PSYC207 Lecture Notes - Lecture 11: Null Hypothesis, Construct Validity, Random Assignment

• School: University of Delaware
• Department: Psychology
• Course: PSYC207
• Professor: Peter Mende-Siedlecki

Document Summary

External to whom or what can the causal claim generalize. Is the difference statistically significant: check the p-value associated with the difference between 2 or more groups, t-test, null hypothesis in the population, there is no difference between the two groups. How large is the effect: che(cid:272)k the effe(cid:272)t size of the diffe(cid:396)e(cid:374)(cid:272)e (cid:271)et(cid:449)ee(cid:374) the t(cid:449)o g(cid:396)oups (cid:894)cohe(cid:374)"s d(cid:895) The clustering of scores around means; sd difference; individual scores being closer/farther from the mean on average: cohens d = difference of 2 means / pooled sd. The way we construct levels of iv may have inherent confounds to it. If a within-subjects design was used, did they control for order effects by counterbalancing: counterbalancing the order in which the conditions/levels are experienced, or other low level aspects of the task as well. Additional validity concerns to be addressed (hopefully) in experiments.