I am currently writing a book chapter and paper about different causal methods and using breast feeding with different outcomes as an illustrative example (largely using published work, including some publications that have used ALSPAC data). I would like to include as an e.g. a negative outcome control study i.e. a study in which I would look at the association of breast feeding with an outcome that biologically it would not be expected to affect (but might be associated with via confounding).

ALSPAC has questions on whether the family home had ever been invaded by rats, dogs, cats, mice, cockroaches, pigeons, woodlice etc. (see picture attached with this email) which would be ideal as negative controls.

In both the book chapter and paper these analyses would be a small part of the whole document

I would look at the simple association of breast feeding with one or more of the 'animal/insect home invasion' responses - depending on how common they are in the cohort.