When we analyse experimental data, we have many choices. What observations do we exclude? What variables do we compare? What statistical tests do we use? And so on. These choices lead us into what is often called the garden of forking paths. The problem is that some paths lead to a “significant” result. Hurrah! Publication on the way. Others lead nowhere. So when we read a published paper, you might ask: did they choose their path because it led to a significant result? One solution to th...
