The Factory Commission and the 1841 Infant Mortality Table
You are a research assistant to a parliamentary committee in 1842 investigating child labour. The committee has two tables for 1841. Millbrook is a textile town with 18,000 inhabitants, 4,200 factory workers and a new poor-law union; its infant mortality is 168 per 1,000 live births. Ashford is a market town with 9,000 inhabitants, 600 factory workers and no large mill; its infant mortality is 112 per 1,000. A factory inspector tells the committee that the difference 'proves the mills destroy children'. A mill owner replies that Millbrook has overcrowded courts, contaminated wells and a large migrant Irish population, while Ashford is a healthy rural parish. The committee must decide whether to recommend a ten-hour limit for children under thirteen. You have the two tables, the inspector's statement and the owner's reply, and no other data.
1. What can the two rates legitimately show, and what can they not show?
Read the worked answer
The rates show an association: in 1841, Millbrook recorded 168 infant deaths per 1,000 live births, while Ashford recorded 112. They do not show that mill work caused the difference. The comparison is between two whole populations, not between children who worked in mills and those who did not. Millbrook is larger, has 4,200 factory workers, overcrowded courts, contaminated wells and a migrant Irish population; Ashford is a small market town with only 600 factory workers. Any of these could affect infant mortality. The inspector's claim commits an ecological fallacy: town-level rates cannot prove individual-level causation. The owner's reply is also interested and incomplete, because it offers alternative explanations without demonstrating them. A good answer notes that infant mortality is shaped by sanitation, housing, income, feeding practices and age structure, not just employment.
2. Which additional piece of evidence would most strengthen a causal claim, and why?
Read the worked answer
The strongest additional evidence would be linked individual-level records: for each household, the occupation of parents, housing conditions, water supply, income and infant deaths. That would allow a comparison between mill-worker and non-mill-worker families within the same town, holding local sanitation partly constant. Alternatively, a before-and-after study of a mill opening or a factory regulation could test whether mortality changed. Another town alone would not help much, because it would add another ecological comparison with new confounders. The key is to move from town-level correlation to a design that can separate mill employment from poverty and environment.
3. Recommend a position to the committee: should the report say that mills cause infant deaths?
Read the worked answer
The committee should not report that mills 'cause' infant deaths on this evidence. It can say that the comparison is consistent with a concern about industrial towns but does not establish causation. If the committee wants a ten-hour limit for children under thirteen, it should justify it on evidence about children's working hours, accidents, schooling and health at the individual level, or on a moral and legal principle, not on these two rates. A defensible report would recommend further local investigation and perhaps sanitary reform alongside factory regulation. This answer shows judgement because it separates a policy recommendation from the limits of the statistical evidence.