LINB29H3 Lecture Notes - Lecture 11: Univariate, Type I And Type Ii Errors
27 Mar 2017
School
Department
Course
Professor
Document Summary
Recall hypotheses: h0: observed freqs = expected freqs, h1: observed freqs =/= expected freqs. Univariate stats: means t-test, non-parametric test, dispersion f-test. Multivariate: freq chi-square, correlation pearson/kendall. Analysis of variance (anova): recall the f-test tests homogeneity of variance, anova compares all cells at once (compares variances not means: anova compares more than 2 means, anova with more than 1 factor, repeated measures. Running multiple t-tests would increase the likelihood of a type i error (rejecting null hyp when it is true) One-way anova: 1 factor that has more than 2 levels (ex. Height high, medium, low: h0: all means are all equal, mean for level 1 is equal to mean for level 2, etc, h1: at least one of the means is different. Pitt & shoaf (2002: 3 conditions: early, mid, late (in experiment, reaction times decrease (they"re getting faster/better) at responding. A classic anova collects one data point from each participant.
