This is a second paper in the "many labs" project trying to replicate seminal or influential studies in psychology and behavioral science. Like the prior many labs, they find that some things replicate, but many do not. In this case they are finding about half the studies replicate and typically at smaller effect sizes than the original publication when they do.
More interestingly, in this version they tried a number of things to get to the questions of whether things don't replicate is because the finding is somehow "fragile". That is, either the phenomena does show up strongly, but only under certain conditions, or it is generally not real "effects". Here is what one of the leading authors, Brian Nosek wrote about this: "
The main purpose of ML2 was to examine heterogeneity across sample & setting. Some heterogeneity was observed. It was mostly in large effects, not in weak effects. The notion that some “fragile” effects are highly sensitive to sample had no support here." That is a very interesting finding.
Another finding from the study was that within large/strong effects, while there was some heterogeneity across populations (e.g., WEIRD vs non-WEIRD populations), the "replicable effects" are usually similar across different populations. Brian Nosek had a nice "tweek storm" about the findings
