Deconstructing Perceived Risk to Predict Suboptimal Food Purchase: A Strategy for Mitigating Food Waste

Abstract

Food waste poses a serious threat to global food sustainability, and consumer rejection of suboptimal food due to perceived risks is a significant factor exacerbating this issue—a phenomenon particularly pronounced in the Chinese context. Using survey data from 1022 Chinese consumers, this study investigates how multidimensional perceived risk and demographic characteristics jointly influence purchase intention toward suboptimal food. The results indicate that perceived quality risk, perceived health risk, and perceived social risk exert significant negative effects on purchase intention, whereas perceived psychological risk shows no significant effect. Moreover, the effect of perceived risk varies significantly across key demographic dimensions. Perceived health risk mediates the relationship between perceived quality risk and purchase intention. A significant interaction also emerges between perceived quality risk and perceived social risk: under conditions of high perceived social risk, high perceived quality risk substantially reduces purchase intention; under low perceived social risk, this negative effect persists but is attenuated. By delineating the differential effects and underlying mechanisms through which distinct risk dimensions shape purchase intention, this study not only advances the theoretical understanding of the interplay between multiple risk perceptions in consumer decision-making but also provides empirical evidence for reducing food waste from the consumption side, offering important implications for promoting sustainable consumption practices.

1. Introduction

Food waste has emerged as a pressing global challenge, attracting increasing scholarly and public attention due to its profound economic, environmental, and ethical implications. Globally, approximately 1.3 billion tons of food are wasted annually [1]. When food is not ultimately consumed, the substantial resources invested throughout its lifecycle—land, water, fertilizers, labor, and energy—are effectively squandered [2,3]. Consequently, food waste is increasingly recognized not only as a critical sustainability challenge but also as a pressing social and ethical issue [4,5]. In China, a key driver of food waste is consumer rejection of suboptimal food due to perceived risks, leading to significant economic loss and environmental pressure [6]. The term “suboptimal food” refers to edible and safe products that deviate from optimal standards due to minor imperfections, commonly categorized into three types: (a) cosmetic defects (e.g., irregular shape, size, or color), (b) products approaching their expiration date, and (c) package damage (e.g., dents or tears) [7,8]. Consumers’ low preference for such food triggers cascading negative effects across the supply chain. Direct waste at the retail level propagates upstream, leading distributors and farmers to reject non-standard produce—resulting in pre-market waste such as field disposal. Downstream, the economic pressure to avoid losses may incentivize the diversion of unsold or near-expired food into less regulated channels, where inadequate storage or handling may further exacerbate food waste.

Research on suboptimal food has been conducted predominantly in Western developed economies [8,9,10,11,12], with limited evidence from other national contexts. Exploratory surveys have also been carried out in Southeast Asian countries such as Malaysia and Indonesia [13,14]. Within mainland China, Loebnitz and Grunert [15] provided an early investigation into consumers’ willingness to purchase irregularly shaped produce. Subsequent studies in Taiwan have extended the Theory of Planned Behavior, identifying predictors such as personal attitudes, sensory appeal, environmental concern, food waste awareness, and perceived risk [16,17]. Further research has examined consumer preferences for product attributes [8,18], informational interventions [19], perceived benefits versus environmental concerns [20], and the interaction between original price and discount effects [21]. Although existing research has advanced our general understanding of suboptimal food acceptance, its direct application to mainland China remains problematic due to reliance on Western theoretical frameworks that overlook culturally specific determinants—particularly “face” (mianzi), which profoundly shapes consumer behavior in Chinese society. Given China’s unique cultural context, developing context-sensitive frameworks is imperative for both academic understanding and intervention design.

Perceived risk refers to consumers’ subjective expectation of negative consequences from purchase decisions, encompassing multiple dimensions such as financial, functional, social, psychological, physical, and time-related risks [22]. Since its introduction by Bauer [23], perceived risk has remained a cornerstone of consumer behavior research, with greater uncertainty elevating perceived risk and suppressing purchase intent [24,25]. This relationship is particularly salient in food choices, where conspicuous defects have been shown to elevate perceived risk and reduce purchase intention [26,27]. Despite its critical relevance, perceived risk in the context of suboptimal food remains underexplored—particularly in China. Beyond price and quality, consumers’ inherent risk perceptions regarding inferior quality or safety [28] may drive rejection alongside or independently of simple dislike [29,30]. However, the specific mechanisms through which perceived risk operates in this context remain unclear. This gap is especially consequential in China, where sociocultural dynamics—particularly “face” (mianzi)—are likely to interact with and amplify perceived social risk, a dimension largely overlooked in the existing literature. A range of established scales exist for measuring perceived risk [31], which researchers typically adapt to suit specific research contexts. To align with the suboptimal food purchase setting in China, this study assessed six foundational dimensions. According to Mitchell’s [32] perceived risk framework, time and financial risks are most salient in high-involvement, high-price purchase contexts that require extensive information search and financial commitment. In contrast, prior research has shown that low-involvement products and price discounts reduce these risk perceptions [33,34]. Consistent with this theoretical perspective, suboptimal food products are typically low-priced, often sold at a discount, and involve routine, low-effort purchase decisions. Furthermore, Chinese consumers rarely engage in post-purchase returns for such everyday food items. Therefore, the associated time and financial risks are theoretically justified as negligible in this study [35]. Informed by the actual consumption context and expert evaluation, this study conceptualizes perceived risk as the subjective concerns and uncertainties associated with purchasing suboptimal food, operationalized through four core dimensions: perceived quality risk, health risk, psychological risk, and social risk.

Sustainable consumption behavior has been extensively studied. White et al. [36] provided the SHIFT framework (Social influence, Habit formation, Individual self, Feelings and Cognition, Tangibility). Schanes et al. [37] systematically reviewed household food waste practices, while Stancu et al. [38] identified two routes to food waste: perceived behavioral control and routines. Vermeir & Verbeke [39] explored the attitude–behavioral intention gap in sustainable food consumption. Poulis et al. [40] recently highlighted the role of social influences and brand value in shaping food waste attitudes. However, research on sustainable consumption behavior regarding suboptimal food in the Chinese context still requires further investigation. The concept of “face” (mianzi)—a deep-seated sociocultural construct governing social standing—is largely absent from Western models. Face consciousness manifests as heightened sensitivity to others’ opinions and desire for external validation. Purchasing “perfect” food can signal positive social identity, while opting for suboptimal food risks conveying negative information about one’s economic standing. However, these models have not yet been systematically integrated with perceived risk theory in the context of suboptimal food.

Taken together, the literature reviewed above reveals three interrelated gaps that motivate the present study. First, prior research has provided comprehensive conceptualizations of perceived risk [22], but these frameworks were developed primarily in Western individualistic contexts. As discussed, the Chinese cultural construct of “face” (mianzi) profoundly shapes social evaluation concerns [35,41], yet existing models do not account for how such cultural specificity may amplify certain risk dimensions—particularly social risk. Second, although prior studies have identified various predictors of suboptimal food acceptance [8,15,16,17,18,19,20,21], they have typically examined main effects in isolation. The discussion above highlights that perceived risk comprises multiple dimensions, but it remains unclear whether these dimensions operate independently, sequentially, or interactively in shaping purchase intention. Third, demographic factors such as gender and age are known to influence consumer behavior generally [6], yet their moderating role in the specific context of suboptimal food risk perception has received limited empirical attention. To address these gaps, this study examines the mechanisms through which perceived risk shapes purchase intention for suboptimal food in the Chinese context. Specifically, it seeks to: (a) assess consumers’ perceived levels across key risk dimensions; (b) analyze the differential effects of each risk dimension and demographic characteristics on purchase intention; and (c) elucidate the underlying mechanisms by testing for mediating and moderating effects among these dimensions. The findings are expected to advance theoretical understanding of how multidimensional perceived risk operates in culturally distinct settings and offer actionable insights for retailers and policymakers seeking to promote suboptimal food acceptance and sustainable consumption.

2. Materials and Methods

To investigate the interrelationships among perceived risk dimensions and their mechanisms of influence on purchase intention, this study employs a quantitative methodology, utilizing correlation and regression analyses to investigate associations and test hypothesized causal pathways.

2.1. Questionnaire Design and Measure

The questionnaire comprised 20 items divided into three sections: perceived risk measurement (11 items), purchase intention measurement (4 items), and demographic questions (5 items), following a fixed sequence of informed consent, study introduction, scenario presentation, purchase intention measurement, perceived risk measurement, and demographic questions.

2.1.1. Perceived Risk

To ensure contextual relevance to Chinese consumers’ suboptimal food purchases, this study conducted group discussions and expert reviews involving five experts in the fields of food safety, consumer behavior, and marketing (three university professors with research expertise in food consumption and two industry practitioners with over five years of experience in food retail management), based on which perceived risk was operationalized through four dimensions: perceived quality risk, perceived health risk, perceived psychological risk, and perceived social risk. All items were measured on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). The initial questionnaire was piloted and refined based on feedback to enhance clarity and relevance, resulting in a final scale of 11 items across four dimensions (see Section 3.2).

2.1.2. Purchase Intention

Purchase intention was measured using a 4-item scale adapted from validated instruments in consumer research [17,42]. To ensure contextual relevance to suboptimal food purchases in China, the scale was refined based on insights from an expert panel. All items were rated on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). All measurement items are provided in Section 3.2.

2.1.3. Socio-Demographic Questionnaire

Participants also reported their sociodemographic characteristics, including gender, location, age, education level, income, and household size.

2.2. Data Collection

2.2.1. Respondent Selection

Existing research on suboptimal food has predominantly focused on consumers with are familiar with and have prior purchasing experience, leaving those who are knowledgeable about but seldom buy such foods largely underexplored. Understanding this segment is critical, as their purchasing behavior represents a dichotomous choice: identifying the reasons for their avoidance and enhancing their purchase intention constitute key pathways to promoting suboptimal food consumption and mitigating food waste. Given that perceived risk serves as a primary barrier to adoption, this study specifically targets consumers with low purchase frequency of suboptimal food. The screening criteria were as follows: (a) participants must be at least 18 years old; (b) participants must be responsible for or regularly involved in household food shopping; (c) participants must report purchasing suboptimal food “rarely” or “never” in the past three months; and (d) participants must demonstrate correct understanding of the suboptimal food concept as verified by a definition recognition question. Using a questionnaire survey, it systematically investigates how perceived risk dimensions shape their consumption decisions.

Participants were recruited through two complementary approaches. First, an initial pool was established by distributing paper questionnaires offline to consumers who met the screening criteria in Xinxiang. Research assistants distributed the questionnaires on-site, and participants completed and returned them immediately. Snowball sampling was then used to expand the sample. Second, field visits were conducted at four retail outlets offering suboptimal food in Henan Province (Zhengzhou, Luoyang, and Kaifeng): two supermarkets, one discount grocery store, and one fresh food market. These outlets were purposefully selected to represent different retail formats and socioeconomic areas (downtown, residential, and suburban) where suboptimal food is commonly available, thereby capturing diverse consumer profiles. At these outlets, potential participants were identified through accompanied shopping interviews. This multi-channel recruitment strategy was employed to access a diverse participant pool, thereby enhancing the depth and breadth of the data.

2.2.2. Experimental Procedure

Data were collected offline from November 2025 to January 2026. Prior to the main survey, a focus group discussion with 12 consumers was conducted in October 2025 to inform questionnaire design, followed by a pilot test with 31 Chinese consumers to assess clarity and relevance. The questionnaire was originally developed in English and translated into Chinese using a back-translation procedure to ensure conceptual equivalence.

Before participation, researchers explained the concept of suboptimal food to potential respondents by presenting three types of suboptimal food: (a) products approaching their expiration date (e.g., bread), (b) package-damaged products (e.g., dented cans), and (c) cosmetically imperfect products (e.g., irregularly shaped potatoes). Researchers then confirmed through screening questions that they belonged to the target population—consumers who rarely or unwillingly purchase such products. All participants provided informed consent and were assured that their data would be used solely for academic purposes. Upon completion, each respondent received a cash incentive of 5 RMB.

2.3. Data Analyses

All statistical analyses were conducted using SPSS Statistics 28 and AMOS 25. The specific roles of each software package were as follows:

SPSS 28 was used for: (1) descriptive statistics (means, standard deviations, frequencies) to summarize sample characteristics; (2) reliability analysis (Cronbach’s α > 0.7) to assess internal consistency; Kaiser-Meyer-Olkin (KMO) measure (>0.7) to assess sampling adequacy for factor analysis; (3) one-way ANOVA to examine demographic differences in perceived risk dimensions; (4) hierarchical regression analysis to test the main effects of perceived risk dimensions on purchase intention; (5) bootstrap analysis (5000 resamples) to examine the mediating effect of perceived health risk in the quality risk–purchase intention relationship; and (6) moderation analysis to test the interaction effect between perceived quality risk and perceived social risk on purchase intention.

AMOS 25 was used for confirmatory factor analysis (CFA) to assess the measurement model. Composite reliability (CR) was assessed with a threshold of >0.80. Convergent validity was supported by factor loadings > 0.50 and AVE > 0.50, and discriminant validity was established using the Fornell–Larcker criterion, whereby the square root of the AVE for each construct was required to exceed its highest correlation with any other construct.

Graphical illustrations were produced with Origin Pro 8.0.

3. Results

3.1. Descriptive Statistical Analysis

Of the 1100 questionnaires distributed (650 via offline paper distribution with snowball sampling and 450 via field visits with accompanied shopping interviews), 1086 were returned. After excluding respondents who failed the suboptimal food identification check (n = 7), did not complete the survey (n = 3), failed the attention check (n = 28), or exhibited inconsistent response patterns (n = 26), a final sample of 1022 valid responses was retained for analysis.

Table 1. Sociodemographic sample characteristics.

Variable	Category	Frequency	Percentage
Gender	Male	509	49.8%
	Female	513	50.2%
Location	Countryside	429	42.0%
	City	593	58.0%
Age	18–29 years	214	20.9%
	30–39 years	337	33.0%
	40–49 years	282	27.6%
	50 years and older	189	18.5%
Education	Secondary schools and below	289	28.3%
	Senior school	315	30.8%
	College degree	240	23.5%
	Graduate or above	178	17.4%
Income	3000 or less	394	38.6%
	3001–5000	289	28.3%
	5001–10,000	245	24.0%
	More than 10,000	94	9.2%
Household size	1 person or 2 persons	192	18.8%
	3 persons	364	35.6%
	More than 3 people	466	45.6%

Note. Income refers to participants’ average monthly income, reported in Chinese Yuan (CNY). The exchange rate at the time of the survey was approximately 1 USD = 7.06 CNY.

summarizes the demographic characteristics of the sample.

3.2. Reliability and Validity

As shown in

Table 2. Results of reliability and validity.

Latent Variable	Item	Cronbach’s α	KMO	CR	AVE
Psychology risk	I feel psychological pressure to purchase suboptimal food.	0.877	0.737	0.878	0.708
	I would not feel wise to purchase suboptimal food.
	I worry about purchasing suboptimal food.
Quality risk	Suboptimal food is of poor quality compared to normal food.	0.835	0.500	0.835	0.717
	Suboptimal food is less nutritious or functional than normal food.
Health risk	Purchasing suboptimal food is associated with a higher risk of illness.	0.905	0.727	0.908	0.769
	Purchasing suboptimal food will damage your health.
	It is not harmful to eat suboptimal food.
Social risk	Purchasing suboptimal food would be laughed at by those around me.	0.886	0.745	0.886	0.722
	Purchasing suboptimal food is considered unwise by those around me.
	Purchasing suboptimal food will affect my image in the eyes of those around me.
Purchase intention	I am highly likely to purchase the suboptimal food mentioned above.	0.936	0.864	0.936	0.786
	I think the suboptimal food described above are worth purchasing.
	I would recommend to my friends to purchase the suboptimal food described above.

, all constructs demonstrated high internal consistency, with Cronbach’s α coefficients ranging from 0.835 to 0.936 and composite reliability (CR) values exceeding the recommended thresholds of 0.70 and 0.80, respectively.

As shown in Table 2, KMO values for all latent variables—with the exception of perceived quality risk (measured by two items)—exceeded the 0.7 threshold, and the average variance extracted (AVE) for each construct surpassed the 0.5 criterion, supporting convergent validity. Discriminant validity was also established, as the square root of the AVE exceeded its highest correlation with any other construct in the model (see

Table 3. Discriminant validity of the measurement scales.

Latent Variable	Psychology Risk	Quality Risk	Health Risk	Social Risk	Purchase Intention
Psychology risk	(0.841)
Quality risk	0.021	(0.847)
Health risk	−0.029	0.491 **	(0.877)
Social risk	0.050	0.468 **	0.158 **	(0.850)
Purchase intention	0.002	−0.682 **	−0.592 **	−0.529 **	(0.887)

Note. Diagonal values in bold are the square roots of the average variance extracted (AVE). Off-diagonal values are the inter-construct correlations. ** p < 0.01.

).

3.3. Results for Perceived Risk Dimensions

To examine whether consumers’ perceptions of risk toward suboptimal food vary across dimensions, a one-way ANOVA was conducted on four perceived risk dimensions: psychological, quality, health, and social. The results showed that there were significant differences among the four perceived risk dimensions (F (3, 4084) = 910.452, p = 0.000). The results are presented in Figure 1.

Figure 1 presents the mean scores for each perceived risk dimension. In descending order, health risk (M = 6.39, SD = 1.01) scored highest, followed by quality risk (M = 5.87, SD = 1.07), social risk (M = 5.29, SD = 1.03), and psychological risk (M = 3.77, SD = 1.35). Notably, psychological risk was the only dimension with a mean below the scale midpoint of 4.00, suggesting that consumers do not experience substantial psychological burden—such as worry or perceived imprudence—when considering suboptimal food purchases. In contrast, health, quality, and social risk all exceeded the midpoint, indicating that these dimensions represent the primary sources of perceived risk in this context. Together, these findings suggest that consumer risk perceptions regarding suboptimal food are predominantly shaped by health, quality, and social considerations, rather than psychological apprehensions.

3.4. Influence of Demographic Variables on Perceived Risk Dimensions

Given that psychological risk did not emerge as a significant concern, subsequent analyses focused on the three dimensions that consumers perceived as most salient: quality risk, health risk, and social risk. Independent samples t-tests were conducted to examine gender and residential area differences, as both variables consisted of two groups. For variables with more than two categories—age (four groups), education level (four groups), monthly income (four groups), and household size (three groups)—one-way ANOVA were performed. Prior to ANOVA, Levene’s test was used to assess homogeneity of variances. Where the homogeneity assumption was met, post hoc pairwise comparisons were conducted using Fisher’s LSD test; in cases where variances were unequal, Tamhane’s T2 procedure was applied. Due to space constraints, only statistically significant results are reported in

Table 4. Gender differences in perceived risk dimensions.

Latent Variable		F	t	Sig.	Difference in Mean
Gender	Health risk	9.252	−2.450	0.014	−0.155
	Quality risk	2.813	−2.488	0.013	−0.167

Note. Independent samples t-tests were conducted. Gender was coded as 1 = male, 2 = female.

and

Table 5. Age, Education, and Household size differences in perceived risk dimensions.

Latent Variable		Test Method	(I)	(J)	Mean Value (I–J)	Significance
Age	Social risk	LSD	18~29 years old	30~39 years old	−0.322 *	0.006
			30~39 years old	≥50 years old	0.331 *	0.005
Education	Social risk	Tamhane	Middle school and below	Undergraduate or junior college	−0.383 *	0.002
			High school or technical secondary school	Undergraduate or junior college	−0.328 *	0.011
Household size	Social risk	LSD	1 or 2 persons	3 persons	−0.243 *	0.036
				≥4 persons	−0.381 *	0.001

Note. * p < 0.05.

.

As shown in Table 4, gender had a significant effect on health risk and quality risk, with female respondents reporting higher levels of both dimensions compared to their male counterparts. In contrast, no significant gender difference was observed for social risk (p = 0.536). Furthermore, place of residence was not significantly associated with any of the perceived risk dimensions (quality risk, p = 0.741; health risk, p = 0.919, social risk: p = 0.384).

Age, education, and household size significantly influenced perceived social risk, but had no significant effect on perceived quality risk or perceived health risk (see Table 5). Specifically, respondents aged 30–39 exhibited significantly higher perceived social risk than those aged 18–29 or ≥50. Regarding education, participants with undergraduate or junior college degrees reported greater perceived social risk than those with middle school and below, as well as those with high school or technical secondary school education. Similarly, respondents from households with three or more members perceived higher social risk than those from one- or two-person households. In contrast, no significant effects of age, education, or household size were observed for perceived quality risk (p = 0.511, 0.667, and 0.618, respectively) or perceived health risk (p = 0.095, 0.947, and 0.765, respectively). Income level also did not significantly affect any of the three risk dimensions (p = 0.929 for quality risk, 0.765 for health risk, and 0.369 for social risk).

3.5. Effects of Perceived Risk Dimensions on Purchase Intention

To examine the relative influence of perceived risk dimensions on purchase intention for suboptimal food, a hierarchical multiple regression analysis was conducted. Perceived health risk, quality risk, social risk, and psychological risk were entered as predictors, along with demographic covariates (gender, location, age, education, income, and household size). The results are presented in

Table 6. Results of regression analysis.

Variables	Model 1			Model 2
	B	Standard Error	p	B	Standard Error	p
Constant	8.395	0.165	0.000	8.692	0.221	0.000
Health risk	−0.402	0.024	0.000	−0.402	0.024	0.000
Quality risk	−0.369	0.025	0.000	−0.366	0.026	0.000
Social risk	−0.258	0.019	0.000	−0.257	0.019	0.000
Psychology risk	0.011	0.016	0.475	0.013	0.016	0.394
Gender				−0.060	0.043	0.164
Location				−0.045	0.048	0.350
Age				−0.022	0.021	0.301
Education				−0.034	0.022	0.115
Income				−0.029	0.022	0.198
Household size				0.011	0.029	0.716
Adjusted R2	0.622				0.623
F	420.622				169.412
p	0.000				0.000
Durbin-Watson		1.638
Mean purchase intention (SD)				2.34 (1.11)
	Dependent variable: suboptimal food purchase intention

.

Perceived health risk, quality risk, and social risk were found to have significant effects on purchase intention, whereas perceived psychological risk and the included sociodemographic variables did not (see Table 6).

3.6. Mediating Role of Perceived Health Risk

The descriptive findings indicate that perceived quality risk, perceived health risk, and perceived social risk significantly influence consumers’ purchase intention regarding suboptimal food. Drawing on perceived risk theory, two dimensions are particularly relevant to the suboptimal food context: physical risk—the potential threat a product may pose to one’s own or others’ health and safety—and functional risk—the possibility that a product fails to meet expected performance standards or underperforms relative to alternatives. In the context of suboptimal food, these risks manifest as concerns about inferior quality, reduced nutritional or functional value, and potential health hazards compared to perfect food products. Such concerns have been identified as significant barriers to consumer acceptance [34].

Importantly, perceived quality risk may not only directly influence purchase intention but also trigger a cascade of health-related concerns. That is, when consumers perceive suboptimal food as being of lower quality, they may infer associated health risks, ultimately leading to reduced purchase intention. This reasoning suggests that perceived health risk may function as a mediating mechanism in the relationship between perceived quality risk and purchase intention. Accordingly, the following hypothesis is proposed:

H1. 

Perceived health risk mediates the relationship between perceived quality risk and purchase intention.

To test the mediating role of perceived health risk in the relationship between perceived quality risk and purchase intention for suboptimal food, a bootstrap mediation analysis was conducted (5000 resamples). Perceived quality risk was specified as the independent variable, perceived health risk as the mediator, and purchase intention as the dependent variable. Control variables included gender, place of residence, age, education level, income, and household size. The results are presented in Figure 2.

As shown in Figure 2, the indirect effect of perceived quality risk on purchase intention for suboptimal foods through perceived health risk (β = −0.171) is significant, as indicated by a 95% CI [−0.228, −0.123] that does not include zero. Meanwhile, the direct effect of perceived quality risk on purchase intention (β = −0.526) is also significant, with a 95% CI [−0.574, −0.478]. These results suggest that perceived health risk partially mediates the relationship between perceived quality risk and purchase intention, thereby supporting Hypothesis H1.

3.7. Moderating Role of Perceived Social Risk

Suboptimal food purchases are embedded in specific social contexts. In China, where concern for social reputation is culturally salient, consumers anticipate potential public scrutiny when considering such products, thereby heightening perceived social risk. According to social identity theory, individuals derive part of their self-concept from membership in social groups and strive to maintain a positive social identity [43]. Consumption choices serve as visible signals through which individuals communicate their social standing and group affiliation. In the Chinese cultural context, “face” (mianzi) represents a deeply internalized concern with social image and external evaluation, functioning as a powerful source of normative pressure [35]. Purchasing aesthetically perfect food signals positive social identity—competence, taste, and economic standing—whereas opting for suboptimal food risks conveying negative information about one’s judgment or financial capability, thereby threatening social identity.

Within our perceived risk framework, “face” operates as a cultural mechanism that amplifies perceived social risk. Specifically, social risk—defined as the anticipated loss of social esteem or approval from reference others—becomes particularly salient in face-conscious cultures. When consumers evaluate suboptimal food, face consciousness heightens their sensitivity to potential public scrutiny, thereby elevating the perceived social risk associated with the purchase decision. Consequently, perceived social risk exerts not only a direct inhibitory effect on purchase intention (alongside perceived quality risk) but also a moderating effect on the relationship between quality risk and purchase intention. Specifically, the negative effect of quality risk on purchase intention may be amplified when social risk is high, as quality concerns become compounded by fears of social disapproval. This integration advances existing perceived risk models by: (1) extending Western-centric frameworks to account for cultural amplification of social risk; (2) revealing that social risk can moderate quality risk, challenging the assumption of independent risk dimensions; and (3) incorporating social and reputational consequences through social identity theory.

Based on this reasoning, the following hypothesis is proposed:

H2. 

Perceived social risk moderates the relationship between perceived quality risk and purchase intention of suboptimal food.

To test Hypothesis 2, this study employed Model 1 in Process 4.2, developed by Hayes, to test the moderating effect. Covariates (gender, location, age, education, income, household size), perceived quality risk (the independent variable), perceived health risk (the moderator), purchase intention as the dependent variable. The results are presented in

Table 7. Results of hierarchical moderated regression analysis.

Variables	B	Standard Error	t	p
Constant	2.716	0.179	15.177	0.000
Gender	−0.081	0.048	−1.698	0.090
Location	−0.030	0.054	−0.562	0.574
Age	−0.014	0.024	−0.608	0.544
Education	−0.027	0.024	−1.128	0.259
Income	−0.029	0.025	−1.155	0.248
Household size	−0.006	0.033	−0.184	0.854
Quality risk	−0.608 ***	0.027	−22.663	0.000
Social risk	−0.238 ***	0.021	−11.263	0.000
Quality risk × Social risk	−0.055 ***	0.012	−4.473	0.000
Adjusted R2		0.529

Note: *** p < 0.001.

.

As shown in Table 7, the interaction between perceived quality risk and perceived social risk was statistically significant (β = −0.055, p < 0.001), indicating that perceived social risk exerts a significant moderating effect on the relationship between perceived quality risk and purchase intention. Thus, Hypothesis 2 is supported.

To probe the nature of this interaction, simple slope analyses were conducted following Aiken and West’s [44] procedure. Perceived social risk was dichotomized into high and low groups based on one standard deviation above and below the mean. Figure 3 illustrates the relationship between perceived quality risk and purchase intention for each group.

As shown in Figure 3, when consumers perceived high social risk, quality risk had a significant negative impact on purchase intention (β = −0.68, p < 0.001). When consumers perceived low social risk, quality risk still had a significant negative impact on purchase intention, but the effect size decreased (β = −0.54, p < 0.001). Therefore, the moderating effect of perceived social risk is more pronounced at higher levels, exerting a stronger dampening effect on the relationship between perceived quality risk and purchase intention.

4. Discussion

4.1. General Discussion

The findings reveal that consumers’ perceived risk toward suboptimal food comprises three primary dimensions—quality risk, health risk, and social risk—each significantly and negatively influencing purchase intention. Notably, psychological risk was not found to be a significant component of consumers’ perceived risk in this context.

4.1.1. Perceived Quality and Health Risk

According to cue utilization theory, consumers rely on extrinsic and intrinsic cues to infer unobservable product attributes such as quality and credibility [45]. Purchase decisions and value perceptions emerge from the interplay between perceived benefits—economic, functional, and psychological—and the resources (e.g., money, time, effort) required to obtain them [46]. In the context of suboptimal food, price reductions—commonly associated with inferior quality [47]—may lead consumers to infer diminished quality in terms of health and taste [46], thereby heightening concerns about safety and potential health risks.

The mediating role of perceived health risk in the relationship between perceived quality risk and purchase intention can be understood through this lens. Perceived quality risk concerns product performance (e.g., taste, freshness, texture), whereas perceived health risk concerns physical safety (e.g., food poisoning, illness). Although a product may be perceived as lower in quality without being unsafe, suboptimal food features (e.g., unusual appearance or near-expiry dates) may lead consumers to infer not only quality deficits but also potential health hazards [28]. Consumers tend to associate atypical product features with lower quality [48]; in some cases, such deviations may even evoke associations with severe risks, such as genetic mutations [15,49].

These risk judgments are shaped by the decoding of specific product cues. Remaining shelf life functions as a primary indicator of freshness and safety [50,51]. Color—a key component of “natural packaging”—informs first impressions and expectations regarding taste, texture, and overall quality; abnormal color changes (e.g., darkened bananas) directly trigger safety concerns [52,53]. Similarly, damaged packaging is interpreted as a signal of potential contamination or quality degradation, amplifying perceived health risk [30,54].

Our finding that perceived quality risk and health risk negatively influence purchase intention is consistent with prior research. Siegrist & Hartmann [55] demonstrated that perceived health and safety risks are primary barriers to accepting atypical food products. Similarly, de Hooge et al. [8] found that consumers associate suboptimal food with lower quality and reduced taste expectations. However, our study extends these findings by revealing that quality risk operates partially through health risk—a mediated pathway not previously examined. This contrasts with studies that treated quality and health risks as independent predictors [55], suggesting that their relationship may be more complex than previously assumed.

These findings suggest a coherent interpretive pathway: through the negative appraisal of observable cues—such as shelf life, color, and packaging condition—consumers develop specific concerns about product safety, freshness, and intrinsic quality. These specific concerns coalesce into a generalized perception of risk associated with the “suboptimal” label, which in turn systematically suppresses purchase intention. This interpretation not only aligns with the existing literature on risk signaling but also extends it by elucidating how multidimensional risk perceptions are formed and integrated in the context of suboptimal food.

4.1.2. The Dual Role of Perceived Social Risk

Perceived social risk emerged not only as a direct negative predictor of purchase intention but also as a significant moderator of the relationship between perceived quality risk and purchase intention. This dual role can be interpreted through the lens of impression management theory [56], which posits that individuals are motivated to cultivate positive social images and avoid negative evaluations from others. In consumption contexts, this translates into concerns that purchasing suboptimal food may deviate from socially desirable norms, inviting disapproval or ridicule from one’s social circle.

It is plausible that this mechanism may be particularly pronounced in the Chinese cultural context, where the concept of “face” (mianzi) plays a foundational role in interpersonal relations and social standing [35]. Face consciousness could manifest as a heightened sensitivity to others’ opinions and a desire for external validation and group belonging. Within this framework, the ability to purchase “perfect” or unblemished food might function as a subtle status signal [57]. Conversely, opting for suboptimal food—especially in public or social settings—could risk conveying negative information about one’s economic standing or judgment, thereby threatening face and inviting social sanctions.

Accordingly, consumers with elevated perceived social risk are not only directly deterred from purchasing suboptimal food but also exhibit heightened sensitivity to quality-related concerns, as the latter may amplify potential social disapproval. This finding extends prior research by demonstrating that social risk does not operate in isolation [33]; rather, it interacts with product-related risk perceptions, magnifying their impact on purchase decisions. The moderating role of perceived social risk thus underscores the importance of embedding consumer decision-making within its broader sociocultural context.

4.1.3. The Non-Significant Role of Perceived Psychological Risk

Contrary to Zhong et al. [24], who found psychological distress negatively affected purchase intention across suboptimal food types, our results show no significant direct effect of perceived psychological risk. This divergence can be explained by three considerations.

First, the conceptual distinction between psychological and social risk is critical. Psychological risk pertains to introspective concerns—such as inner unease, worry, or regret arising from one’s own decisions—independent of external evaluation. Social risk, by contrast, centers on individuals’ perceptions of how their behavior is viewed by others, carrying a strong interpersonal orientation. In the Chinese cultural context, where “face” (mianzi) concerns are paramount, individuals tend to prioritize external social approval over internal psychological states [35]. Thus, the inhibitory effect of psychological risk is likely overshadowed by the more salient social risk—suggesting that mianzi operates as a cultural antecedent that amplifies the weight of social risk in decision-making.

Second, the cultural dynamic of “prioritizing face over inner self” offers a compelling explanation. In traditional Chinese thought, “face” (mianzi) and “inner self” (lizi) are conceptually distinct yet interrelated: the inner self serves as the substantive foundation of face, while face functions as its external manifestation [58,59]. Mianzi is predominantly other-oriented, representing public self-image and reputation, whereas lizi is self-oriented, representing inner reality and personal morality [58]. However, in everyday social practice, individuals are socialized to prioritize the preservation of external image over introspection on internal feelings [60]. This orientation extends to consumption decisions, wherein consumers are more attuned to whether their behavior appears appropriate and respectable in the eyes of others than to the potential psychological consequences for themselves [60]. This cultural predisposition may attenuate the role of psychological risk in shaping purchase intention.

Third, situational factors embedded in the research context may further suppress psychological risk. Price discounts—a common marketing strategy for suboptimal food—enhance perceived economic value, thereby offsetting potential concerns about product imperfections. Moreover, purchasing suboptimal food carries altruistic connotations: by reducing food waste and contributing to environmental sustainability, consumers engage in behavior with positive societal externalities [61]. This moral framing legitimizes the consumption act, positioning it as a “wise and responsible” choice. Under such conditions, consumers’ attentional focus is directed away from introspective psychological concerns toward utilitarian and ethical considerations, further accounting for the non-significant effect of perceived psychological risk.

Collectively, these explanations indicate that the non-significance of psychological risk is not a methodological artifact but a theoretically meaningful finding, underscoring the context-dependent and culturally moderated nature of risk perception. Specifically, mianzi functions as a cultural antecedent that amplifies social risk salience rather than serving as an independent mechanism or being fully subsumed under social risk.

4.2. Theoretical Contributions

This study advances the literature on suboptimal food consumption by developing a culturally grounded, multi-dimensional framework of perceived risk in the Chinese context. Moving beyond previous research that predominantly employed unidimensional risk measures, this study systematically examines the differential effects of distinct risk dimensions on purchase intention and elucidates the structural relationships among them. Several theoretical contributions emerge from this investigation.

First, by disaggregating perceived risk into its constituent dimensions, this study reveals that health risk, quality risk, and social risk are the primary drivers of purchase intention, whereas psychological risk exerts no significant direct effect. This finding challenges the assumption that psychological cost is a universally salient component of perceived risk [22] and suggests that, in certain cultural and product contexts, its influence may be attenuated or subsumed by more socially oriented concerns.

Second, this study extends the understanding of demographic heterogeneity in risk perception. The results indicate that gender differences are domain-specific: female consumers exhibit heightened sensitivity to quality and health risks, but not to social or psychological risks. In contrast, age, education, and household size only significantly shape perceived social risk. These nuanced patterns underscore the importance of disaggregating both risk dimensions and demographic categories, moving beyond generalized claims about demographic effects.

Third, and most importantly, this study delineates the structural architecture linking the risk dimensions. Perceived quality risk not only directly suppresses purchase intention but also operates through perceived health risk as a mediating mechanism. This mediation pathway is further moderated by perceived social risk, which amplifies the negative impact of quality risk when social risk is salient. By uncovering this moderated mediation model, the study provides a more complete account of how multiple risk perceptions jointly shape consumer decisions—a complexity that prior research, focused on isolated main effects, has largely overlooked.

Collectively, these contributions advance theoretical understanding of perceived risk as a multidimensional construct embedded in cultural context. They also lay a foundation for future research to explore cross-cultural variations in the salience and interrelationships of specific risk dimensions.

4.3. Practical Implications

The findings offer several actionable implications for retailers, marketers, and policymakers seeking to promote the consumption of suboptimal food and reduce food waste.

First, addressing quality and health concerns through transparent communication. Given that perceived quality risk and health risk are the two most salient barriers to purchase, retailers should prioritize communicating product safety and quality assurance in their marketing efforts. Messaging should explicitly convey that suboptimal food—despite aesthetic imperfections—meets the same safety and nutritional standards as perfect produce. For example, retailers can emphasize that taste, nutritional value, and freshness remain uncompromised, or are only minimally affected in ways that do not detract from the consumption experience. Leveraging endorsements from credible authorities (e.g., food safety experts, nutritionists) through in-store digital displays or packaging labels can further enhance consumer trust and reduce uncertainty.

Second, leveraging social risk to reframe the purchase narrative. Perceived social risk emerged as both a direct inhibitor and a moderator of purchase intention. To counter this, retailers should strategically align suboptimal food purchases with consumers’ desire for social recognition and face gain. Sales personnel can frame the purchase as a discerning and socially conscious choice—one that demonstrates knowledge, resourcefulness, and commitment to environmental sustainability. Emphasizing the prosocial benefits of reducing food waste can transform what might otherwise be seen as a mianzi—threatening act into a source of pride and social acknowledgment. This reframing is particularly important in cultural contexts where face consciousness is salient, such as China and other East Asian societies.

Third, tailoring risk communication to specific consumer segments. The demographic variations observed in this study underscore the need for differentiated communication strategies. From a policy perspective, regulators and public health authorities should: For female consumers (higher perceived quality and health risk): Mandate or encourage messaging that reassures food safety, nutritional integrity, and health implications, while de-emphasizing social or psychological risks. For consumers aged 30–39, those with associate or bachelor’s degrees, and one- or two-person households (higher social risk sensitivity): Promote campaigns emphasizing the social acceptability and prosocial value of suboptimal food purchases—highlighting waste reduction and environmental benefits to mitigate social image concerns. These segment-specific policy interventions can more effectively reduce perceived risk barriers and encourage sustainable food consumption at the population level.

These implications provide a roadmap for converting consumer resistance into acceptance. By addressing the specific risk dimensions that matter most to different consumer groups, and by framing suboptimal food purchases in ways that resonate with cultural values and social motivations, retailers can more effectively promote sustainable consumption behaviors. While grounded in the Chinese context, the dynamics of face and social risk extend to other markets where social image shapes consumer behavior, as evidenced across East Asian and Western settings [41]. Accordingly, the proposed strategies—particularly those addressing social risk—offer valuable guidance for food retailers operating in or entering such markets.

4.4. Limitations and Future Research

Despite these limitations, this study offers valuable directions for future research. First, suboptimal food was examined as a broad category; future research should distinguish between specific types (e.g., imperfect fresh produce vs. near-expiry packaged goods) to assess whether perceived risk and purchase intention differ across subcategories. Second, the reliance on self-report measures may limit ecological validity. Subsequent studies could employ observational methods to test these effects in real-world settings. Third, snowball sampling may have introduced selection bias, as participants recruited through social networks often share similar characteristics, limiting sample representativeness. Consequently, findings are most directly generalizable to populations resembling the initial seed participants (e.g., urban consumers with active social media ties), whereas generalization to broader or more diverse groups (e.g., rural populations, older adults) should be made with caution. Future research should employ probability sampling to enhance external validity. Fourth, the cultural specificity of the sample presents both a boundary condition and an opportunity for theoretical advancement. While “face” (perceived social pressure) is particularly salient in East Asia, its role in sustainable consumption remains underexplored elsewhere. Cross-cultural comparisons could test the generalizability of this framework and integrate cultural psychology and behavioral economics to examine how face-related motives interact with contextual factors across societies, thereby enriching theory in this domain.

5. Conclusions

This study investigates the multifaceted role of perceived risk in shaping Chinese consumers’ purchase intention toward suboptimal food. Analysis of survey data from 1022 respondents reveals that perceived quality, health, and social risks significantly deter purchase intention, while perceived psychological risk does not. Perceived health risk mediates the quality risk–purchase intention relationship, and a significant interaction emerges between quality risk and social risk: the negative effect of quality risk is attenuated under high (vs. low) social risk. Additionally, demographic differences moderate risk perceptions. Female consumers report higher perceived quality risk and health risk than males, whereas no significant gender difference was found for social risk. Perceived social risk varies significantly by age, with consumers aged 30–39 exhibiting higher levels than those aged 18–29 or 50 and above, while age has no significant effect on perceived quality risk or perceived health risk. Within the scope of this sample, the findings elucidate differential pathways and interactive mechanisms of risk perceptions, offering preliminary theoretical insights and practical guidance for retailers to mitigate risk perceptions, facilitate suboptimal food acceptance, and reduce food waste.