Is There a Root Cause for This Low Replicability?

If we use common tests and look at p-values less than 0.05, then we should expect one in 20 studies doesn't replicate. We're seeing a much more frequent rate. Most people in the reformist movement in psychological science tend to blame the questionable research practices for this low replication rate. The good practitioners of machine learning know about all of these pitfalls and usually they experience it in their own skins.