How product attributes and consumer attitudes affect purchase prices of japonica rice in China

This study focuses on japonica rice (Oryza sativa L. ssp. japonica) and examines how product attributes, consumer attitudes, consumer demographics, and regional differences influence different categories of the daily purchase prices of japonica rice in China. Online questionnaires were distributed to 492 respondents from the inland, coastal, and northern regions of China. The study applied an ordered logit regression and a multinomial logit regression to analyze the survey data and understand the differences among various categories based on purchase prices. Consumers with high income and educational levels, who are more concerned with the environment and their health and have greater trust in government supervision of rice quality, as well as inland consumers, are more likely to purchase high‐priced japonica rice, whereas the price factor has a negative relationship with such purchases. The results provide sufficient information to the japonica rice industry to expand its market and improve its profitability.


INTRODUCTION
Rice (Oryza sativa L. ssp. indica and japonica) is among the most important staple foods worldwide. More than 60% of the population of China consumes rice, with japonica rice accounting for 30% of rice consumption. Normally, the low price elasticity of rice ensures that its price is not highly volatile. However, income growth, convenience, marketization, and urbanization have significantly changed the rice consumption structure (Aoyagi, 2012;Wu, Wu, Yin, & Chien, 2020). In particular, more and more consumers in the north of China prefer to consume japonica rice instead of wheat (Triticum aestivum L.), and an increasing number of consumers in the south purchase japonica rice instead of indica rice. Another significant change is in the quality of rice; increasing numbers of consumers favor natural, safe, healthy, environmentally friendly, functional (e.g., low protein), and This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Agricultural & Environmental Letters published by Wiley Periodicals LLC on behalf of American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America organic high-quality rice. Consequently, the market share of japonica rice is continually growing, and the price range of rice has been fluctuating wildly. For example, the average market price for normal japonica rice is approximately CNY 8 (yuan [CNY]) kg −1 , but for certain japonica rice varieties it can reach up to CNY 80 kg −1 (Zeng & Xu, 2014).
Previous studies examined the influence of awareness and knowledge on the purchasing behavior or willingness to pay for rice; however, there is limited research on japonica rice (Wu, Zhou, & Chien, 2019;Xu, Su, & Lone, 2018;Zhou, Liu, Mao, & Yu, 2017). Until now, consumer perceptions have remained a largely unexplored mechanism for empirically understanding price structures in the japonica rice market. Clarifying the preference and characteristics of consumers who purchase high-priced agricultural products can provide useful information to producers and enable product differentiation. This study examines how product attributes, consumer attitudes, consumer demographics, and regional differences influence different categories of the daily purchase prices of japonica rice in China.

DATA AND METHODS
Given the existence of regional differences in the Chinese rice market, our survey randomly selected three different regions across 10 provinces (Beijing, Shanghai, Yunnan, Sichuan, Guangdong, Jiangsu, Hubei, Fujian, Liaoning, and Heilongjiang) and targeted respondents who consume japonica rice as their main staple food. The main data collection method was an online questionnaire, conducted November-December 2017. A third-party platform database agency (Baidu) was used to distribute questionnaires to potential respondents (>24 years old). All other procedures (e.g., survey design and data analysis) were performed by the authors. The questionnaire focused on four components: consumer demographics (age, gender, education, and household income), product attributes (brand and place of production), consumer attitudes (price, environment, health, and supervision), and purchase prices of japonica rice. Forms used a fivepoint rating scale. Product attributes and consumer attitudes (price) were measured from very unimportant (1) to very important (5) by the answer to the question "To what extent do you think the brand, place of production (place) or price of japonica rice are important?" Consumer attitudes regarding environment and health were measured from not concerned (1) to very concerned (5) by the answer to the question "To what extent are you concerned about environmental impact or health effect while purchasing japonica rice?" Consumer attitude of supervision was measured from no trust (1) to very high trust (5) by the answer to the question: "To what extent do you trust the food supervision system?" The purchase price of japonica rice was measured using the question ''What is your daily purchase price of japonica rice (yuan/kg)?"; respondents were given choices of purchase prices (four ordered categories) ( Table 1 gives details for each category).
The final set of respondents consisted of 492 randomly selected respondents: 150, 187, and 155 respondents from the inland (region I: Hubei, Yunnan, and Sichuan), coastal (region C: Guangdong, Shanghai, Jiangsu, and Fujian), and northern (region N: Beijing, Liaoning, and Heilongjiang) regions, respectively.
To measure the ordinary dependent variables, we used an ordered logit regression (Model 1) and multinomial logit regression (Model 2). The ordered logit regression is based on the following formula:

Core Ideas
• Product attributes and consumer attitudes influence japonica rice prices in China. • Consumers concerned about the environment are more likely to purchase high-priced japonica rice. • Consumers with more trust in the supervision system are more likely to purchase high-priced japonica rice. • High-income, highly educated consumers are also more likely to purchase high-priced japonica rice. • Price has a negative relationship with purchase of high-priced japonica rice.
where k = 1, 2 . . . , k (the outcome category), X indicates the matrix of independent variables, α k and β are the parameters to be estimated, P (Y ≤ k) is the probability of Y falling at or below a given k, and p 1 , p 2 , p k are the probabilities of Y (an ordinal response). The ordered logit regression in the logit form is as follows: where β indexes the effect of p independent variables on the logit, α k is not used to interpret the results and is only related to the out category, and X = (X 1 , X 2 , X 3 , ..., X 12 ) is a set of exploratory variables.
For the multinomial logistic regression, the dependent variable contains four categories (A, B, C, and D). Choosing category A as a baseline, the equations of each category are as follows: where β 10 , β 11 , β 12 , ..., β 3p are the parameters to be estimated, and X = (X 1 , X 2 , X 3 , ..., X 12 ) is a set of exploratory variables. Table 2 presents a detailed description of the variables and a summary of sample statistics used in Models 1 and 2.  Table 3 presents the results of the ordered logit regression. The results of Model 1 were proved to be robust and statistically significant by examining the Brant test and their variance inflation factors (Brant, 1990). Education; household income; consumer attitudes toward the environment, health, and supervision; and region I were positive and statistically significant, whereas price was negative and statistically significant. Regarding consumer attitudes, if consumers consider price an important factor, they are less likely to buy higher-priced japonica rice. For every one-unit increase in product attribute (price), the odds of being more likely (category B, C, or D vs. category A) to buy was 18.01% lower, holding all other variables constant. Consumers are more likely to purchase highpriced japonica rice if they focus on environment and health when making purchase decisions. In addition, if consumers have greater trust in government supervision of rice quality, they are more likely to buy higher-priced japonica rice. For every one-unit increase in consumer attitudes (environment, health, and supervision), the odds of being more likely to buy was multiplied 1.27, 1.23, and 1.17 times, respectively (i.e., increased by 27.40, 22.93, and 17.15%, respectively), holding all other variables constant.

RESULTS
Consumers from inland areas were more likely to purchase higher-priced japonica rice; their odds of purchasing was 2.13 times that of consumers in the coastal or northern regions, holding all other variables constant. One possible explanation is that inland consumers face higher logistic costs; this is because the inland region only produces a small amount of japonica rice compared with coastal or northern regions. Among demographic factors, if consumers have a high education level or high household income, they are more likely to purchase high-priced japonica rice. For every oneunit increase in consumers' education level, the odds of being more likely to buy was multiplied 1.38 times (i.e., increased by 38.02%), holding all other variables constant. And for every one-unit increase in consumers' household income, the odds of being more likely to buy was multiplied 1.19 times (i.e., increased by 18.71%). Table 4 shows the results of the multinomial logit regression regarding the specific differences between the reference category A and categories B, C, or D. Category A is the reference group compared with category B (Case 1), category C (Case 2), and category D (Case 3), respectively. Case 1 shows that category A and category B have similar values, and only the environment and supervision factors were statistically significant. This is because price differences were not large in Case 1. The results of Case 2 and Case 3 were very similar to Model 1. However, Case 3 shows that the age factor has statistical significance; that is, older consumers were more likely to buy high-priced japonica rice. Finally, consumer attitude toward supervision was not significant.

CONCLUSION
The Chinese agriculture market is facing the issue of agricultural product similarity, partly because producers often ignore consumer demand. Analyzing consumer preferences and regional characteristics could contribute to developing a product differentiation strategy for producers. This study focused on japonica rice, examining how product attributes, consumer attitudes, consumer demographics, and regional differences influence different categories of the daily purchase prices of japonica rice. Consumers with high income and educational levels, who are more concerned with the environment and their health, and have greater trust in government supervision over rice quality, as well as older consumers and inland consumers, are more likely to purchase high-priced japonica rice. These results not only provide an insight into the food market but also provide sufficient information to the japonica rice industry to expand its market and improve its profitability. More important, the results could contribute to the

C O N F L I C T O F I N T E R E S T
The authors declare no conflicts of interest O R C I D Wenhao Wu https://orcid.org/0000-0003-1405-6470