PSY319H5 Lecture Notes - Lecture 6: Type I And Type Ii Errors, Analysis Of Variance, Depression In Childhood And Adolescence
9 Apr 2016
School
Department
Course
Professor
Document Summary
Lecture 6 data analysis ii: factorial anovas, linear regression. Used when we have >1 categorical iv o o o. 2-way anova or 2 2 factorial anova o. Why not just do two independent samples t-tests: multiple tests increase the probability of making a type i error, doing an anova keeps the probability of making a type i error at 5% Post-hoc tests should also correct for multiple comparison: you can see if there is an interaction. Main effects of attachment, gender and ethnicity. 2-way interactions between attachment gender, attachment ethnicity, gender ethnicity. 3-way interaction between: attachment gender ethnicity. E. g. , the interaction between attachment and gender might change depending on ethnicity. Independence of observations: dv must be continuous. 3: no significant outliers explore function, dv normally distributed shapiro-wilks; q-q plots (no sig. , homogeneity of variances levene"s statistic (no sig. ) Look at shapiro-wilk, histograms and q-q plots to determine if dv is normally distributed.
