PSYC19H3 Study Guide - Midterm Guide: Brian Nosek, Precognition, Publication Bias
Document Summary
Brian nosek and colleagues repeated 100 published psychological experiments to see if they could get the same results the second time around. Reproducibility project: 1st big systematic attempt to answer questions about what proportion of results in psychology field is reliable. Signs that reliability in psychology research is low: Failure to replicate results of classic textbook experiments. Publication bias: jour(cid:374)als te(cid:374)d to o(cid:374)l(cid:455) pu(cid:271)lish positi(cid:448)e results that (cid:272)o(cid:374)fir(cid:373) the resear(cid:272)her"s hypothesis p-hacking: attempts to torture positive results out of ambiguous data ex. Only reporting the results of successful experiments, checking to see if they had statistically significant results before deciding whether to collect more data. 97% of 100 studies originally reported significant results but just 36% of the replications did. A result is statistically significant if the p-value is less than 0. 05 which means if you did your study again, your odds of fluking your way to the same results would be less than 1 in 20.
