Offering a Higher Credit Card Limit: Working with Noncompliance

Let's continue with the CreditCo dataset described in the previous exercise. The CFO has assigned you to figure out the average treatment effect of taking a credit line increase offer on late payments (default_post), and he wants to see something before lunch. You realize that you can give him a simple result: you just need to do some quick checks for balance in your samples and compute a "naive" average treatment effect to get the analysis started.

A simulated version of this dataset, CreditCo, is available in the workspace. With that data:

1) See if treatment and control groups (offered) are of equal size.
2) Compute the fraction of people who opted in (opt_in) to the credit offer.
3) Compute a naive average treatment effect on late payments (default_post).

Exercise

Offering a Higher Credit Card Limit: Working with Noncompliance

Instructions