abstract Gregory Francis

Gregory Francis (Purdue University, U.S.A.)

The psychology of replication and replication in psychology

Experimental replication is widely believed to be the gold standard for validating scientific findings. However, when replication is based on statistical analysis it is possible to have too much replication success. I will show how to detect such situations for several findings in psychological science. I further show that for two prominent journals that publish research on psychology, over 80% of the articles with four or more experimental results suffer from results that appear to be "too good to be true", relative to the magnitudes of the reported effects. Such findings undermine the theoretical conclusions in those articles and cast doubt on much of the scientific research that is reported in top journals.