Would there be confounding variables though?
Users who have the app are anyways likely to be the ones who use Reddit more so naturally it has much higher engagement than on a browser.
Users who have the app are anyways likely to be the ones who use Reddit more so naturally it has much higher engagement than on a browser.