PSYB04H3 Lecture Notes - Lecture 8: Repeated Measures Design, Internal Validity, Selection Bias
Document Summary
Experiments part ii: three things needed to establish causality, temporal order must be correct, variables must covary, no other variable is causing the outcome. Solution: counterbalancing: maturation: processes such as growing older, growing stronger, and even growing tired and impatient. E. g. repeated measures design used to examine effect of student discussion groups on attitudes toward school safety. Non-discussion group comes first, then there is news of a school shooting. Solution: comparison group: regression to the mean. Unusual scores followed by scores closer to group mean. Solution: use comparison group: attrition (mortality) Solution: carefully examine the data, remove scores if: testing appropriate. Solution: post-test only, recalibration, interrater calculations: observer bias (experimenter bias) Experimenter treats subjects differently by treatment type: demand characteristics. The participants behave differently based on the condition they"re in: placebo effects. Participants in control group behave as though they were in experimental: solution for all: blind or double blind procedure, participant bias.