2.1 Food Labeling

The food labeling system is used to satisfy consumers’ safety-seeking needs by providing accurate food-related information to ensure consumers’ health and safe eating habits (Lee, 2019). Food labeling can be used for systematic tracking and management of production and distribution information from the suppliers, with the advantage that consumers can choose foods safely and rationally based on hygiene- and safety-related information through food labeling.

The food labeling system in Korea is operated according to the classification standards and is divided into mandatory and voluntary labeling. In the case of mandatory labeling, the law stipulates labeling to protect consumers, while voluntary labeling is operated through a certification system to encourage production (Kim et al., 2018). Mandatory label information includes essential information on nutrition, consumer caution/warning, traceability, indication of sodium content, GMO and non-GMO content, irradiation, and freshness. In contrast, voluntary labeling includes information such as hazard analysis critical control point (HACCP), Good Agricultural Practices (GAP), certifications of being eco-friendly, health functional foods, children’s preference food quality, low-carbon agricultural products, animal welfare, geographical indication, traditional food quality, alcohol quality, as well as Korean Industrial Standard (KS) certifications.

In addition, if food labels are divided by attributes according to the system’s operational purpose, attributes such as quality standards, hygiene and safety, environmental ethics, and nutritional health are considered (Kim et al., 2018). The quality standard attributes include certifications like traditional food quality, alcohol quality, KS, agricultural standard, and geographical indication. The hygiene and safety attributes include HACCP, GAP, irradiated food labeling, and traceability management. The environmental ethics attributes include certifications in organic processing, organic products, organic livestock products, pesticide-free agricultural products, antibioticfree livestock products, low-carbon livestock production, and animal welfare. Finally, the nutritional health attributes include sodium content, healthful functional food, and children’s preferred food quality certifications.

Most previous research on food labeling was limited to individual labeling systems, such as genetically modified food (Hwang & Nam, 2021; Nam & Lee, 2022), health supplements (Moss, 2006; Vincent, 2017), and nutrition labeling (Brecher et al., 2000; Wang et al., 1995). This study, however, is not limited to analyzing consumers’ perceptions and characteristics of individual specific food labels but instead examines differences in consumers’ perceptions of the overall food labeling system.

2.2 Perceptions of Food Labeling (PFL)

Food labeling is an essential source of information in food selection and purchasing behavior and serves as a criterion for consumer purchasing choices. It effectively addresses and solves the consumer information asymmetry in the food selection market (Choi & Kang, 2015). Therefore, the Korean government has implemented policies to manage, promote, and educate consumers about foodrelated labeling and certification systems. Nevertheless, if consumers do not confirm or understand food labeling information, or if there is no change in purchasing behavior by food label recognition, it is difficult to evaluate the impact of the food labeling policy. Generally, consumers do not use food labeling properly (Miller & Cassady, 2015). Jackey et al. (2017) showed that although most consumers were familiar with food labels, fewer than half were able to interpret food labels accurately. According to the 2017 Food Behavior Survey by the Korea Rural Economic Research Institute, 51.4% of Korean adults asserted the importance of food labeling information in choosing food; however, only 18.3% answered that they use it.

In today's evolving consumer environment, where preferences and concerns about food safety are everchanging, PFL is increasingly being recognized as an advanced level of understanding and a fundamental initial stage of knowledge. This stage encompasses awareness and familiarity with labeling information, which is essential in shaping consumer attitudes and behaviors toward food choices. At this initial stage, PFL includes recognizing labels and gaining a general understanding of their meanings, such as the implications of organic certification or non-GMO labels (Visschers & Siegrist, 2015). This early recognition and understanding lay the groundwork for more informed consumer decisions, with consumers often relying on their initial perceptions of labels to make quick judgments (Miller & Cassady, 2015).

The impact of sociodemographic factors on consumer interactions with food labeling information is crucial, with education level, age, and gender as significant determinants. The role of education is evident, as higher education levels aer associated with an enhanced understanding and utilization of food labels, suggesting that educational attainment is a critical factor in promoting label literacy (Drichoutis et al., 2006). This underscores the potential of education to equip consumers with the skills to interpret food labels effectively, thereby enabling more informed food choices.

Age differences also play a role, with younger and older consumers displaying varying preferences and understandings, reflecting generational shifts in foodrelated concerns and awareness (Shamim et al., 2022; Van der Merwe et al., 2013). Younger individuals are less likely to check food labels (Ryu, 2016). Gender differences further influence the examination of food labels, with studies showing that women are more likely to scrutinize food labels than men (Ryu, 2016). This gender disparity suggests a deeper concern among women regarding nutritional information and food safety, potentially driving more health-conscious or informed food selection behaviors (Lee, 2019). Understanding these demographic nuances is essential for developing effective food labeling policies and education programs that address the diverse needs and behaviors of the consumer base, thereby fostering a more informed and health-conscious society.

Recent studies have further emphasized the importance of PFL in bridging the gap between essential label perception and more informed, information-based decisionmaking. Even a fundamental level of label awareness can significantly impact consumer purchasing behavior, underscoring the vital role of PFL in enabling informed consumer choices (Jacobs et al., 2011; Lundeberg et al., 2018). It has been observed that consumers' knowledge of food label information is closely linked with their level of informedness about nutrition, thereby affecting their purchasing decisions (Shamim et al., 2022).

2.3 Perception of Food Risk (PFR)

Perception of Food Risk (PFR) revolves around the subjective interpretation of potential dangers in food consumption. Risk, as characterized by Slovic (1993), pertains to the probability or uncertainty of facing an undesirable event. This notion emphasizes that personal subjectivity shapes risk perception more than objective data. The amalgamation of personal judgment and various cognitive and situational factors, as discussed by authors like Siegrist (2000) and Sjoberg (2001), further complicates this individualized perception. Consequently, this complexity necessitates that government entities, serving as risk managers, are attuned to the diverse consumer concerns regarding food risks.

Food risks can be classified into various categories depending on the criteria used. Kramers and Ravenswaay (1989) broadly divide food risks into chemical and biological hazards. The chemical hazards are subdivided into natural toxins and chemical additives (such as agrochemicals, food additives, altered GMOs, etc.) and environmental contaminants (including heavy metals, antibiotics, hormones, pesticide residues, etc.). On the other hand, biological hazards are split into bacteria (food poisoning bacteria, various germs, etc.) and zoonotic diseases (such as Bovine Spongiform Encephalopathy - BSE, etc.). The food risks mentioned vary over time; while some factors like food additives, pesticide residues, and GMOs are consistently discussed, others such as BSE, radioactive contamination, avian influenza, and foot-and-mouth disease are highlighted only during specific periods when they become relevant social issues.

This study considers food risk factors to include foreign substances, pesticide residues, the use of antibiotics in livestock and fish, inherent toxicity, food additives, heavy metals and endocrine disruptors, food poisoning bacteria, diseases in livestock, genetically modified organisms (GMOs), irradiated food, packaging hazards, and allergens. This classification allows for a comprehensive understanding of the various risks associated with food consumption, facilitating a more targeted approach to managing and mitigating these hazards.

Studies exploring the factors influencing PFR have indicated that objective knowledge about food can significantly reduce consumers’ psychological perception of risk (Klerck & Sweeney, 2007). This understanding implies a vital role for consumer education in alleviating undue anxiety about food risks. Research has shown that consumers' responses to food risks, such as fear or rejection, are often rooted in a lack of familiarity or perceived severity of these risks (Fife-Schaw & Rowe, 1996; Tucker et al., 2006). Additionally, the influence of specific factors like harmful ingredients or genetic modification on consumer decisions has been noted as overshadowing the impact of demographic variables (Gifford & Bernard, 2007).

However, the role of demographic factors in shaping PFR is undeniable. Research indicates that individuals with higher income or education levels generally exhibit lower anxiety about food risks (Hogan & Berning, 2012). In contrast, studies have found that women, particularly older ones or those with higher education, tend to have higher levels of anxiety regarding food hazards (Jun, 2014). This increased concern is linked to a heightened responsibility for family health. Specifically, research focusing on women raising children has highlighted their significant anxiety levels about food risks, underscoring their proactive approach to food risk management for their family's health (Davidson & Freudenburg, 1996).

These findings call for the need for customized communication strategies in food safety. These strategies should be designed to address various demographic groups' concerns and information needs. For instance, educational interventions could be pivotal in mitigating anxiety among higher-income or more educated groups, while targeted information could specifically address the concerns of women, especially those raising children.

The moderating effects of PFR on the relationship between PFL and food purchasing decisions are highly significant. Research has shown that when individuals perceive high levels of food risk, they are more inclined to scrutinize food labels, leading to more informed and cautious food choices (Siegrist, 2000; Slovic, 1993). For instance, studies such as Kang et al. (2021) have observed a change in consumer concerns about food additives, reflecting how evolving PFR, influenced by increased awareness and education, can change in how consumers interact with food labels and, ultimately, their purchasing decisions. Furthermore, the research underscores the critical role of information, mainly as conveyed through food labeling, in molding the PFR and subsequently affecting consumer behavior (Miles & Frewer, 2003). This emphasizes the importance of food labels in providing consumers with the necessary information to make informed decisions regarding food safety, thereby highlighting the integral link between adequate labeling, consumer risk perception, and behavioral responses. These findings suggest that PFR is a critical moderator in the relationship between PFL and food purchases. This indicates that strategies focused on enhancing food label clarity and consumer education can effectively influence food purchasing decisions toward safer choices.

2.4 Safe Food Purchase

The exploration of consumer behavior in food purchasing within this study is methodically segmented into two core aspects: the decision-making process behind buying of labeled food(BLF) and the intention to paying a high price for safe food(IHPS). This bifurcated approach provides a detailed insight into the complex interplay between food labeling and consumer purchase decisions, emphasizing the role of both informational and psychological factors in shaping these behaviors.

The inclination to opt for labeled food is significantly influenced by the information on these labels, ranging from nutritional content to provenance details. Labels are critical signals of quality, safety, and other desirable attributes, steering consumers toward certain products (Urala & Lähteenmäki, 2007). The success of these labels in transmitting valuable information hinges on their clarity and reliability, which directly affects consumer choice, highlighting the necessity for transparent and accurate food labeling (Caswell & Mojduszka, 1996).

On the other hand, IHPS for the assurance of food safety zeroes in on a distinct facet of consumer behavior that revolves around the quest for safety. This willingness is not only guided by the presence of safety labels but also by individual perceptions of risk associated with food consumption (Verbeke, 2005). The premium placed on safety, often molded by personal experiences and beliefs regarding food risks, prompts consumers to invest more in products perceived as be safer or that adhere to higher safety standards (Lobb et al., 2007).

Consumers actively seek out products with extended shelf lives, organic credentials, reputable brands, locally sourced ingredients, and non-GMO certifications as a strategy to mitigate the risks posed by hazardous foods (Jun, 2017). The perception of food-related risks has been positively linked to the inclination towards selecting safer food options (Song & Yoo, 2008), with consumer anxiety about food risks also favoring the purchase of eco-friendly products (Jun, 2014). Additionally, evidence suggests that consumers attentive to nutritional labels tend to make healthier food choices, as seen in preferences for products and menus with better nutritional profiles (Drichoutis et al., 2005; Krieger et al., 2013). This is further supported by findings that individuals who consult nutritional labels show a significantly higher intake of fruits and vegetables and a lower intake of fats (Satia et al., 2005), alongside more healthful consumption patterns in terms of calories, fats, cholesterol, sodium, dietary fibers, and sugars (Ollberding et al., 2011).

However, there is a contrasting narrative in some studies that suggests a reluctance among consumers to engage with nutrition and food safety labels, with no notable behavioral change post-consultation. Instances include adolescents who, despite their skepticism towards the safety of food additives, disregard nutritional education in favor of impulsive and simplistic buying patterns (Song & Choi, 2013). Similarly, the provision of calorie information in fast-food outlets was reported to have no significant effect on altering food selections across various demographics, including teenagers, adults, and the economically disadvantaged, possibly due to the perception that nutritional labeling demands a behavior change (Grunet & Wills, 2007).

This nuanced exploration maintains the original content’s integrity while ensuring a logical and cohesive flow, thereby effectively mapping the intricate relationship between food labeling and consumer purchasing behaviors.

3.1 Data

This study utilized the 2020 Consumer Behavior Survey for Food (CBSF) data from the Korea Rural Economic Institute. The data was collected via face-to-face interviews from June 10 to August 21, 2020. The CBSF included diverse questions encompassing aspects such as food purchasing habits, preferences for specific food types, attitudes toward food labeling and safety, and factors influencing food choices. A stratified sampling method was employed to ensure the survey's representativeness. This approach involved dividing the broader consumer population into subgroups based on age, gender, geographical location, and income level. The CBSF employed a stratified sampling method targeting consumers aged between 19 and 75 years residing across 16 metropolitan cities. This study focused on analyzing responses from 6,355 consumers who primarily handled food purchasing decisions. The overall demographic profiles of these participants are shown in Table 1.

Table 1.

Demographic Characteristics of Participants

3.2 Measures

The perception of food labeling (PFL) was rated as “2” if well known, “1” if heard of, and “0” if not known. The score s for 10 food labeling items (e.g., traditional food certification, KS for processed food, HACCP, country of origin, certified organic, geographic indication, traceability, GAP, GMO, and animal welfare certification) were summed, resulting in a scale ranging from 0 to 20 points.

The perception of food risk (PFR) is defined as a consumer’s subjective judgment about the likelihood of hazardous situations occurring in food consumption (Jun, 2017). It was measured using a 5-point Likert scale, ranging from “not concerned at all” to “very concerned,” across 12 food safety-related items. These items include foreign substances, pesticide residues, antibiotic use in livestock and fish, inherent toxicity, food additives, heavy metals and endocrine disruptors, food poisoning bacteria, diseases in livestock, genetically modified organisms (GMOs), irradiated food, packaging hazards, and allergens. The Cronbach's α for this scale was 0.922. This study employed a median split method to categorize participants based on their PFR and PFL. Specifically, the median values were established at 44.0 for PFR and 12.0 for PFL. Accordingly, participants were divided into two distinct groups for each perception category. For PFR, individuals with scores above the median were classified into the high PFR group, indicating a higher awareness and concern for food-related risks, while those with scores below the median were placed into the low PFR group, suggesting a lower perception of food risk. Similarly, for PFL, participants with scores above the median were categorized into the high PFL group, reflecting a significant value placed on food labeling information in their purchasing decisions. In contrast, those with scores below the median fell into the low PFL group, indicating a lesser reliance on or importance attributed to food labeling. This classification method facilitated a clear differentiation among participants, allowing for targeted analyses of how varying levels of food risk and labeling perceptions influence consumer behavior and safety considerations in food purchasing.

The buying of labeled food (BLF) was rated as “1” for never purchasing products, “2” for purchasing once or twice, “3” for occasional purchases, and “4” for frequent purchases. This rating was applied for the purchase experience of the 10 items measured in PFL. The intention to pay a high price for safe food (IHPS) was measured using a 5-point Likert scale ranging from “not at all” to “very positive.”

3.3 Analysis

SPSS ver. 21.0 (IBM) was used to analyze the data. A chi-squared test was employed to assess whether respondents’ characteristics were related to PFL and PFR.

A general linear model (GLM) was used to examine the main and interactional effects of PFL and PFR on BLF and IHPS. A multiple linear regression, an extension of simple linear regression to multiple independent variables, is a particular case of the general linear models, bound to a single dependent variable. The standard formula for multiple linear regression is given as:

Y_i = β₀ + β₁X_i1 + β₂X_i2 +...+ β_pX_ip + ε_i

for each observation i = 1, ..., n.

In the formula above, we consider n observations of single dependent variable and p independent variables. Thus, Y_i symbolizes the i^th observation of the dependent variable, while X_ij represents the i^th observation of the j^th independent variable, with j ranging from 1 to p. The β_j values are parameters that require estimation, and ε_i denotes the i^th independent error term, assumed to be normally distributed. Within the broader context of general linear regression, for each of the m > 1 dependent variables, a similar equation is employed, all sharing the same explanatory variables. These are concurrently estimated:

Y_ij = β_0j + β_1jX_i1 + β_2jX_i2 +...+ β_pjX_ip + ε_ij

for all observations indexed as i = 1, ..., n and for all dependent variables indexed as j = 1, ..., m.

The GLM is a foundational statistical tool that underpins a wide range of analyses commonly used in applied and social sciences (Nelder & Wedderburn, 1972). It is the basis for various statistical procedures such as the t-test, Analysis of Variance (ANOVA), Analysis of Covariance (ANCOVA), and regression analysis. Moreover, the GLM allows the inclusion of one or more control variables in the model, enabling the control of their influence while estimating the pure effects of the main independent variables. This type of analysis can also take the form of ANCOVA, an extension of ANOVA. However, the use of GLM provides a more general and flexible approach, allowing for handling more extensive and complex data structures (Huitema, 2011). This flexibility makes it particularly suitable for factorial designs where both main and interaction effects must be verified, as in social science research. Consequently, this study conducted an ANOVA-based GLM analysis to examine the impact of PFL and PFR on BLF and IHPS.

4.1 PFL and PFR according to Demographics

Table 2 shows the PFL and PFR according to participants’ demographics. Regarding PFL, gender, age, education, and food expenditure were all statistically significant. The low group for men and the high group for women showed a high frequency. With an increase in age, the frequency in the low group increased, while that in the high group decreased. The results of education were inconsistent, but as food expenditure increased, the frequency of the low group decreased, whereas, the frequency of the high group increased.

Table 2.

Perceptions of Food Risk and Food Labeling by Demographics Characteristics

Regarding PFR, factors including gender, age, education, and food expenditure were all statistically significant. Both men and women showed high frequency in the low group, and the results of education were inconsistent. In contrast, as age and food expenditure increased, the frequency in the low group increased, while that in the high group decreased.

4.2 Results of the GLM for BLF

Table 3 and Figure 1 show the result of the GLM for BLF. The main effects of gender, education, PFL, and PFR were statistically significant. Significant main effects were observed for gender, educational attainment (high school graduation or higher), PFL, and PFR. However, the moderating effect of PFR did not show statistical significance. Notably, PFL exerted the most substantial relative influence when considering factors such as gender, age, and education. For PFR, the average score was lower in the group with a high PFR (2.289) compared to the group with a low PFR (2.389).

Table 3.

GLM Results for the Buying of Labeled Food

Figure 1.

The effects of perception of food labeling and food risk on buying of labeled food.

4.3 Result of the GLM for IHPS

Table 4 and Figure 2 show the results of the GLM for IHPS. Except for age, the main effects of gender, education, food expenditure, PFL, and PFR were all statistically significant. The main effects(F = 40.943, p=.000) and interaction effect(F = 60.943, p=.000) of PFL and PFR were all statistically significant.

Table 4.

Result of GLM for Intention to Pay a High Prices for Safe Food

Figure 2.

The effects of perception of food labeling and food risk on buying of the intention to pay a high price for safe for food.

Specifically, in the group with low PFR, a notable difference in scores was observed between the low and high PFL groups (low PFL=3.387, high PFL=3.569). In contrast, within the high PFR group, the difference between the low and high PFL groups was minimal (low PFL=3.588, high PFL=3.599), indicating that the moderating effect of PFR was prominent in the low PFL group but not in the high PFL group.