WEEK 11 FALSE POSITIVES

Publication bias

Result of an experiment affects the likelyhood of a paper being published

File drawer problem

Studies with significant results are more likely to be published - accumulation of unpublished null effects

Lack of transparency

Researchers not giving all info about design, results, methods, data processing - affects results replication

Researcher degrees of freedom

Researchers can analyse their data many times in different ways - increases likelyhood that one of the analyses yields results they want to see

Can report effect that isn’t true

Transparency

Relevant info that was omitted ?

False positive psychology

Tendency to observe an effect that doesn’t exist