Skip to main content


Front. Sustain. Food Syst., 16 December 2022
Sec. Agro-Food Safety
Volume 6 - 2022 |

Consumer preference and willingness to pay for low-residue vegetables: Evidence from discrete choice experiments in China

Jian Wang1 Liangru Zhou2 Zhilong Ni3 Wenhao Wu4 Guoxiang Liu2 Wenqi Fu2 Xin Zhang2* Jing Tian5*
  • 1Harbin Medical University Cancer Hospital, Harbin, China
  • 2School of Health Management, Harbin Medical University, Harbin, China
  • 3Department of Medical Administration, Tianjin Cancer Hospital Airport Hospital, Tianjin, China
  • 4Department of Scientific Research, Peking University Shenzhen Hospital, Shenzhen, China
  • 5China National Center for Food Safety Risk Assessment, Beijing, China

Introduction: This study aims to investigate consumers' cognition, preference and willingness to pay (WTP) for celery with low pesticide residues, and to provide evidence from a consumer perspective for government food safety regulation.

Method: A survey was conducted on the population over the age of 18 in 6 provinces of Shanghai, Guangdong, Sichuan, Hunan, Hebei and Heilongjiang in China, in order to improve the representativeness of sample. The study carried out a survey of respondents' cognitive attitudes towards low-residue vegetables, and applied a discrete choice model (DCE) to simulate different scenarios of consumers buying vegetables. The DCE included residue level, appearance, taste, and price and finally constructed 24 choice sets. Respondents' preference for low-residue celery and WTP were analyzed using a mixed logit model.

Result: A total of 1292 respondents were surveyed. The model results showed that consumers had the highest positive preference for pesticide-free celery. For the Chinese consumers, price was the most important attribute, followed by the residue level, taste, and appearance. Consumer WTP for pesticide-free celery was11.17CNY/500g. Factors affecting consumer preferences were age, gender, income, education, whether they had children, or paid attention to pesticide residue in vegetables, and related reports of pesticide residue exceed the standard in vegetables.

Conclusion: Our study is more finely divided at the residue level, and the findings provide useful information for producers and policy makers.

1. Introduction

Food safety has become a global public health problem. Globally, around 600 million people fall ill each year after eating contaminated food, resulting in 420,000 deaths and a loss of 33 million Disabilty Adjust Life Years (DALYs), according to the (World Health Organization, 2015). Between 2003 and 2017, China's National Foodborne Disease Outbreak Surveillance System reported a total of 19,517 food-borne disease outbreaks, resulting in 1,457 deaths. In the 13,307 incidents with known causes, 4.8% were related to pesticides (Li et al., 2020). In the analysis of the causes of pesticide poisoning in food-borne diseases, excessive pesticide residues accounted for about 35.59% (Zhuang et al., 2021). The food safety supervision and sampling inspection announcement issued by the China State Administration for Market Regulation in 2020 shows that, among the unqualified items detected, the pesticide and veterinary drug residue exceeding the standard limit, accounted for 35.31% of the total unqualified samples (State Administration for Market Regulation of the People's Republic of China, 2020). Among the sampling results in the first half of 2021, agricultural veterinary drug residue exceeding the standard value, accounted for 37.29% of the total unqualified samples (State Administration for Market Regulation of the People's Republic of China, 2021).

From an economic point of view, the occurrence of food safety problems is mainly due to market failure caused by information asymmetry and externalities, resulting in insufficient food safety and effective supply, excessive harmful substances in food, and excessive risks to human health. In this case, if the government formulates food safety regulations to reduce the level of food-borne health risks, public welfare can be improved.

The use of pesticides has greatly improved crop yields, but also has an adverse effect on the environment and human health (World Health Organization., 2002). Policymakers in various countries have also begun to formulate relevant food safety regulations to control the impact of pesticides on the environment and health, such as adjusting the limit of pesticide residues in food or restricting the use of certain pesticides in crops (Van Ravenswaay and Hoehn, 1991; Buzby et al., 1995). The new regulations will inevitably have an impact on the main players in the food chain. For example, the reduction of available pesticides may lead to a decline in crop yield or quality, which in turn affects producer profits and product prices. From the perspective of cost-benefit analysis, although regulation reduces food safety risks, it also has a negative impact on social welfare (Jones et al., 1999).

In the process of improving food safety levels, policymakers need to weigh the benefits and costs of food safety regulation to ensure that regulations maximize the net benefit of food safety, that is, making the marginal benefit of safer food equal the marginal cost of achieving food safety goals. Cost-benefit analysis is considered an indispensable tool in policy design and decision-making (OECD, 2006). A cost-benefit analysis reflects the costs and benefits of all beneficiaries and losers, and provides a rational model for determining the net benefit optimum. At the same time, the cost-benefit analysis focuses on consumer preferences and is relatively fair. Consumer willingness to pay (WTP) for lower pesticide residue levels can be used as an indicator of food safety needs (Fu et al., 1999). The monetary value of changes in pesticide residue levels can reflect the preferences and perceptions of those exposed to risk. The lack of directly observable prices requires the use of non-market techniques to monetize individual preferences (Bateman et al., 2002). Non-market assessment techniques can be classified as revealed preference (RP) or stated preference (SP) techniques. Revealed preferences are based on information about actual decisions in the real market to derive the monetary value of risk changes, whereas, stated preferences require respondents to make decisions in a hypothetical market (Tago et al., 2014). Identifying consumers' WTP for low-residue vegetables is critical for both vegetable producers and policy makers. For producers, it is necessary to learn whether consumers are willing to pay for low-residue vegetables, whether producers can obtain additional benefits, and if the benefits can offset the costs incurred. Establishing vegetable preferences help in obtaining valuable information in terms of policy formulation process and cost control, and based on it, reasonable and effective food safety policies can be formulated.

A large number of studies in different countries show that consumers are willing to pay a premium for low-residue or organic food (Cecchini et al., 2018), and the main factors that affect consumers' WTP were divided into demographic and socioeconomic factors, and risk perceptions (Haghiri et al., 2009; Suhaimi et al., 2021). Weaver et al. assessed consumer WTP for pesticide-free tomatoes. The results show that most respondents were willing to pay 10% more (Weaver et al., 1992). Other scholars have also conducted research on pesticide-free tomatoes in Turkey, Tanzania, and other countries. The results show that most consumers willing to pay a small premium for residue-free tomatoes. Factors such as gender and education affect the WTP (Sedef et al., 2001; Bayramoglu and Göktolga, 2009; Alphonce and Alfnes, 2012). Studies on organic apples, Japanese mustard, and eggs done in France, Japan, and Italy also show that consumers are willing to pay a premium for low-residue products (Stéphan et al., 2012; Seo et al., 2019; Yeh et al., 2020). Demographic factors such as age, gender, nationality, education, family size, and income, as well as environmental concerns and emphasis on health, are important factors affecting WTP (Stefano and Michele, 2000; Morteza et al., 2007; Haghiri et al., 2009; Haghjou et al., 2013; Muhammad et al., 2015).

Scholars in China analyzed consumers' WTP for meat and vegetables that were low-residue, organic, or traceable agricultural products, by using conditional value evaluation and choice experiment. The results show that most consumers are willing to pay a certain price according to their preferences, and the premium for various products ranges from 42.11 to 335%. The main influencing factors were demographic factors, consumers' subjective knowledge, trust in government agencies, and awareness regarding the health and environmental impact of the use of pesticides (Dai et al., 2006; Zhou, 2006; Guo, 2013; Ge, 2018; Ma and Yao, 2018). At present, most of the studies on consumers' preference for low-residue vegetables in China are conducted in one city, which lacks overall representation.

Based on the above background, this study aims to evaluate Chinese consumers' cognition, preference and willingness to pay for low-residue vegetables. In this study, based on the per capita gross domestic product (GDP) level, we selected six provinces in China to conduct a survey, obtained consumers' concerns about pesticide residues in vegetables through a self-made questionnaire, and used discrete choice experiments to obtain consumer preferences and willingness to pay. Empirical analyses were conducted using mixed logistic models and explored preference heterogeneity among groups with different demographic characteristics. Information on consumer preferences and willingness to pay is crucial for analyzing the effects of policy implementation. This study estimates the benefits of food safety risk reduction, combined with the cost of regulation, to determine whether regulation can generate social benefits, thus providing a reference for government management decisions.

2. Materials and methods

2.1. Study design

According to the per capita GDP of each province in 2020 released by the National Bureau of Statistics of China, we divided the GDP per capita of 31 provinces in mainland China into three levels, high, middle and low, and considering the geographical location of the provinces, selected Shanghai, Guangdong, Sichuan, Hunan, Hebei and Heilongjiang for investigation. The geographical location and per capita GDP of the provinces are shown in Supplementary Table S1. The survey was officially conducted in July 2021. A self-designed questionnaire was used to collect data on Wenjuanxing, an online research platform. The questionnaire included three parts: basic personal information, cognitive situation and discrete choice experiment. Basic personal information includes demographic information and personal health status; cognition includes whether to pay attention to pesticide residues, sources of information, frequency of active acquisition, etc. In order to check the fluency and readability of the questionnaire, grasp the survey time, and improve the response efficiency of the respondents, a pre-survey was conducted before the formal experiment was carried out. The pre-survey was conducted by face-to-face survey. The respondents were instructed on site to complete the questionnaire, and feedback the problems encountered in questionnaire filling. The pre-survey was conducted in Harbin, Heilongjiang Province, with a sample size of about 15 people. In the formal survey, in order to improve the enthusiasm of the respondents, each respondent is given a reward of 5 yuan. In order to better control the quality of the questionnaires, the questionnaires whose filling time was too short were deleted, and the questionnaires were logically checked. Prior to data analysis, we removed respondents with obvious protest responses, that is, questionnaires in which all choose out-opt or one side of the alternative items (Sufyan et al., 2019b).

2.2. Choice sets design

There are two main reasons for taking celery as the research object in this study: First. Celery is commonly consumed by the chinese people. In recent years, in the sampling monitoring of commercially available celery, the detection rate of pesticides is about 70%, and the exceeding rate is between 16.4 and 33.3%. There are relatively serious pesticide residues (Wu et al., 2010; Fang et al., 2015; Xu et al., 2018; Qin et al., 2020; Zhang and Li, 2020). The main purpose of the research is to determine consumers' willingness to pay for celery with different pesticide residue levels, therefore, the pesticide residue level is included as the main attribute. The price attribute was included in the study as a tool to measure willingness to pay (Zhang and Jakku, 2020). At the same time, the taste (Malone and Lusk, 2017) and appearance (Alfnes et al., 2006) attributes that consumers are concerned about are included. Among these four attributes, the residual level and price can be quantified, so we divide the level more finely. The prices of common celery and organic celery in the market are confirmed as the upper and lower limits of price attributes. Taste and appearance are relatively subjective and difficult to quantify, so the taste attribute divided into three levels: superior, equal and inferior to ordinary celery. The appearance attribute are divided into three levels: marked, mild and scar-free (see Table 1).


Table 1. Attributes and levels used in the discrete choice experiment (DCE).

After determining the attributes and levels, next step was to design the choice set. The choice and quantity were closely related to the number of attributes and levels. There were four attributes in this study. Two attributes were at five levels and two attributes were at three levels. If a full factorial design is used, (5 × 3 × 3 × 5) 2 = 50,625 choice sets will be generated. Therefore, the D-efficiency program of stata16.0 is used to generate choice sets. Each choice set includes an out-opt. A total of 24 choice sets were generated, which were divided into four versions to reduce the respondent's response burden and improve response efficiency (Reed Johnson et al., 2013). Each version sets a repeated choice set based on six choice sets. An example of a choice set is given in Table 2.


Table 2. Example of discrete experiment choice set.

2.3. Theoretical framework and data analysis

This study uses discrete choice experiments to simulate consumer purchases of vegetables. In recent years, discrete choice experiments have been widely used as an emerging preference measurement tool in the fields of food, environment, hygiene, and transportation (Flügel et al., 2015; Barrowclough and Alwang, 2018; Thøgersen et al., 2018; Livingstone et al., 2020; Phillips et al., 2021). Discrete choice applies to a range of choice scenarios, where an individual chooses one from a set of alternatives, and the alternative is represented by a set of attributes, thereby revealing the important attributes/levels that influence the individual's choice (Nakatani et al., 2014). The basis of the discrete choice model is the random utility theory and the theory of characteristic value. The theory of characteristic value shows that the utility consumers obtain from a product is a function of product attributes (Lancaster, 1966), and under budget constraints, consumers rationally choose products to maximize utility. Unjt represents the utility that the decision maker n obtains from the consumption of vegetable j under the choice scenario t, specified as a function of price Pnjt and other non-monetary attributes xnjt (Hole and Kolstad, 2012). The utility model is as follows:

Unjt=αnpnjt+βnxnjt+enjt    (1)

In the formula, αn and βn are random among the decision makers, assuming that enjt is an independent identically distributed (IID) type I extreme value distribution (EV I) distribution. The variance of enjt is different for different decision makers: Var(enjt)=kn2(π2/6), where kn is the scale parameter of decision maker n.

Dividing the utility function (1) by kn does not affect the behavior (Train and Weeks, 2005), but produces a new error term, which obeys the IID extreme value distribution, and the variance is equal to π26:

Unjt=-(αn/kn)pnjt+(βn/kn)xnjt+εnjt    (2)

The utility coefficient is defined as λn = αn/kn and cn = βn/kn, and the utility is written as:

Unjt=-λnpnjt+cnxnjt+εnjt    (3)

Equation (3) is called the utility model in the preference space. The willingness to pay for an attribute is the ratio of the attribute coefficient to the price coefficient: wn = cnn. Using this definition, the utility function can be rewritten as:

Unjt=-λnpnjt+(λnwn)xnjt+εnjt    (4)

Equation (4) is called the utility model in the willingness to pay space. Under this parameterization, changes in willingness to pay (independent of scale) and changes in price coefficients (including scale) are distinguished.

In discrete choice experiments, consumers are usually faced with multiple combined scenarios with different attributes and are asked to make a choice among these scenarios. According to the results of the choice, we can simulate and estimate the consumers' preference parameters for these characteristics or attributes by establishing a certain measurement model, thereby explaining the consumer's choice behavior. It is important to consider these individual preferences and heterogeneity in the modeling process. The mix logit model is one of the methods to explain the heterogeneity of the interviewees' preferences. Alternative specific constant (asc) terms are set to analyze intrinsic, property-independent preferences. We set the constant term for alternative1 and alternative2 to 1, and the out-opt to 0 (Si et al., 2019; Sufyan et al., 2019a). It allows the parameters to vary randomly among the individuals, and is characterized by the heterogeneity as a continuous function of the parameters (McFadden and Train, 2000). The probability that individual n chooses alternative j from the choice set sequence I is:

Pnj=exp(βsXnj)j=1Jexp(βsXnj)    (5)

The program was developed by Hole in stata16.0 for data analysis (Hole, 2007). This study uses a mixed logit model to estimate the main effects, and an interaction term estimation model to assess whether there were potential differences in preferences among groups with different sociodemographic characteristics, including gender, age, income, education, degree, whether they had children, and the degree of concern about pesticide residue in vegetables. The interaction terms are fixed effect parameters, and the main attribute coefficients are random coefficients. For a list of variables, see Supplementary Tables S2, S3.

3. Results and discussion

3.1. Sample characters

A total of 1,307 respondents were investigated in this study, and 1,292 samples were finally retained after the questionnaire of obvious protest response and logic problem was deleted. 31.81% were registered in rural areas. The sample sizes of Shanghai, Guangdong, Hunan, Sichuan, Hebei and Heilongjiang were 239, 239, 204, 212, 203, and 195„ respectively. Women accounted for 48.76% of all respondents, the average age was 31.2 years, 79.95% had a job, 72.83% had a college degree or above, 62% were married, 81.27% had an average monthly household income of 5,000 yuan and above, and 81.42% of households spent 1,000 yuan and above on food every month. Most respondents believed that they were in good health. 61.38% of the respondents had children under 15. The basic information of the sample population is shown in Table 3.


Table 3. Socio-demographic characteristics of the sample.

3.2. Consumer cognition

The survey results showed that 80.19% of consumers were worried about pesticide residue in vegetables. Television, the Internet, and food safety agency publicity are the main channels through which consumers obtained information on pesticide residue in vegetables. Most consumers learn relevant information 1–3 times a week, 86.07% would screen the information, 79.72% would continue to pay attention to information, and 82.82% of consumers were concerned about reports of excessive pesticide residue in vegetables. 55.81% of the people strongly agreed or agreed that they could not identify the levels of pesticide residues in vegetables, and 41.56% of the people strongly agreed or agreed that they lacked the relevant knowledge about the health effects of pesticide residues. The main impact of these reports was reduced purchase or more cautious purchase (see Table 4).


Table 4. Consumers' cognition of vegetables with low residue.

The results of our survey showed that consumers are concerned about pesticide residues and lack of relevant knowledge and information. At the same time, studies have shown that some consumers do not understand the health hazards of pesticide residues and cannot identify the pesticide residue levels of vegetables sold in the market (Vidogbéna et al., 2015). Therefore, it's necessary to fully inform consumers the residue level through disclosing information or puting certification labels on low vegetables.

3.3. Preference weights and relative importance of attributes and levels

Table 5 reports the results of the mixed logit model in preference space. The coefficient of Mean asc is significantly positive, indicating that consumers are more inclined to choose low-residue vegetables than to maintain the status quo. The coefficient of SD asc is significant, indicating that there is significant heterogeneity in consumer preferences. Whether in urban or rural areas, the mean coefficients of all attributes are statistically significant and are expected signs, indicating that all attributes included in DCE have an impact on low-residue vegetable purchasing decisions. From the standard deviation of the regression coefficients, it can be seen that among urban and rural consumers, residue4, residue5, taste3 and appe3 have a significant impact on consumers' vegetable choices, and different respondents have different preferences for these attributes. Negative price coefficients indicate that both rural and urban consumers prefer lower-priced vegetables.


Table 5. Mixed logit model estimates result (main effects).

The preference weights are shown in Figure 1. The colored sphere represents the mean value of the preference coefficient, and the gray sphere represents the 95% confidence interval. In both urban or rural areas, consumers had the highest positive preference for celery without pesticide residues (Rural: β = 2.27, SE = 0.12, p = 0.000; Urban: β = 2.31, SE = 0.21, p = 0.000). The preference weight of cities is slightly higher than that of rural areas. Price is a negative preference, and the absolute value of the preference weight in the countryside is slightly higher than that in the city (Rural: β = −0.28, SE=0.02, p = 0.000; Urban: β = −0.24, SE=0.01, p = 0.000).


Figure 1. Rural and urban consumer preference weights with a 95% CI.

Figure 2 shows the relative importance scores of attributes of rural and urban consumers. The relative importance score of each attribute was determined by the ratio of the maximum utility to the total utility of the attribute. Among rural consumers, the relative importance score of price attributes was 83.15%, followed by residue attributes at 9.56%, taste attributes at 5.43%, and appearance attributes at 1.82%. The relative importance score of prices among urban consumers was 84.36%, followed by 9.01, 4.27, and 2.36% for residue, taste and appearance„ respectively.


Figure 2. Rural and urban consumer relative importance score of the attributes.

In general, among the four attributes, the relative importance of price is the highest, and Chinese consumers' vegetable purchase preference is affected by consumption level. Consumers preferred low residue attributes over taste and appearance, and there were similarities in preferences between urban and rural residents. The highest positive preference for no pesticide residue in the preference weight also shows that reducing pesticide residue levels is more important to consumers relative to taste and appearance.

3.4. Preference heterogeneity

The heterogeneity analysis of the preferences in different regions is presented in Table 6. Among rural consumers, gender, income, education level, and concerns about pesticide residue in vegetables have an impact on purchasing decisions for low-residue vegetables. Compared with consumers who are not concerned about pesticide residue, concerned consumers place more importance on residue levels, but less on appearance. The higher the education level, the greater the emphasis on appearance. The higher the income level, the higher the emphasis on taste.


Table 6. Results of the preference heterogeneity analysis.

Age, gender, income, education level, whether the consumers had children, whether they paid attention to pesticide residue in vegetables, and whether they paid attention to reports of pesticide residue exceeding the standard in vegetables plays an important role in the urban consumers' decision to purchase low-residue vegetables. Compared with female consumers, male consumers are less concerned about residue levels and taste but are more concerned about price. Consumers with higher income place more emphasis on the residual levels. Consumers who paid attention to reports are more interested in residual levels and prices, as compared to those who do not. Older consumers are less concerned with appearance and taste, but are more concerned with price. Consumers who are concerned about pesticide residues in vegetables pay less attention to appearance.

In order to better analyze the heterogeneity of preferences among provinces, we added the province interaction term, and the specific results are shown in Supplementary Table S4. Among the interaction terms between province and each attribute level, only the interaction term between price attribute and province is significant, indicating that consumers in different provinces have different preferences for price attribute.

Heterogeneity analysis revealed the effects of demographic factors and risk concerns on preferences. Consistent with previous studies, our study confirmed that income level has a significant impact on consumers' vegetable purchasing decisions, and people with higher incomes tend to buy vegetables with lower pesticide residues (Dai et al., 2006; Haghjou et al., 2013; Muhammad et al., 2015). This may be due to a higher household income and higher affordability of consumers, the more aware they are of the potential health hazards of pesticide use (Haghiri et al., 2009). Consumers with higher risk concerns place more emphasis on the attribute of residue level, and conversely, they place less emphasis on appearance (Stefano and Michele, 2000). We also found that older consumers are less concerned with residue level, appearance, and taste, and are more concerned with price.

3.5. Model estimates and WTP calculation

Table 7 shows the willingness to pay for each attribute level. The result of willingness to pay comes from a mix logit model with no interaction. Both urban and rural consumers have the highest WTP for celery without pesticide residue, with the prices 9.82CNY/500 g in rural areas and 11.72CNY/500 g in urban areas. The rural consumers' WTP for low-residue celery with pesticide residues of 60, 40, and 20% of normal celery is 3.43, 5.73, and 9.23 CNY/500 g„ respectively, the WTP for taste attribute is 2.65 and 5.46CNY/500 g„ respectively, the WTP for appearance attribute is 2.23 and 2.78CNY/500 g„ respectively. The urban consumers' WTP for low-residue celery with pesticide residues 60, 40, and 20 of ordinary celery is 3.97, 6.95, and 10.32 CNY/500 g, respectively, the WTP for taste attribute is 2.88 and 5.23 CNY/500 g, respectively, the WTP for appearance attribute is 2.32 and 3.64 CNY/500 g, respectively.


Table 7. Mixed logit model in WTP space.

The results of WTP analysis showed that, consumers' WTP for pesticide-free celery is slightly higher than that of low-residue celery with 20% residues of ordinary celery, and much higher than that of low-residue celery with residue levels of 40% and 60% of ordinary celery. When the pesticide residue level is high, the difference of consumers' WTP between different pesticide residue levels was large. The difference was smaller when pesticide residue levels were relatively low. That is, when consumers perceive low residue level and low food safety risk, they are not willing to pay more money to further improve food safety level. It is acceptable for all consumers to pay a higher price to meet their own needs, there is a potential demand for higher levels of food safety regulations, and consumers want to obtain less residual food. The low-residue vegetable market has great potential for development in China (Certification Accreditation Administration of the People's Republic of China, 2019).

The analysis results from the urban and rural subgroups show that urban and rural consumers have significant differences in basic information such as age, education level, marriage, and average monthly household income, and these differences are also reflected in consumers' WTP. Rural consumers pay a premium of about 327% for pesticide-free celery and urban consumers about 390%. The WTP of urban consumers for each residue level is higher than rural consumers, and rural consumers are willing to pay less money to improve food safety when pesticide residue levels are low. In the two attributes of taste and appearance, both urban and rural consumers are willing to pay for celery with a better taste; urban consumers' WTP for perfect appearance is 3.636CNY/500 g (a premium of 121%), while rural consumers are not willing to pay a premium for that.

Based on data from six provinces, this study analyzed Chinese consumers' concern about pesticide residues in vegetables, and evaluated Chinese consumers' preference and WTP for low-residue vegetables using discrete choice experiment. The results of WTP can be used to indirectly measure the net benefit of consumers avoiding important sources of health risks as a result of improved food safety. Compared with previous studies on Chinese consumers, the data in this study are more representative. In previous studies on consumers' WTP for low-residue foods, the residue-related attributes are usually set as whether organic or different certification labels, and the residue attributes are not subdivided into different levels (Sakagami and Haas, 2012; Wang et al., 2019; Carzedda et al., 2021; Van Loo et al., 2021). In actual food safety regulation, there is not a single “yes” or “no” issue. In this study, residue attributes are divided into five levels, which can provide more detailed information for food safety regulation evaluation and decision-making.

This study had certain limitations. First, DCE is a declarative preference, and the choice of scenario is based on assumptions, which does not necessarily reflect the consumer choices in real scenarios. Based on literature research and pre-survey, we screen the important attributes that affect consumer choice and set up a more realistic choice scenario. Second, due to the complexity of real market decision-making, the attributes in the research cannot contain all the attributes contained in the commodity in reality, therefore, the four main attributes of residue level, appearance taste and price are included in this study.

4. Conclusion

Consumer preference is an important market information andis crucial for the government to formulate relevant regulatory policies. Previous studies have shown that consumers are willing to pay a premium for residue-free or organic vegetables. This study further refined the residue levels and analyzed the willingness to pay of consumers with different residue levels. The results show that consumers have a strong positive preference and higher willingness to pay for low-residue vegetables. This indicates that consumers are willing to pay a certain cost for safe vegetables with less residue, and there is a potential demand for higher food safety levels. It can also provide market incentives for producers to improve their production methods and provide better and safer food, thereby influencing the development of relevant markets. And when the residue level is 60%-40% of common vegetables, consumers are willing to pay a relatively high premium to reduce the residue level. When the residue level is 20% of common vegetables, that is, the food safety risk is relatively low, consumers are willing to pay a relatively low premium to further improve the food safety level. Our results can provide data support for the cost-benefit analysis of the revision of pesticide residue standard system.

The results of heterogeneity analysis showed that demographic factors such as gender, age and income, and whether to pay attention to pesticide residues and reports of pesticide residues exceeding the standard in vegetables have a significant impact on Chinese consumers' preference for low-residue vegetable attributes, which suggests that we can develop market segments based on demographic characteristics to better promote the development of organic vegetable market.

Finally, we suggest that future research should analyze the consistency between real market data and hypothetical markets. Applying willingness-to-pay data to a cost-benefit analysis of food safety regulation revisions, conduct ex ante and ex post regulatory impact assessments, to estimate welfare changes under different regulatory scenarios and provide decision support for the development and adjustment of food safety regulations.

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions

Conceptualization: JW, LZ, ZN, WW, GL, WF, XZ, and JT. Data curation: LZ and ZN. Formal analysis: JW. Investigation, project administration, and writing—original draft: JW, ZN, and WW. Methodology, supervision, and writing—review and editing: XZ and JT. Validation: LZ. Visualization: WW and WF. All authors contributed to the article and approved the submitted version.


This research was funded by National Key R&D Program of China (2019YFC1605201 and 2019YFC1605200).

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at:


Alfnes, F., Guttormsen, A. G., Steine, G., and Kolstad, K. (2006). Consumers' Willingness to Pay for the Color of Salmon: A Choice Experiment with Real Economic Incentives. Am. J. Agric. Econ. 88, 1050–1061. doi: 10.1111/j.1467-8276.2006.00915.x

CrossRef Full Text | Google Scholar

Alphonce, R., and Alfnes, F. (2012). Consumer willingness to pay for food safety in Tanzania: an incentive-aligned conjoint analysis. Int. J. Consum. Stud. 36, 394–400. doi: 10.1111/j.1470-6431.2011.01067.x

CrossRef Full Text | Google Scholar

Barrowclough, M. J., and Alwang, J. (2018). Conservation agriculture in Ecuador's highlands: a discrete choice experiment. Environ. Dev. Sustain. 20, 2681–2705. doi: 10.1007/s10668-017-0011-0

CrossRef Full Text | Google Scholar

Bateman, I. J., Carson, R. T., Day, B., Hanemann, M., Hanley, N., Hett, T., et al. (2002). Economic Valuation with Stated Preference Techniques: A Manual. Edward Elgar: Cheltenham, UK.

Google Scholar

Bayramoglu, Z., and Göktolga, Z. (2009). “Consumer willingness to pay for pesticide free tomatoes in Turkey,” in First International Syposium on Sustainable Development, Sarajevo, Bosnia and Herzegovina. Available online at: (accessed on 9–10 June 2009).

Google Scholar

Buzby, J., Ready, R., and Skees, J. (1995). Contingent Valuation in Food Policy Analysis: A Case Study of a Pesticide-Residue Risk Reduction. J. Agric. Appl. Econ. 27, 613–625. doi: 10.1017/S1074070800028637

CrossRef Full Text | Google Scholar

Carzedda, M., Gallenti, G., Troiano, S., Cosmina, M., Marangon, F., de Luca, G., et al. (2021). Consumer preferences for origin and organic attributes of extra virgin olive oil: a choice experiment in the italian market. Foods (Basel, Switzerland), 10, 994. doi: 10.3390/foods10050994

PubMed Abstract | CrossRef Full Text | Google Scholar

Cecchini, L., Torquati, B., and Chiorri, M. (2018). Sustainable agri-food products: A review of consumer preference studies through experimental economics. Agric. Econ. 12, 554–565. doi: 10.17221/272/2017-AGRICECON

CrossRef Full Text | Google Scholar

Certification Accreditation Administration of the People's Republic of China (2019). China Organic Product Certification and Organic Industry Development Report. Available online at:

Google Scholar

Dai, Y. C., Zhu, B., and Ying, R. Y. (2006). Consumers' willingness to choose food safety—a case study of organic vegetable consumption in Nanjing. J. Nanjing Agric. University (Social Sciences Edition), 1, 47–52. doi: 10.3969/j.issn.1671-7465.2006.01.009

CrossRef Full Text | Google Scholar

Fang, L., Zhang, S., Chen, Z., Du, H., Zhu, Q., Dong, Z., et al. (2015). Risk assessment of pesticide residues in dietary intake of celery in China. Regulat. Toxicol. Pharmacol. 73, 578–586. doi: 10.1016/j.yrtph.2015.08.009

PubMed Abstract | CrossRef Full Text | Google Scholar

Flügel, S., Elvik, R., Veisten, K., Rizzi, L., Meyer, S., Ramjerdi, F., et al. (2015). Asymmetric preferences for road safety: Evidence from a stated choice experiment among car drivers. Transport. Res. Part F Traffic Psychol. Behav. 31, 112–123. doi: 10.1016/j.trf.2015.04.001

CrossRef Full Text | Google Scholar

Fu, T. T., Liu, J. T., and Hammitt, J. K. (1999). Consumer willingness to pay for low-pesticide fresh produce in Taiwan. J. Agric. Econ., 50, 220–233. doi: 10.1111/j.1477-9552.1999.tb00809.x

CrossRef Full Text | Google Scholar

Ge, J. Y. (2018). Consumer Decision Behavior Based on Willingness to Pay: An Agricultural Product Safety Attribute Perspective. Wuxi, China: JiangNan University.

Google Scholar

Guo, Y. Y. (2013). Empirical research of Shanghai consumer s' willingness to pay for food safety–Pork consumption as an example. J. Southwest Minzu Univer. 039, 605–610. doi: 10.3969/j.issn.1003-4271.2013.04.27

CrossRef Full Text | Google Scholar

Haghiri, M., Hobbs, J. E., and Mcnamara, M. L. (2009). Assessing consumer preferences for organically grown fresh fruit and vegetables in eastern new brunswick. Int. Food Agribus. Manage. Rev. 12, 81–99.

Google Scholar

Haghjou, M., Hayati, B., Pishbahar, E., Mohammadrezaei, R., and Dashti, G. (2013). Factors Affecting Consumers' Potential Willingness to Pay for Organic Food Products in Iran: Case Study of Tabriz. J. Agric. Sci. Technol. 15, 191–202.

Google Scholar

Hole, A. R. (2007). Fitting mixed logit models by using maximum simulated likelihood. Stata J. 7, 388–401. doi: 10.1177/1536867X0700700306

PubMed Abstract | CrossRef Full Text | Google Scholar

Hole, A. R., and Kolstad, J. R. (2012). Mixed logit estimation of willingness to pay distributions: a comparison of models in preference and WTP space using data from a health-related choice experiment. Empir. Econ. 42, 445–469. doi: 10.1007/s00181-011-0500-1

CrossRef Full Text | Google Scholar

Jones, W., Bureau, J. C., and Marette, S. (1999). Food safety: Protection or protectionism? OECD Observer 1999, 27–30.

Google Scholar

Lancaster, K. J. (1966). A new approach to consumer theory. J. Pol. Econ. 74, 132–157. doi: 10.1086/259131

CrossRef Full Text | Google Scholar

Li, W. W., Pires, S., Liu, Z. T., Ma, X. C., Liang, J. J., Jiang, Y. Y., et al. (2020). Surveillance of foodborne disease outbreaks in China, 2003–2017. Food Control 118, 107359. doi: 10.1016/j.foodcont.2020.107359

CrossRef Full Text | Google Scholar

Livingstone, K. M., Lamb, K. E., Abbott, G., Worsley, T., and McNaughton, S. A. (2020). Ranking of meal preferences and interactions with demographic characteristics: a discrete choice experiment in young adults. Int. J. Behav. Nutr. Phys. Activity 17, 157. doi: 10.1186/s12966-020-01059-7

PubMed Abstract | CrossRef Full Text | Google Scholar

Ma, R. Y., and Yao, Z. Z. (2018). Study on Consumers' Preference for Vegetable Attributes and Willingness to pay—A consumer survey based on choice test method. Jiangsu Agric. Sci. 46, 319–323. doi: 10.15889/j.issn.1002-1302.2018.10.076

CrossRef Full Text | Google Scholar

Malone, T., and Lusk, J. L. (2017). Taste trumps health and safety: incorporating consumer perceptions into a discrete choice experiment for meat. J. Agric. Appl. Econ. 49, 139–157. doi: 10.1017/aae.2016.33

CrossRef Full Text | Google Scholar

McFadden, D., and Train, K. (2000). Mixed MNL models for discrete response. J. Appl. Econ. 15, 447–470. doi: 10.1002/1099-1255(200009/10)15:5<447::AID-JAE570>3.0.CO;2-1

PubMed Abstract | CrossRef Full Text | Google Scholar

Morteza, H., Meaghan, L., and McNamara. (2007). Predicting consumers' acceptability of pesticide-free fresh produce in Canada's maritime provinces. J. Int. Food Agribus. Market. 19, 45-59. doi: 10.1300/J047v19n04_04

CrossRef Full Text | Google Scholar

Muhammad, S., Fathelrahman, E., and Ullah, R. U. T. (2015). Factors affecting consumers' willingness to pay for certififed organic food products in United Arab Emirates. J. Food Distrib. Res. 46, 1–9. doi: 10.22004/ag.econ.199045

CrossRef Full Text | Google Scholar

OECD (2006). Cost-Benefit Analysis and the Environment: Recent Developments. Paris: OECD Publishing.

Google Scholar

Phillips, E. A., Himmler, S. F., and Schreyögg, J. (2021). Preferences for e-mental health interventions in germany: a discrete choice experiment. Value Health J. Int. Soc. PharmacoEcon. Outcomes Res. 24, 421–430. doi: 10.1016/j.jval.2020.09.018

PubMed Abstract | CrossRef Full Text | Google Scholar

Qin, G. F., He, F. R., Zhang, W., Yang, B. X., Zuo, B., Li, Y. B., et al. (2020). Investigation on 11 fungicides residues in 130 vegetables in Shaanxi province in 2018. Modern Prev. Med. 47, 604–607.

Google Scholar

Reed Johnson, F., Lancsar, E., Marshall, D., Kilambi, V., Mühlbacher, A., Regier, D. A., et al. (2013). Constructing experimental designs for discrete-choice experiments: report of the ISPOR Conjoint Analysis Experimental Design Good Research Practices Task Force. Value Health J. Int. Soc. PharmacoEcon. Outcomes Res. 16, 3–13. doi: 10.1016/j.jval.2012.08.2223

PubMed Abstract | CrossRef Full Text | Google Scholar

Sakagami, M., and Haas, R. (2012). Consumer preferences for organic products in austria using stated preference methods. Curr. Nutr. Food Sci. 8, 122–125. doi: 10.2174/157340112800840844

CrossRef Full Text | Google Scholar

Sedef, A., BÜLent, M., and Canan, A. (2001). Consumer willingness to pay for food safety labels in urban Turkey. J. Int. Food Agribus. Market. 12, 91–107. doi: 10.1300/J047v12n01_05

CrossRef Full Text | Google Scholar

Seo, Y., Someya, Y., and Dowaki, K. (2019). Environmental impacts and consumer preference for sustainably cultivated Japanese mustard spinach, komatsuna. J. Environ Manage. 231, 364–369. doi: 10.1016/j.jenvman.2018.10.077

PubMed Abstract | CrossRef Full Text | Google Scholar

Si, L., Tu, L. D., Xie, Y., Andrew, J. P., Gu, Y. Y., Zhen., et al. (2019). Chinese patients' preference for pharmaceutical treatments of osteoporosis: a discrete choice experiment. Arch Osteoporos. 14, 85. doi: 10.1007/s11657-019-0624-z

PubMed Abstract | CrossRef Full Text | Google Scholar

State Administration for Market Regulation of the People's Republic of China (2020). Circular of the State Administration for Market Regulation on spot Inspection of Food safety supervision by Market Supervision departments in 2020. Available online at: (accessed on January 10, 2022).

Google Scholar

State Administration for Market Regulation of the People's Republic of China (2021). Circular of the State Administration for Market Regulation on spot Inspection of Food safety supervision by market supervision departments in the first half of 2021. Available online at: (accessed on January 10, 2022).

Google Scholar

Stefano, B., and Michele, N. (2000). Consumer willingness to pay for pesticide-free fresh fruit and vegetables in Italy. Int. Food Agribus. Manage. Rev. 3, 297–310. doi: 10.1016/S1096-7508(01)00049-0

CrossRef Full Text | Google Scholar

Stéphan, M., Antoine, M., and Guy, M. (2012). Consumers' willingness to pay for eco-friendly apples under different labels: Evidences from a lab experiment. Food Policy, 37, 151–161. doi: 10.1016/j.foodpol.2011.12.001

CrossRef Full Text | Google Scholar

Sufyan, U. K., Imran, K., Zhao, M. J., Arshad, A. K., and Muhammad, A. S. A. (2019b). Valuation of ecosystem services using choice experiment with preference heterogeneity: A benefit transfer analysis across inland river basin. Sci. Total Environ. 679, 126–135. doi: 10.1016/j.scitotenv.2019.05.049

PubMed Abstract | CrossRef Full Text | Google Scholar

Sufyan, U. K., Imran, K., Zhao, M. J., Hsiaoping, C., Lu, Q., Muhammad, A. S. A., et al. (2019a). Spatial heterogeneity of ecosystem services: a distance decay approach to quantify willingness to pay for improvements in Heihe River Basin ecosystems. Environ. Sci. Pollut. Res. 26, 25247–25261. doi: 10.1007/s11356-019-05691-0

PubMed Abstract | CrossRef Full Text | Google Scholar

Suhaimi, A., Othman, A., Sundram, V., and Ghazali, A. (2021). Consumers' purchase decision based on intrinsic and extrinsic factors related to food safety issues: A review. IOP Conf. Ser. Earth Environ. Sci. 756, 012010. doi: 10.1088/1755-1315/756/1/012010

CrossRef Full Text | Google Scholar

Tago, D., Andersson, H., and Treich, N. (2014). Pesticides and health: a review of evidence on health effects, valuation of risks, and benefit-cost analysis. Adv. Health Econ. Health Serv. Res. 24, 203–295. doi: 10.1108/S0731-219920140000024006

PubMed Abstract | CrossRef Full Text | Google Scholar

Thøgersen, J., Pedersen, S., and Aschemann-Witzel, J. (2018). The impact of organic certification and country of origin on consumer food choice in developed and emerging economies. Food Qual. Pref. 72, 10–30. doi: 10.1016/j.foodqual.2018.09.003

CrossRef Full Text | Google Scholar

Train, K., and Weeks, M. (2005). “Discrete choice models in preference space and willingness-to-pay space,” in Applications of Simulation Methods in Environmental and Resource Economics. The Economics of Non-Market Goods and Resources, eds. R. Scarpa, A., Alberini (Berlin/Heidelberg, Germany: Springer, Dordrecht), 1–16

PubMed Abstract | Google Scholar

Van Loo, E. J., Minnens, F., and Verbeke, W. (2021). Consumer preferences for private label brand vs. national brand organic juice and eggs: a latent class approach. Sustainability, 13, 7028. doi: 10.3390/su13137028

CrossRef Full Text | Google Scholar

Van Ravenswaay, E. O., and Hoehn, J. P. (1991). “The impact of health risk information on food demand: a case study of alar and apples,” in Economics of Food Safety, eds. J. A. Caswell (Berlin/Heidelberg, Germany: Springer, Dordrecht), 155–174.

Google Scholar

Vidogbéna, F., Adégbidi, A., Tossou, R., Assogba-Komlan, F., Martin, T., Ngouajio, M., et al. (2015). consumers' willingness to pay for cabbage with minimized pesticide residues in Southern Benin. Environments 2, 449–470. doi: 10.3390/environments2040449

CrossRef Full Text | Google Scholar

Wang, E. P., Gao, Z. F., Heng, Y., and Shi, L. J. (2019). Chinese consumers'preferences for food quality test/measurement indicators and cues of milk powder: a case of Zhengzhou, China. Food Policy 89, 101791. doi: 10.1016/j.foodpol.2019.101791

CrossRef Full Text | Google Scholar

Weaver, R. D., Evans, D. J., and Luloff, A. E. (1992). Pesticide use in tomato production: consumer concerns and willingness-to-pay. Agribusiness 8, 131–142. doi: 10.1002/1520-6297(199203)8:2<131::AID-AGR2720080205>3.0.CO;2-W

CrossRef Full Text | Google Scholar

World Health Organization (2015). WHO Estimates of the Global Burden of Foodborne Diseases: Foodborne Disease Burden Epidemiology Reference Group 2007-2015. Geneva, Switzerland: World Health Organization. Available online at:

Google Scholar

World Health Organization. (2002). Reducing and Eliminating the Use of Persistent Organic Pesticides: Guidance on Alternative Strategies for Sustainable Pest and Vector Management. Available online at: (accessed on January 10, 2022).

Google Scholar

Wu, P. W., Hou, R. L., and Ma, Y. J. (2010). Analysis on the reasons for excessive pesticide residues in celery and research on countermeasures. Pest. Sci. Adm. 31, 35–37. doi: 10.3969/j.issn.1002-5480.2010.05.010

CrossRef Full Text | Google Scholar

Xu, X., Li, L., Huang, X., Lin, H., Liu, G., Xu, D., et al. (2018). Survey of four groups of cumulative pesticide residues in 12 vegetables in 15 Provinces in China. J. Food Protect. 81, 377–385. doi: 10.4315/0362-028X.JFP-17-197

PubMed Abstract | CrossRef Full Text | Google Scholar

Yeh, C. H., Menozzi, D., and Török, Á. (2020). Eliciting egg consumer preferences for organic labels and omega 3 claims in italy and hungary. Foods, 9, 1212. doi: 10.3390/foods9091212

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhang, A., and Jakku, E. (2020). Australian consumers' preferences for food attributes: a latent profile analysis. Foods, 10, 56. doi: 10.3390/foods10010056

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhang, Q., and Li, X. (2020). Analysis of pesticide residues in celery in Changzhi area. Agric. Technol. 40, 57–58. doi: 10.19754/j.nyyjs.20200615018

CrossRef Full Text

Zhou, Y. H. (2006). Consumer willingness to pay for food safety in Jiangsu Province China: a case study of reduced pesticide residues B. chinensis. China Econ. Q. 5, 1319–1342.

Zhuang, Z., Guo, Y. C., Yang, S. X., Geng, X. F., Li, W. W., Li, J. J., et al. (2021). Analysis of food-borne pesticide poisoning incidents in China from 2002 to 2017. Chin. J. Food Hyg. 33, 373–378. doi: 10.13590/j.cjfh.2021.03.024

CrossRef Full Text

Keywords: willingness to pay, discrete choice experiment, consumer preference, organic food, low-pesticide residue vegetables, Chinese consumer

Citation: Wang J, Zhou L, Ni Z, Wu W, Liu G, Fu W, Zhang X and Tian J (2022) Consumer preference and willingness to pay for low-residue vegetables: Evidence from discrete choice experiments in China. Front. Sustain. Food Syst. 6:1019372. doi: 10.3389/fsufs.2022.1019372

Received: 15 August 2022; Accepted: 30 November 2022;
Published: 16 December 2022.

Edited by:

Delia Grace, University of Greenwich, United Kingdom

Reviewed by:

Arshad Ahmad Khan, Northwest A&F University, China
Yingjun Xu, Qufu Normal University, China

Copyright © 2022 Wang, Zhou, Ni, Wu, Liu, Fu, Zhang and Tian. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Xin Zhang, yes; Jing Tian, yes