July 2015 - Explaining how Family Structure Influences Children's Growth

Post date: Jan 15, 2016 6:13:29 PM

Many studies offer explanations on the change of children’s growth pattern in history, and these numerous explanations can be easily divided into two categories.

1. The first set of factors influence children’s health by directly contributing to the total amount of food available for one household.

2. The second set includes factors determining how food, resource and workload are distributed among children within a household.

There is a large body of literature that shows how the first category drives up children’s growth in history, but our focus here lies in the latter.

The newly collected dataset from the North Surrey School District (NSSD) is so far the only nineteenth century individual-level dataset that allows children’s growth pattern to be reconstructed. It contains information on heights and weights for 3,842 children, both girls and boys (1:1.35), who were born between 1866 -1892. More importantly, 53% of the children enrolled in the NSSD were along with their siblings, which provides us a unique opportunity to assess how various household structures may influence the allocation of resource among children, therefore their growth pattern, in the late 19th century Britain.

To begin with, preliminary analysis using NSSD data show that there is a large within-household variation in terms of z-score [1] for both height and weight, indicating that children do not necessarily have similar health conditions within a same family. What factors may affect parent’s intentional or unintentional decisions on food, other resources and workload distribution among their children? Presumably, gender, the number of siblings, age gap between siblings and order of birth all have potentials impacts.

Holding the amount of resource constant for a household, the most obvious factors that affect the resource distribution include the size of the family and the age-band among children. We start our analysis by first comparing families with different number of sibling. Surprisingly, in this exercise, no trade-off between the quantity and quality of children has been observed. Instead, children who enrolled in the school alone had worse health condition comparing with children who had other siblings admitted to school together. The correlation between a child’s health and the number of siblings is not linear, but an inversed U-shape, reaching its tipping point when the sibling number is 3. Of course, this finding requires further rigorous investigation, since the number of siblings we observe in the data is only a proxy for the actual household size. Parents’ decision on how many children and which specific child should be sent to schools are not random. Therefore, a systematic measurement error might exist if we adopt the number of siblings who enrolled in school as an indicator for household size.[2] When we turn to another aspect of family structure -- age band between children, defined by the average age difference between a child’s immediate siblings, the findings show that controlled for the sibling numbers, the larger the age gap is, the worse off a child could be. It also contradicts with our intuition that the larger age gap between siblings may allow parents to accumulate more savings or other resource to support a higher living standard for their children.

Apart from comparing cross families, we also adopt family fix effect to explore individual features that may influence on food allocation under one roof. First, boys did not perform better than the girls for a given household. Even in the most extreme cases that there was only one boy with all other female siblings, no boy preference has been detected. Furthermore, the importance of birth order in explaining a child’s health condition emerges. Apparently, later born child, regardless boy or girl, tends to be worse off comparing with his/her older siblings, and this is probably due to insufficient attention and food allocation.

Figure 1:

In short, children’s growth is deeply influenced by household structure, which determines how food, resource and housework are allocated to each child for a given household. Further investigation will be pursued and the data we newly collected will facilitate our exploration in our next step.

Pei Gao, July 2015

[1] Z-scores are a standard deviation classification system provided by WHO, which is widely recognised as the best system for analysis and presentation of anthropometric data.

[2] For instance, one thing we interesting shows in the NSSD dataset is that girls are more likely to be sent away to school when the family size is small, while with the household get larger, more boys were sent to school, implying that girls may stay at home and take care housework.