Within parapsychology, there is a tendency to accept any positive replications but to dismiss failures to replicate if the procedures followed have not been exactly duplicated.
—Chris French, on why his article showing ESP doesn’t exist, an article replicating a previous study showing ESP does exist, was not published in the same journal. An illustration of why many “scientific” journal should not consider itself scientific.
Now, I must acknowledge that you can't publish 100 studies all showing ESP or ghosts don't exist, but once you publish one that says they do exist, you are obliged to publish at least a few others that show they don't exist.
If there is a sample selection problem, the sample is biased. Thus, the Journal of Personality and Social Psychology is biased.