Analysis of variance for BMI Z scores by sex and change in MDM consumption status from IHDS-1 to IHDS-2.

Recent empirical evidence emphasizes the higher prevalence of overweight and obesity for women, especially in developing countries. However, the potential link between gender inequality and obesity has rarely been investigated. Using longitudinal data from India (IHDS 2005-11), we implement Hausman-Taylor and fixed-effect models to estimate the effect of different dimensions of gender inequalities on female overweight. This study demonstrates that the form of gender inequality or women’s mistreatment differently affects female bodyweight. Indeed, we show that some forms of women’s mistreatments (such as perceived community violence and age difference with husband) increase the risk of female overweight, whereas more severe forms of abuse such as child marriage increase the risk of underweight. Moreover, we also find that higher decision-making power and autonomy about outings are risk factors of weight gain and obesity, especially in urban settings, perhaps indicating a higher exposure to urban obesogenic lifestyles. To conclude, our results suggest that, although improving women’s status in society may be a key action to address the epidemic of obesity, policies must also target hazardous habits that emancipation may imply in urban (obesogenic) environments. These meta-data include: (i) the merged database from the two waves of IHDS we used in the study (.dta in Stata format); (ii) the codes used for data treatment and analysis (.do in Stata format). Original IHDS data are freely available on: https://ihds.umd.edu. Further details about our methods and results will be published soon in a scientific journal and will reference these meta-data.

Keywords: India; Gender inequality; Obesity; Hausman-Taylor estimations; Fixed effects estimations. JEL codes: I14 I15 J16

Null models: With District Effects (Model 1) and District and Community Effects (Model 2).

OLS estimates (unstandardized β coefficients), IHDS-II.

The three levels random slope model.

Analysis of variance for BMI Z scores in IHDS-2 and MDM consumption status from IHDS-1 to IHDS-2.

Underweight prevalence among school-aged children by socioeconomic characteristics and change in MDM consumption status from IHDS-1 (2004–2005) to IHDS-2 (2011–2012).

Prevalence rate of asthma for diagnosed cases** and reported cases*, India and sub-region, 2011–12.

Socio-demographic characteristics and growth outcomes of the study population, IHDS-1 (2004–2005) and IHDS-2 (2011–2012).

Note-# Reference groupsMultiple linear regression predicting menarcheal age among ever-married women (15–49 y) in India, IHDS, 2004–2005.

Original set indicators and dimensions of empowerment as self-compassion in the theoretical model (before CFA).

Pattern of neurocognitive performance by cognitive domain based on MoCA and IHDS, N(397).

Oaxaca decomposition:contribution by endowments, coefficients and interaction to Male-Female difference in health care expenditure, 2011–2012.

Multilevel mixed effects models of treatment seeking.

Note that though the census tables enumerate the entire population of the country at the macro-level, we sample at random from its joint distribution a smaller micro-data sample of size noted in the table, which we use later to integrate the IHDS.