- Can recognize pre-post and other repeated measures designs (and distinguish within from between designs).
- Can make the decision when to enter a predictor variable as a fixed effect or a random effect.
- Can compute and interpret an intraclass correlation based on statistical output.
- Know how to use syntax to run a linear mixed model.
- Can interpret the output of a simple linear mixed model analysis for any within design, predict values, calculate confidence intervals and do hypothesis testing.
- Can report a simple linear mixed model analysis in APA format.
- Know when to use a nonparametric alternative for a linear mixed model.
- Know what nonparametric alternatives are available for linear mixed models.
- Know when to perform a logistic regression analysis, know how to do it, interpret and report the results APA style.
- Know how to compute probabilities from logoddsratios and vice versa.
This Research Methods study unit in module 3 expands the linear model to the generalized linear mixed model. First we introduce linear mixed models and random effects. Departing from analyzing pre-post designs, we extend this to repeated measures analyses in general, including analysing mixed designs (within+between mixed designs). Students are familiarized with nonparametric alternatives tests: Friedman’s and Wilcoxon’s test. Next, generalized linear models are introduced, but limited to analysing dichotomous outcome data through logistic regression. Great care is giving to choosing the right analysis and reshaping the data matrix appropriately. Through the entire course we will make active use of R and students will learn how to interpret the output properly The interpretation of results and their generalization to populations are also emphasized. Students practise with APA-style reporting.
This study unit is part of the module Cognition and Development (202000330). A module is offered as one educational unity and students take it as such.