Name the confounder: number of churches and number of bars

Question

An analyst plots data across many towns and finds that towns with more churches also tend to have more bars. A columnist writes that religion and drinking must go hand in hand.

What third variable drives both counts, producing the correlation? Name it and explain.

Accepted Answer

The confounder is the town's population. Bigger towns have more people, so they support more churches and more bars, more of nearly every kind of establishment. Population is a common cause driving both counts. Churches and bars are not linked to each other; they simply both scale with how many residents a town has. The fix. Convert to per-capita rates, churches per 1,000 people and bars per 1,000 people, and the correlation collapses (or even flips). The raw counts are confounded by town size, which is why analysts normalize by population before comparing. Takeaway. Counts of almost anything…

Name the confounder: number of churches and number of bars

Your answer

More Statistics questions