Dietary patterns in China have changed dramatically over the past few decades as the Internet has become rapidly available. Based on data from the China Health and Nutrition Survey (2006–2011), we use a two-way fixed effects model and an instrumental variable approach to determine the impact of Internet use on the dietary quality of rural residents. The results indicate that Internet use could significantly improve the dietary quality of Chinese rural residents, with an increase of about 10.4% in the China Food Pagoda Score (CFPS), mainly due to the increase in the dietary quality score for five food groups: fruits, meats, eggs, oil, and salt. We also found that Internet use significantly increased the consumption amounts of milk and its products (4 g), fruits (31 g), eggs (8 g), and vegetables (34 g), while also decreasing the intake of salts (2 g) and oil (6 g). A possible mechanism is that Internet use improves the dietary knowledge of rural residents, thus optimizing their dietary structure. Moreover, the effect of the Internet was greater among females and those who prepare food for a family. Rural residents without a college degree enjoyed more benefits. In summary, governments should further promote Internet penetration in rural areas for health purposes.

Fresh milk represents a major type of dairy product in China, while it remains at a low level in terms of consumption. Consumers’ preferences are a crucial factor determining consumers’ attitudes and behavior towards fresh milk consumption. However, little is known about what drives consumers’ preferences for fresh milk. This study intends to fill this gap based on a survey with a sample of 1248 respondents in ten cities in China. The best–worst scaling (BWS) method was employed to measure the importance of different attributes that consumers consider when selecting fresh milk. The BWS result indicates that safety certification, shelf-life, and nutrition were ranked as the most important attributes of fresh milk, whereas origin, purchasing location, and package were found to be the least preferred attributes. Furthermore, the study also explored heterogeneities by dividing all samples into a developed area and a less developed area, and a latent class model (LCM) was then applied to classify consumers based on their preferences and demographics in these two areas, respectively. In developed areas, consumers were divided into “Safety First” and “Taste Sensitive”, and Gender and Age were significant determinants of class membership. In less developed areas, consumers were classified as “Quality Sensitive” and “Brand Sensitive”, while Education, Pregnancy, and Health condition were significant determinants of class membership. The findings might be useful for the government in terms of encouraging dairy companies to implement safety management certification, and suggest that companies should design differentiated strategies in different areas.