Probability models – The daily blog of behavioral and cognitive economics

Why you should be using linear models for your RCTS with binary outcomes

If you think this is controversial, this is the blog for you.When working in behavioural insights, the outcome we want to change is often measured with a binary variable. Binary outcomes are events with only two possible values, like whether someone submitted a tax return or not. When we model these outcomes, we usually include the two categories as ones and zeros where 1 = the target outcome (i.e. submitting their tax return) and 0 = not the target outcome (i.e. not submitting their tax return). ….[READ]