PSY 0035 Lecture Notes - Lecture 17: Random Assignment, Observer-Expectancy Effect, Blind Experiment
Document Summary
Internal validity: ability to rule out alternaive explanaions for causal relaionship. Threats: soluions: design confound: control for other variables, selecion efect: random assignment, larger sample, order efect: counterbalance, tesing/pracice/faigue: comparison group, maturaion: comparison group. Experimental group changes over ime only because of natural development. History: experimental group changes over ime because of external factor/event; afects most/all members of the group: atriion: do analysis only on paricipants that don"t drop. Experimental group changes b/c most extreme cases systemaically drop out and scores aren"t included in postest: regression to the mean: extreme comparison & experimental. Experimental group whose avg is extremely low/high at pretest will get beter/worse over ime b/c random events that caused extreme pretest scores don"t occur again in the postest: instrumentaion: use the same forms/scales throughout. Experimental group changes b/c measurement technique changes: observer bias: blinded design. Experimental group raings difer from comparison group because researcher expects raings to difer: demand characterisic: between subject or double-blind design.
