A confounder is a variable that influences both the dependent variable and independent variable. 1

One way to mitigate the effect of potential confounders is to split samples into subgroups with respect to the confounder (see Simpson’s paradox). However, subgroup analysis does not guarantee to find the problem, and increasing the number of comparison can even create other problems, such as inflated false positives due to multiple comparisons. 1

Footnotes

  1. Confounding - Wikipedia 2