# Matching and Weighting Methods

Learn about matching and weighting methods for causal inference.

StartPropensity Scores

Lesson 1 of 1

- 1One of the main assumptions in causal inference is known as the assumption of
*conditional exchangeability*. This assumption states that, so long as we account for confounders (the non-treatment, n… - 2The first step in a propensity score analysis is to check how similar the treatment and control groups are at baseline, before using propensity score methods. There are two measures that are common…
- 3Suppose we are interested in determining whether the practice of meditation increases the amount of sleep that university students get per night. To gather more information, we surveyed 250 student…
- 4Distribution plots are great for numeric variables, but we need a different type of plot for categorical variables. Fortunately, we can use the exact same bal.plot() function from cobalt with no ne…
- 5While visual assessments of balance are definitely helpful, we can also assess overlap and balance numerically using the standardized mean difference (SMD) and variance ratio for each variable. Ob…
- 6Returning to our student sleep data, we are interested in the effect of meditation on sleep. It seems intuitive that people with high levels of stress might struggle with sleep AND might be less li…
- 7Now that you know about the basics of propensity scores, let’s talk about some possible applications. Propensity scores often show up in matching and stratification. However, we will focus on _pr…
- 8If this seems like a lot of work, don’t worry! The WeightIt package in R has functions to model the propensity scores and simultaneously perform propensity score weighting. We don’t need to make a …
- 9If propensity score weighting is successful, we expect the distribution of propensity scores in the treatment group to be similar to that of the control group. To check the overall balance of pro…
- 10As you may have noticed, propensity score methods are an iterative process: we check variable balance, model propensity scores, perform weighting, then check balance again. If imbalance still exist…
- 11Now that we have a good balance, we can proceed to the last step of a propensity score analysis: estimating the causal treatment effect. If we think back to the beginning of our lesson, the motiv…
- 12Congratulations! You have learned a lot about propensity score methods in this lesson and are quickly becoming a master of causation (not just correlation). In this lesson, you learned: - There …

## What you'll create

Portfolio projects that showcase your new skills

## How you'll master it

Stress-test your knowledge with quizzes that help commit syntax to memory