-of-sample criteria is essential. In this study, the out-of-sample predictive power of the research model was evaluated using a cross-correlated predictive ability test (CVPAT) procedure. Sharma et al. [36] introduced a novel out-of-sample statistical tool called CVPAT, which compares the overall predictive capabilities of two theoretically competing PLS-SEM models (the original model and the alternative). This method investigated the existence of an alternative model that provides better predictions than the original model. Within the scope of this study, the average loss difference between the PLS and the alternative models must be negative and statistically significant.

Table 6. PLS-Predict and CVPAT PLS-SEM analysis of results.

	PLS Loss	IA Loss	Average Loss Difference	t Value	p Value
Attitude Towards GM Foods	1.331	1.772	−0.441	6.001	0.000
Intention to Consume GM Foods	1.451	1.605	−0.154	3.600	0.000
Perceived Benefits	1.599	1.888	−0.289	3.727	0.000
Perceived Risks	0.902	1.303	−0.402	4.444	0.000
Overall	1.342	1.661	−0.319	5.257	0.000

shows that the average losses of the PLS model are smaller than those of the alternative model. In other words, the research model works better than the alternative model created by software.

5. Discussion

Various debates have occurred regarding the introduction of GM foods, their consumer use, benefits, risks, environmental and health safety, and limitations. In today’s rapidly evolving world, innovations in food technology have often sparked concerns among consumers. Consumers’ increased knowledge of food technologies can significantly influence their purchasing decisions. This study analyzed the effects of consumers’ fear of food technology innovation and perceived knowledge on their attitudes toward GM foods and purchase intentions within the framework of a model. Data collected from consumers aged 18 and above in Turkey were analyzed using PLS-SEM. The analyses revealed that food technology neophobia negatively impacts perceived benefits and has a positive effect on perceived risks.

The findings indicate that perceived knowledge has a positive effect on perceived benefits and a negative effect on perceived risk. While perceived benefits positively influenced attitudes toward GM foods, perceived risks hurt these attitudes. Finally, a positive attitude towards GM foods positively affects the intention to consume these products. All hypotheses established within the scope of the study were accepted, and the findings were generally consistent with the relevant literature [21,22,37,38,39,40]. Perceived risk may lead to skepticism and distrust, especially when consumers lack trust in and understanding of new food technologies [37]. Zhu et al. [38] found a significant negative relationship between consumers’ perceived risks and attitudes. According to Ngo et al. [39], despite having a positive correlation with purchase intentions, trust in GM foods greatly increases perceived advantages but has no discernible impact on risk assessment.

Additionally, the findings showed that greater awareness of GM foods is associated with higher views of advantages and lower perceptions of hazards, which in turn raises purchasing intentions. Chen and Li [21] found that consumer knowledge does not significantly increase the perceived benefits of applying gene technology to produce food products. According to Hassan et al. [40], customized educational interventions can significantly increase knowledge and favorably affect attitudes regarding GM foods, especially for groups with lower initial knowledge levels, such as older people, those without graduate degrees, and those with health-related educational backgrounds. Rabbanee et al. [23] revealed that consumers’ awareness of the benefits of GM foods increases their purchase of GM foods. Research by Caataneo et al. [41] suggests that individuals with lower levels of education tend to be more neophobic, indicating that resistance to genetically modified (GM) foods may also be influenced by social class or structural factors

In line with SDG 12, the study’s conclusions also have important implications for stakeholders seeking to reduce food waste and promote sustainable patterns of production and consumption. GM foods can help reduce post-harvest losses and food spoilage throughout the supply chain by enhancing crop resilience, extending shelf life, and improving resistance to pests and diseases. At various stages of production, distribution, and consumption, these technical developments can help reduce food waste and enhance the efficiency of food systems. Therefore, promoting the adoption of such innovations through favorable consumer perceptions can support larger initiatives aimed at achieving the SDG 12 goals. The study also emphasizes the importance of sustainability-focused education in shaping consumer attitudes and the adoption of food innovations, such as genetically modified foods. The findings show that consumers’ assessments of risks and advantages, which in turn influence their attitudes and purchase intentions, are strongly influenced by their perceived level of knowledge. This highlights the importance of developing educational initiatives and communication plans that enhance public awareness and understanding of sustainable food technologies. Targeted educational initiatives can reduce technology neophobia, increase trust, and encourage informed decision-making—all of which will eventually lead to responsible consumption practices that support sustainability objectives. This is especially true for groups with low baseline knowledge.

By addressing these concerns, this study contributes to the body of knowledge in academia while also providing valuable insights for educators, industry stakeholders, and policymakers seeking to promote sustainable food systems and consumption habits.

6. Conclusions

These results highlight the importance of addressing psychological obstacles such as neophobia and information gaps. This study provides insightful information about consumer acceptability of food technology developments, particularly in genetically modified foods. More broadly, this study advances several of the United Nations’ Sustainable Development Goals (SDGs). It contributes to Goal 2: Zero Hunger by providing insights into possible tactics to improve food availability and resilience by illuminating the elements that affect the adoption of genetically modified foods. Additionally, it supports Goal 12: Responsible Consumption and Production by emphasizing how communication and perceived knowledge influence consumer decisions. Finally, our study indirectly supports Goal 13: Climate Action, since GM food technology helps improve agricultural productivity and lessen environmental impacts. Creating inclusive, successful, and sustainable food policies requires an understanding of consumer behavior in this setting.

One of the study’s most original contributions is that the model it employs is directly linked to the Sustainable Development Goals (SDGs). More concretely, within the scope of SDG 12 (Responsible Consumption and Production), the model reveals how consumers’ level of knowledge and fear of innovation influence their attitude towards sustainable food technologies, especially GM foods. In this respect, the study contributes to consumer education, risk communication, and the development of policies that will encourage the adoption of sustainable agricultural practices. Consequently, sustainable education policies aimed at enhancing food literacy should intentionally encompass socioeconomically disadvantaged groups. Such inclusive approaches would support the fulfillment of Sustainable Development Goal 4 (Quality Education) and contribute to achieving Sustainable Development Goal 10 (Reduced Inequalities). This would promote educational equity and social inclusion in the context of the acceptance of food technology.

The study findings provide useful information for GM food managers and decision-makers. Government organizations, private biotechnology industries, policymakers, scientists, and professionals can benefit from understanding the key factors influencing consumers’ purchase intentions for GM foods. The research findings shed light on consumers’ attitudes and behaviors toward food technologies, underscoring the need for effective public policies that support the achievement of the Sustainable Development Goals (SDGs)—particularly SDG 12 (Responsible Consumption and Production), SDG 2 (Zero Hunger), SDG 3 (Good Health and Well-Being), and SDG 13 (Climate Action). These insights are also valuable for private sector actors, especially in the food and biotechnology industries, as they can inform the development of communication and product strategies aligned with consumer expectations, aimed at reducing food waste and fostering sustainable production. Educational initiatives underpinned by scientific literacy, transparency in food technologies, and the development of critical thinking competencies constitute a fundamental mechanism for alleviating innovation-related apprehensions (neophobia) and diminishing perceived risks, while concurrently enhancing the perceived benefits associated with technological advancements. Consumers’ self-assessed levels of knowledge exert a significant influence not only on individual decision-making processes but also on the formation of collective behavioral patterns at the societal level. Accordingly, sustainable educational strategies should extend beyond the confines of formal educational institutions and be complemented by multifaceted dissemination channels, including mass media, public service announcements, social advocacy campaigns, and digital platforms. The systematic incorporation of such knowledge-based approaches into both public awareness efforts and institutional education frameworks is consistent with the objectives of Sustainable Development Goal 4 (Quality Education). It contributes to the establishment of long-term behavioral change. This is particularly salient in reshaping the perceptions and attitudes of younger cohorts toward food technologies, thereby promoting informed, critical, and socially responsible consumption practices that support overarching sustainability agendas.

Therefore, this study’s findings contribute to a better understanding of consumer behavior toward GM foods, which can help create a desirable product market and sustainability in the GM food industry.

While this study provides valuable insights, it is important to acknowledge certain limitations. The sample used in the research does not fully reflect the general population in terms of age and gender distribution compared to national data from the Turkish Statistical Institute. Additionally, the use of online surveys and convenience sampling can over-represent certain demographic and socioeconomic groups, such as individuals with higher education levels and internet access. These factors limit the generalizability of the findings and introduce constraints to the study’s external validity.

This research contributes to sustainable food consumption goals by examining the effect of food technology neophobia on consumer attitudes and purchase intentions toward genetically modified foods using quantitative data. Additionally, the study employs an interdisciplinary approach by combining the disciplines of psychology, food science, and sustainability. This study has developed several recommendations for future studies. In future studies, the moderating effects of demographic variables such as age and education on the relationships within the model could be further investigated. Such analyses may contribute to the development of targeted marketing strategies for GM products tailored to different consumer segments. Additionally, cultural and regional differences (e.g., rural versus urban areas) can be examined comparatively to explore how they shape consumer attitudes and purchase intentions towards GM foods. Beyond perceived risk and benefit, future research may test alternative models that incorporate various psychosocial variables, such as trust, information source credibility, and social norms. Finally, rather than focusing solely on general GM products, future studies could investigate how consumer preferences and perceptions differ for specific GM product types (e.g., GM corn, soy, tomatoes).