Effect sizes and power in cognitive neuroscience and psychology literature

For those not in the know, John Ioannidis is responsible for much of the current focus on reproducibility in science, following his seminal 2005 paper "Why Most Published Research Findings are False." (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1182327/)

I'd love someone more adept at the statistics than me to shed some light on the veracity and implications of this study.

Based on the abstract it seems pretty damning for cognitive science research conducted over the last 5 years (the period covered in the study). Perhaps it extends even farther back if most of those studies were built on previous research.

Post-hoc power assessments? c'mon e.g. Hoenig & Heisey (2001) provide evidence that several different approaches for calculating post-hoc power are flawed and can produce misleading conclusions. Once a confidence interval has been computed, there is no additional information that a post hoc power calculation can provide.

Nevertheless cog neuroscience has had a bad habit of using small samples, sample sizes that developmental cognitive neuroscientists have criticized.

Full title "Empirical assessment of published effect sizes and power in the recent cognitive neuroscience and psychology literature" was too long. Authors: Denes Szucs and John Ioannidis