Does online food shopping boost dietary diversity? Application of an endogenous switching model with a count outcome variable

Abstract

Increasingly, rural households in developing countries are shopping for food online, and the COVID-19 pandemic has accelerated this trend. In parallel, dietary guidelines worldwide recommend eating a balanced and healthy diet. With this in mind, this study explores whether online food shopping boosts dietary diversity—defined as the number of distinct food groups consumed—among rural households in China. Because people choose to shop for food online, it is important to account for the self-selection bias inherent in online food shopping. Accordingly, we estimate the treatment effects of online food shopping on dietary diversity using the endogenous switching model with a count outcome variable. The results indicate that online food shopping increases dietary diversity by 7.34%. We also find that education, asset ownership, and knowing the government’s dietary guidelines are the main factors driving rural households’ decisions to shop for food online.

Analysis of incidence, intensity, and gender perspective of multidimensional urban poverty in Kenya

“…Based on these findings, it can be deduced that multidimensional poverty varies across different poverty cutoff thresholds, signifying that the adjusted headcount ratio (MPI) is highly sensitive to the choice of cutoff value. These findings agree with the assertions by Ma et al [ 102 ] and Tigre [ 107 ], who deduced that poverty cutoff and MPI value vary in different and opposite directions (see Fig. 6 ).…”

“…Consistent with studies such as Rono et al [ 99 ] and Koomson et al [ 100 ], this paper resolves the endogeneity problem in the study by applying IV-Probit model regression whereby the product of household head age and gender is applied as the instrumental variable. As indicated in Table B in the Appendix section, the falsification results show that the association between the selected IV and the response variable (household poverty, which is dichotomous) is weak, while the selected IV is significantly associated with the treatment variable (food insecurity) [ 101 , 102 ]. The results support the validity of the selected IV [ 102 ].…”

“…As indicated in Table B in the Appendix section, the falsification results show that the association between the selected IV and the response variable (household poverty, which is dichotomous) is weak, while the selected IV is significantly associated with the treatment variable (food insecurity) [ 101 , 102 ]. The results support the validity of the selected IV [ 102 ]. In addition, the paper utilizes the selected IV in testing for the consistency of results using Lewbel's [2012] 2SLS technique, which has been extensively used in literature to address the endogeneity problem [ 101 , 103 , 104 ].…”

Does nutrition knowledge training improve dietary diversity and nutrition intake? Insights from rural China

“…The selection of control variables is drawn upon existing literature related to nutrition knowledge programs (e.g., Davidson et al, 2021;Hou et al, 2021;Mwale et al, 2022) and household food consumption and nutrition intake (e.g., Korir et al, 2023;Ma et al, 2022aMa et al, , 2022bQin et al, 2023). Specifically, the selected control variables capture respondents' characteristics (e.g., age, gender, education level, and health knowledge), household-level characteristics (household size, children ratio, and asset ownership), and villagelevel characteristics (e.g., income level, number of food stores, express service outlet, ICT facility and distance to the food market).…”

“…In this study, dietary diversity refers to the number of food groups (breakfast, lunch, and dinner) consumed in the last 3 days (72 h), which is also called the household dietary diversity score (HDDS) (Ma et al, 2022b; Mulenga et al, 2021; Shahzad & Abdulai, 2021; Sibhatu et al, 2022). Concisely, a higher dietary diversity core signified a diversified diet pattern.…”