When experiments fail to mask, the effect size is estimated to be inflated by 0.28σ to 0.91σ.

Use the sliders to explore how sample size and bias (modeled here as a positive increase to the true effect size) influence the amount of false positives.