Multi-armed bandits, and the Duolingo example

Duolingo, the company behind the famous language learning app, face a similar challenge to the majority of companies whose existence largely depends on usage of an app or website. They want to promote "user engagement". That's to say it's in their interest - and fortunately in this case also in the interest of their customers … Continue reading Multi-armed bandits, and the Duolingo example →

How to evaluate the results of an experiment early and often without increasing false positives

Most data folk I know love experiments. They're the ideal way to use data to answer the question of not only whether A is associated with B, but also if A causes B. Randomised Controlled Trials are a subset of experiments that most interested people seem to agree are the gold standard in, for instance, … Continue reading How to evaluate the results of an experiment early and often without increasing false positives →

Create similar test and control groups by randomising participants with blocking in R

In the classic randomised experiment we randomly assign participants to at least two groups, test and control, by metaphorically tossing a coin to allocate them to one or the other. However, in reality sometimes slightly more sophisticated methods can be useful. One such method is blocking. Here, you first create "blocks" of participants, usually based … Continue reading Create similar test and control groups by randomising participants with blocking in R →