Strategies to address confounding
As discussed, addressing confounding is essential to ensure that estimated effects genuinely reflect the true causal relationship, devoid of influence from extraneous factors. In the following subsections, we present two prevalent statistical methods that are often employed to tackle confounding: regression adjustment and propensity score methods.
Regression adjustment
This method is a staple in controlling for confounding in observational studies. Luckily, we have already applied this method in this chapter. As you have seen previously (in the Individual treatment effect section), the primary idea is to integrate potential confounders as covariates into a regression model, thereby separating the effect of the treatment or exposure of interest from the influences of the confounders. We saw the impact of a treatment 
on an outcome 
, alongside a collection of confounders 
. A typical linear regression model might be formulated as follows:
...
