Quality and safety concerns in agrifoods present intrinsic hazards due to unpredictability, potentially resulting in food waste and increased healthcare costs. When customers identify such faults and consider the product inedible, they are likely to dispose of it, hence increasing the total cost of food consumption. Consuming products with quality and safety issues can exacerbate healthcare costs. Consequently, worries regarding quality and safety in pro-poor agrifoods can lower consumers’ incomes, thereby reducing purchases, consumption, and utility theories. The income effect resulting from these risks causes alterations in customers’ purchasing patterns to conform to revised budget limitations. Individuals endeavor to optimize utility through their consumption patterns; hence, when confronted with quality and safety hazards in pro-poor agrifoods, they may pursue alternatives that offer superior utility. Ultimately, the elevated costs and the ensuing income and substitution impacts lead to a diminished demand for pro-poor agrifoods.

Assuming customers exclusively purchase conventional agrifoods g and pro-poor agrifoods f, with their income being constant at I1 in the short term, quality and safety issues induce substitution and income effects, hence diminishing consumer demand. As illustrated in Figure 1, the escalation in costs attributable to quality and safety issues alters the consumers’ effective budget line from I1 to I2. The demand for pro-poor agrifoods transitions from f1 to f2, indicating a general decline in consumer purchasing desire due to income and substitution impacts. This is exemplified by the transition from point A to point B on the utility maximization curve. A compensation budget line I3 is established using the tangent translation of I2 and U1, where f2f3 denotes the income effect, and f3f1 signifies the substitution effect. Figure 1 illustrates that quality and safety issues exert an “income effect” and a “substitution effect” on consumers, diminishing their propensity to repurchase.

Food quality and safety constitute integral components of the consumer experience; generally, customers favor high-quality products, with quality and safety being pivotal variables influencing their food choices and decisions (Feng et al., 2012; Grunert, 2005; Röhr et al., 2005). The prevalence of food quality and safety issues markedly heightens consumer apprehensions regarding products (Yin et al., 2016), leading consumers to perceive an elevated risk associated with pro-poor edible goods when purchasing such items following experiences with quality and safety concerns. Due to risk aversion, consumers will exercise greater caution in their selections and decisions about pro-poor agrifoods. The consumer experience is crucial to product consumption as it affects the enduring relationship between the product and its use (Siqueira et al., 2019). In instances of recurring purchases, the consumer’s utility function accounts for the “memory effect” (Yuan et al., 2023). The recollection of product consumption and its distinction from competing products influences consumer utility; specifically, the utility derived from the initial consumption experience impacts the decision to repurchase the product (Robson and Samuelson, 2011; Warlop et al., 2005). Consequently, consumers acquiring pro-poor agrifoods face quality and safety issues, leading to a negative experience that diminishes the utility derived from their consumption. This reduction in decision-making utility subsequently decreases consumers’ willingness to repurchase. Consequently, the first hypothesis is posited:

Consumer purchase intention is also affected by information (Xu et al., 2023); due to information asymmetry, consumers are unable to understand the product situation fully and may waver on the purchase decision. The professionalism and trustworthiness of information can increase consumer intention by promoting consumer confidence, and information can affect consumer purchase intention by influencing consumer perception (Meng et al., 2021). Consumer quality and safety evaluation attention and brand attention are the degree of access to and attention to evaluation and brand information. Quality and safety evaluation attention is the way consumers search for information. Social media has become an indispensable communication medium through which consumers seek product information and discuss and share their consumption experiences. As a result, consumer evaluations can also serve as endorsements or recommendations to other customers, which in turn influence the purchase intentions of these customers (Wang et al., 2012). On the one hand, consumers’ attention to product evaluation will weaken their experience memory, and they will form new product perceptions of the purchased pro-poor agrifoods, which will provide information for the decision-making of the next purchase, thus affecting the consumers’ willingness to repurchase. Therefore, hypothesis 2 is proposed:

Branding is a specific quality cue because consumers can use it to identify products (Crawford, 1997; Grunert, 2022). Brands have been shown to influence the perceived quality of a product significantly (Dodds and Monroe, 1985; Jacoby et al., 1971), e.g., as a basis for expected diet quality and expected health quality (Bredahl, 2004). When consumers pay more attention to the brand, they will have a more comprehensive understanding of the information and quality of the product, which will form a new perception in combination with the experience of the pro-poor agrifoods they have purchased, attenuating the solidified memories of the previous pro-poor agrifoods they have purchased, and thus affecting the consumers’ willingness to repurchase. Hypothesis 3 is thus proposed:

Based on social preference theory, individuals are more willing to help or donate to lower-income people (Bolton and Ockenfels, 2000; Charness and Holder, 2019). So, for pro-poor edible produce, consumers are more willing to buy and pay a premium, and this altruism of giving back to the community will make them feel the warm glow effect (Andreoni, 1989). The warm glow generated by repurchasing pro-poor produce is more significant for more public-spirited consumers. Therefore, the utility of purchasing pro-poor agrifoods is higher for public-spirited consumers. This can be partially offset by the reduced utility of food quality and safety issues, thus affecting consumers’ willingness to repurchase pro-poor agrifoods. Therefore, hypothesis 4 is proposed.

The research team collected data for the study through a consumer behavioral data survey. The questionnaire was conducted from January to March 2022 through the online questionnaire survey Questionstar platform.¹ The questionnaire was open to consumers in China, who were randomly invited to fill out the questionnaire through locations such as supermarkets and markets in municipalities directly under the central government and provincial capitals and through online dissemination. The survey employed stratified sampling based on population proportions across provinces, with random sampling conducted in each provincial capital city. The questionnaire covered consumers’ characteristics, household characteristics, and consumption behavior of pro-poor agrifoods. To address potential non-participating and inattentive participants, we excluded respondents whose questionnaires needed longer. We removed survey responses from the dataset that took less than 5 min to complete, with the average time for all responses being 13.8 min. Our study population was consumers who had purchased pro-poor agrifoods, so we excluded consumers who had not. In order to deal with missing values in the questionnaire, respondents with missing for many important variables in the survey were removed from the dataset. Finally, 533 observations were retained and used in the study. The sample of respondents was drawn from every province in China (except two regions, Tibet and Taiwan). The sample sources cover economically developed large and medium-sized cities and represent the study of consumer behavior. The distribution of the sample is shown in Table 1.

The research sample includes more women than men because most agrifoods home buyers are women. More than 60% of the sample is married, and a significantly higher proportion is highly educated, which may be related to the fact that the survey was in municipalities and provincial capitals. Household size was concentrated between three and five members, and household income was between 3,000 and 15,000 yuan.

How do quality and safety issues of pro-poor agrifoods affect consumers’ willingness to repurchase? The article uses a probit model to analyze this. This model is mainly used to estimate the situation in which consumer I chooses between two mutually exclusive alternatives J. In this study, consumers’ willingness to repurchase pro-poor agrifoods under different circumstances are “yes” and “no.” In this study, consumers’ willingness to repurchase pro-poor agrifoods under different circumstances is “yes” and “no.” Based on the individual utility function assumption proposed by McFadden (McFadden, 1972), it is assumed that quality and safety attributes and individual characteristics influence consumer utility. If a consumer chooses item j out of two choices, his utility function is:

In Equation 1, i denotes the ith individual consumer; j denotes the jth choice in the choice set consisting of two mutually exclusive alternatives (J = 0 or J = 1); Q denotes the product quality; Xv¨ denotes a series of individual feature vectors for the ith individual; and εij is a randomized perturbation term.

Individual consumers seeking to maximize utility implies that when Uij>Uik(k≠j), individual i will choose j. The probability that individual i chooses j can be expressed as:

In Equation 2, pi is the probability that consumers choose to buy pro-poor agrifoods. Quality and safety When consumers are willing to buy pro-poor agrifoods under the constraints of food quality and safety and individual characteristics, then consumers have higher utility.

Since consumers’ willingness to buy pro-poor food products is a (0, 1) dichotomous variable and depends on the research needs of its influence factor explanatory variables, this study uses a probit model for econometric analysis. The primary expression of the probit model is as follows:

In Equation 3, μ is a perturbation term that obeys a standard normal distribution so that the binary discrete choice model that affects consumers’ purchase of pro-poor agrifoods can be expressed as follows:

In Equation 4, Φ is the standard normal cumulative distribution function; Y∗ is the unobservable latent variable, and Y is the actual observed dependent variable, which is used to indicate whether consumers choose to repurchase pro-poor agrifoods; 0 means “do not buy,” and one means “buy”; X is food quality and safety issues, and x is the actual observed influencing factor. Therefore, the probit model of the influence of quality and safety issues on consumers’ decision to purchase pro-poor agrifoods can be built as follows:

In Equation 5, prob(Y=1∣Xi) is the probability that a consumer will purchase a pro-poor product (i.e., Y = 1). Xi is the independent variable; Xi indicates whether the purchase of pro-poor agrifoods has encountered a food safety problem; Xn represents the control variables, consumer personal characteristics, household characteristics, and purchasing channel; α0,β1,βn is the parameter to be estimated; εn is the disturbance term, i.e., the effect of the other independent variables that are not included (which has a mean of 0, has or does not have equal variance and ε is statistically independent of each other and its distribution is normal). β1 depends on the linear relationship between the independent variable and the dependent variable, reflecting the magnitude of the independent variable’s influence on the probability of the dependent variable. A positive β1 indicates that an increase in the independent variable raises the probability of the event occurring; a negative value indicates the opposite.

Benchmark regression explores the impact of pro-poor product quality and safety issues on consumers’ repurchase intentions. Table 3 shows the results of regression coefficients and marginal effects before and after adding control variables. According to the results in Table 3, encountering quality and safety problems reduces consumers’ repurchase intention, which is significant at a 1% significance level. According to the results of marginal effects, encountering quality and safety problems decreases consumers’ willingness to repurchase by 12.3% of consumers. After adding the control variables of consumers’ personal and family characteristics, the effect of experiencing quality and safety problems on consumers’ willingness to repurchase is still negative and significant at a 1% significance level. The marginal effect of adding the control variables shows that encountering quality and safety problems will reduce the willingness to repurchase for 11.8% of consumers.

This study conducts a robustness test in two aspects to verify the robustness of the results on the impact of encountering quality and safety problems on consumers’ willingness to purchase pro-poor agrifoods. The first is to conduct regression analyses of consumers’ encounters with quality standards, labeling problems, adulteration and counterfeiting, mold and rot, excessive pathogenic microorganisms, excessive heavy metals, excessive pesticides and veterinary drugs, and quarantine failures as explanatory variables, respectively. The second is to test the average treatment effect of the impact of quality and safety issues on consumer purchase intentions using propensity score matching.

Tables 4, 5 show the effects of different kinds of quality and safety problems on consumers’ willingness to repurchase, and the effects of adulteration and mold and rot on consumers’ willingness to repurchase when buying pro-poor agrifoods are all negatively significant at the 1% significance level. It further indicates that consumers’ willingness to repurchase pro-poor agrifoods will be reduced when they encounter adulteration and moldy and rotten conditions. Combined with the marginal coefficients, consumers’ willingness to repurchase decreases by 16.81% in the case of adulteration and 25.36% in the case of mold and rot. This result further validates the robustness of the benchmark regression.

However, the effects of different quality and safety problems on consumers’ willingness to repurchase differ. The regression results in Tables 4, 5 show that the effects of quality and safety problems encountered by consumers purchasing pro-poor agrifoods, such as quality standards, labeling problems, exceeding of disease-causing microorganisms, exceeding of heavy metals, exceeding of pesticide and veterinary drugs, and failing of quarantine, on the willingness of consumers to repurchase pro-poor agrifoods are not statistically significant. There are two main reasons for this. Firstly, consumers have a deeper understanding of quality and safety issues that can be empirically determined and significantly impact consumers’ willingness to repurchase. These quality and safety issues are categorized into quality and safety issues that consumers can perceive explicitly and quality and safety issues that consumers cannot easily perceive. Food safety refers to all those chronic or acute hazards that can make food products damaging to consumers’ health. Food quality includes all other attributes that affect the product’s value to the consumer. This distinction between safety and quality has significant implications for policy formation (Lusk et al., 2014). Consumers can rely on their own experience to determine the quality standards, adulteration, and mold and rot these three quality and safety issues if they encounter these problems of bad consumer experience to consumers to bring more negative impact. For consumers, it is not easy to perceive the quality and safety issues, labeling problems, pathogenic microorganisms, heavy metals, pesticides, veterinary drugs qua, routine failure, and other issues require professional knowledge and professional tools to assist in order to conclude; the average consumer is rarely equipped with this professional knowledge and equipment to judge, consumers do not have a profound experience of the judgment of these quality and safety issues and just the existence of mistrust. Therefore, the effect on consumers’ repurchase intention is not significant. Secondly, for consumers, the impact of food safety issues on purchase intention is more significant than the impact of quality issues, so for the quality and safety issues that consumers can quickly identify the quality standard, it is more about the size, appearance, freshness, and taste of the product, which does not have any actual negative impact on the consumers, and therefore the impact on the consumers’ willingness to repurchase is also not significant.

Propensity score matching (PSM) can effectively help researchers control for confounders by comparing the characteristics of the control and treatment samples to ensure that extraneous factors can more accurately excluded from estimating treatment effects. By matching subjects (individuals who received the treatment) with non-subjects (individuals who did not), the researcher can more accurately assess the actual impact of the treatment on the outcome. In this study, propensity score matching (PSM) was used to analyze the effect of consumers experiencing quality and safety problems on purchase intentions, and the results are displayed in Table 6. Based on the results of nearest neighbor matching, the treatment group’s mean value is reduced by 0.159 compared to the mean illumination value, and the t-value is significant at a 1% significance level. To further verify that consumers encountering quality and safety problems reduce consumers’ repurchase intention, caliper matching, kernel matching, and local linear regression matching are used to verify the robustness of the results and corroborate the robustness of the underlying regression.

Based on the baseline regression, three moderating variables, brand attention, evaluation attention, and altruism, and the interaction term with pro-poor agricultural product quality and safety issues are added to explore the moderating roles of brand attention, evaluation attention, and altruism in the influence of pro-poor agricultural product quality and safety issues on consumers’ purchase intention. The results are displayed in Table 7.

According to the results of model (1) in Table 7, the effect of quality and safety issues on consumers’ repurchase intention is negative. The effect of the interaction term between quality and safety issues and brand attention is positive and significant at the 1% significance level, meaning that brand attention attenuates the negative effect of quality and safety issues on repurchase intention. According to model (2) in Table 7, the effect of the interaction term of quality and safety issues and evaluation attention on repurchase intention is positive and significant at the 10% significance level, which further proves that the evaluation attention of consumers on pro-poor agrifoods attenuates the negative effect of quality and safety issues on repurchase intention of consumers. According to the model (3) in Table 7, the results show that the interaction term between quality and safety issues and public interest has a positive effect on consumers’ willingness to repurchase, which is significant at the 1% level of significance, and further proves that the stronger the consumers’ public interest is can weaken the negative effect of quality and safety issues on consumers’ willingness to repurchase. To a certain extent, consumers’ concern for brand and product evaluation and altruism can repair the harm caused by quality and safety problems in purchasing poverty-alleviating agrifoods.

According to the classification of different agrifoods, the effect of quality and safety problems on the willingness to repurchase six categories of pro-poor agrifoods, such as fruits, vegetables, grains and oils, tea, fresh meat, and fresh milk, was analyzed for heterogeneity. According to the results of the heterogeneity analysis in Table 8, the quality and safety problems encountered by consumers of two categories of agrifoods, fruits, and vegetables significantly reduce consumers’ willingness to repurchase. In comparison, the regression results for four categories of products, namely, grains and oils, tea, fresh meat, and fresh milk, are insignificant. This may be because fruits and vegetables are the most frequently purchased agrifoods for poverty alleviation, and the risk of quality and safety problems in fruits and vegetables is more and more intuitive, which has the most significant impact on consumers’ consumption experience and has the greatest influence on consumers’ willingness to repurchase.

Studies have shown that quality and safety issues can frustrate consumers’ public interest. Suppose consumers encounter quality and safety problems when purchasing pro-poor agrifoods. In that case, their willingness to repurchase significantly decreases by 12%, which means that food safety incidents force consumers to reduce their consumption of this type of product and look for alternatives with similar functions in other categories, which is consistent with the findings of other scholars (Bai et al., 2024; Barnett et al., 2016). For different kinds of food quality and safety problems, due to consumers’ limited knowledge and skills in judging quality and safety problems, they experience and feel more deeply about quality and safety problems than their existing knowledge and experience can judge, so adulteration and mold and decay have a more significant negative impact on consumers’ willingness to repurchase pro-poor agrifoods. At the same time, consumers’ negative perception of pro-poor agrifoods is more related to safety rather than quality. Hence, the effect of quality standards on consumers’ willingness to repurchase is insignificant.

This lousy experience for consumers is not static and will gradually change their perception of pro-poor agrifoods, according to their information acquisition. This is especially true for consumers who pay attention to brand and product evaluation because the brand can bring certain high-quality information and reputation assurance; consumers’ attention to product evaluation can also update the lousy influence of previous products. Therefore, the greater the consumers’ attention to brand and product evaluation, the more negative impact on repurchase intention can be attenuated when consumers encounter quality and safety problems.

Altruism can make consumers prefer agrifoods with altruism or warmth (Gangadharan et al., 2023). That is, consumers with high altruism are more willing to buy pro-poor agrifoods with altruism and will still be willing to support these products despite some defects. Therefore, the higher the level of altruism of consumers, the lower the negative impact of consumer behavior due to quality and safety problems encountered in consumption.

There is a significant difference in the negative effect of quality and safety issues on consumers’ repurchase intention for two categories of agrifoods: fruits and vegetables. In comparison, the regression results for four categories of products, grain and oil, tea, fresh meat, and fresh milk, are insignificant. There are two reasons for the difference. First, consumers usually buy fruits and vegetables periodically, so they are more sensitive to the quality and safety issues of fruits and vegetables than other products. Second, the quality and safety issues of products such as food and oil, tea, fresh meat, and fresh milk may be relatively abstract or infrequent factors for consumers to consider; therefore, the impact of quality issues of fruits and vegetables on consumers’ consumption experience may be more direct and obvious, and the impact on consumers’ repurchase intention is more significant.

The findings of this study align with previous research on consumer behavior, product quality and safety, and altruistic consumption in their overall direction, while also revealing several unique differences and theoretical extensions. Existing studies generally hold that perceptions of product quality and safety are core variables influencing consumer purchasing decisions, and that adverse incidents significantly weaken purchase intent and brand trust. Our empirical results similarly validate this conclusion: when quality and safety issues arise with poverty-alleviation agricultural products, consumer repurchase intent declines by 12%. This indicates that even in altruistic consumption contexts with public welfare connotations, rational evaluation still dominates certain purchasing decisions. Similar to general market goods, consumer concern for product reliability remains the foundation for sustaining long-term purchasing relationships. This study further reveals the moderating roles of altruism and information attention during quality safety shocks. Consumers’ altruistic tendencies mitigate the negative impact of quality safety concerns on repeat purchases. This suggests that in socially responsible consumption scenarios, such as poverty-alleviation products, moral motivations can partially substitute for functional trust, thereby maintaining purchase stability. When consumers place greater emphasis on brand and product review information, their sensitivity to quality issues is relatively lower. Brand trust and external reputation mechanisms can partially substitute for product attribute trust, thereby mitigating the adverse effects of risk perception. This aligns with perspectives from trust transfer and brand attachment theories. Additionally, another innovation of this study compared to previous literature lies in revealing differential impacts across different categories of agricultural and sideline products. Findings indicate that safety concerns significantly diminish repurchase intent for fruits and vegetables—products characterized by high perceptibility of quality and elevated risk exposure—while standardized goods like grains, cooking oils, tea, fresh meat, and fresh milk experience limited impact. This result supplements previous research shortcomings regarding the “product homogeneity” assumption, highlighting the interactive relationship between perceived quality risks and product characteristics.