PSY 350 Lecture Notes - Lecture 58: Type I And Type Ii Errors, Analysis Of Variance, Psy
PSY 350 4-30-18
Monday, April 30, 2018
8:59 AM
Lecture: Post Hoc Analyses
Objectives
-differentiate post hoc analyses from planned analyses
-identify situations in which post hoc analyses are appropriate
Post Hoc Analyses
-extra analyses to rule out alternative explanations
-usually reported in the discussion section, when you are considering those
alternative explanations
-include the same details you would have included in the results
-not part of your hypotheses
-not the same a post hoc tests that follow an ANOVA
-those are reported in the results
-if you are using ANOVA with more than 2 levels, you know you need to do
post hoc tests
Why do post hoc analyses?
-post hoc analyses help you address the "what ifs"
-what if there were more conscientious people in the treatment condition
-we have some data about conscientiousness, let's see if there was any
difference using post hoc analyses
If no: that alternative explanation cannot account for your results
If yes: that alternative explanation is plausible
Risks of post hoc analyses
-may not have exactly the data you want
-indirect or proxy measures
-less control
-your conclusions won't be as strong
-risk of capitalizing on chance
-essentially, making a Type I error
-finding a relationship or a difference based on a chance quirk of your sample,
without good theory to support it
-need to view any significant findings with a lot of caution
Responsible Post Hoc Analysis
find more resources at oneclass.com
find more resources at oneclass.com
