PSYC 316 Lecture Notes - Lecture 5: Statistical Parameter, Design Of Experiments, Summary Statistics
Document Summary
Moving forward from nhst: nhst was not banned, but 3 suggestions were made, report effect sizes (es), report es confidence intervals, use graphics. Hedges" g: variance accounted for es: i. ii. iii. Useful for looking at the association among variables. 2 ( eta squared : more common to see these es corrected : Corrected to account for sampling error variance: smaller sample size, more variables measured, small population effect size, report effect sizes, report exactly what type of es you are presenting: i. Hedges" g? r? d = 0. 05 r2 = 0. 05: many conversion formulas available that differ. iii. No particular type of es is endorsed. b) Interpret es in context (if no prior context, use benchmarks): i. The judgment regarding the interpretation of an es rests with: Experimental design (t1 vs. t2; pre vs. post). Societal concerns: report es for all results (significant or non-significant). It is far more complicated to calculate cis for es: