This project contains the files for the letter I sent to the Journal of Epidemiology and Community Health about Fancourt and colleagues's [2019 BMJ](https://doi.org/10.1136/bmj.l6377) and [2020 JECH](http://dx.doi.org/10.1136/jech-2019-213029) papers on arts engagement, premature mortality and dementia risk. I use a negative control outcome design ([Tilling et al., 2017](https://doi.org/10.1093/ije/dyw314)) to show their results are still likely biased by unobserved confounding. I will add a link to the letter if it is published.
Data cleaning and regressions were carried out in Stata (see the `.do` file). The plot was made in R using ggplot (see the `.R` file). The `.log` file contains the regression output from Stata.