Background
The United Nations’ Millennium Development Goals Report, 2015, documents that, since 1990, the number of stunted children in sub-Saharan Africa has increased by 33% even though it has fallen in all other world regions. Recognizing this, in 2011 the Government of Uganda implemented a 5-year Nutrition Action Plan. One important tenet of the Plan is to lessen malnutrition in young children by discouraging over-consumption of nutritionally deficient, but plentiful, staple foods, which it defines as a type of food insecurity.

Methods
We use a sample of 6101 observations on 3427 children age five or less compiled from three annual waves of the Uganda National Panel Survey to measure undernourishment. We also use the World Health Organization’s Child Growth Standards to create a binary variable indicating stunting and another indicating wasting for each child in each year. We then use random effects to estimate binary logistic regressions that show that greater staple food concentrations affect the probability of stunting and wasting.

Results
The estimated coefficients are used to compute adjusted odds ratios (OR) that estimate the effect of greater staple food concentration on the likelihood of stunting and the likelihood of wasting. Controlling for other relevant covariates, these odds ratios show that a greater proportion of staple foods in a child’s diet increases the likelihood of stunting (OR = 1.007, p = 0.005) as well as wasting (OR = 1.011, p = 0.034). Stunting is confirmed with subsamples of males only (OR = 1.006, p = 0.05) and females only (OR = 1.008, p = 0.027), suggesting that the finding is not gender specific. Another subsample of children aged 12 months or less, most of whom do not yet consume solid food, shows no statistically significant relationship, thus supporting the validity of the other findings.

Conclusion
Diets containing larger proportions of staple foods are associated with greater likelihoods of both stunting and wasting in Ugandan children. Other causes of stunting and wasting identified in past research are also confirmed with the Uganda data. Finally, the analysis provides clues to other possible causes of undernourishment in young children.