Background
Maternal education is strongly associated with young child nutrition outcomes. However, the threshold of the level of maternal education that reduces the level of undernutrition in children is not well established. This paper investigates the level of threshold of maternal education that influences child nutrition outcomes using Demographic and Health Survey data from Malawi (2010), Tanzania (2009–10) and Zimbabwe (2005–06).

Methods
The total number of children (weighted sample) was 4,563 in Malawi; 4,821 children in Tanzania; and 3,473 children in Zimbabwe Demographic and Health Surveys. Using three measures of child nutritional status: stunting, wasting and underweight, we employ a survey logistic regression to analyse the influence of various levels of maternal education on child nutrition outcomes.

Results
In Malawi, 45 % of the children were stunted, 42 % in Tanzania and 33 % in Zimbabwe. There were 12 % children underweight in Malawi and Zimbabwe and 16 % in Tanzania.The level of wasting was 6 % of children in Malawi, 5 % in Tanzania and 4 % in Zimbabwe. Stunting was significantly (p values < 0.0001) associated with mother’s educational level in all the three countries. Higher levels of maternal education reduced the odds of child stunting, underweight and wasting in the three countries. The maternal threshold for stunting is more than ten years of schooling. Wasting and underweight have lower threshold levels.

Conclusion
These results imply that the free primary education in the three African countries may not be sufficient and policies to keep girls in school beyond primary school hold more promise of addressing child undernutrition.