Last year, I had a very informative and enjoyable blog dialogue, or diablog, with Dan Simons about the reliability and validity of replication attempts. Unfortunately, there was never an easy way for anyone to access this diablog. It has only occurred to me today (!) that I could remedy this situation by creating a meta-post. Here it is.
In my first post on the topic, I argued that it is important to consider to consider not only the reliability but also the validity of replication attempts because it might be problematic if we try to replicate a flawed experiment.
Dan Simons responded to this, arguing that deviations from the original experiment, while interesting, would not allow us to determine the reliability of the original finding.
I then had some more thoughts.
To which Dan wrote another constructive response.
My final point was that direct replications should be augmented with systematic variations of the original experiment.