Why does correlation not equal causation?

In this blog post I will explain why it makes sense to think about causal explanations when we see correlation, why correlation does not always imply causation, and which alternative causal models to consider when you are trying to figure out why two variables in your data are correlated. The blog post is intended as a non-technical introduction. I use only words and pictures to explain all concepts and logic, and I use a hypothetical example from health psychology to illustrate. The post should be relevant to anyone who is interested in using data to understand causal mechanisms.

What is Worth Replicating? - SIPS 2020 Unconference session

Given that most research is original and we have limited resources available for replication, we need guidelines for study selection in replication research. But what makes a study worth replicating? In our unconference session at SIPS this year, we had the pleasure of discussing this problem with fifty-some enthusiastic scholars. Here is a quick summary of what they taught us. First off, we went into this session with our own prior beliefs about what makes a study worth replicating.

Fifty ways to leave your model

A method for calculating the number of alternative causal models, given a number of variables/nodes.

Meta-research at the Psychological Science Accelerator

Summary of unconference session at Tilburg Meta-research day 2019.

Mixed model equivalence test using R and PANGEA

How to calculate an equivalence test, and power for an equivalence test, for a fixed effect in a mixed effects model using R and PANGEA.