PS2024 Statistics - ANOVA 0.0 / 5 ? PsychologyANOVAUniversityNone Created by: maisie2592Created on: 13-01-20 12:07 what does degrees of freedom actually mean? how many numbers in a set are free to vary (usually N-1) 1 of 34 What is the relationship between ** and MS? MS = **/df 2 of 34 what does the F ratio actually calculate? treatment effects 3 of 34 If there were no treatment effects, what should F be in theory? 1 4 of 34 What are the 3 assumptions of BS ANOVA? normality, independence, and homogeneity of variance 5 of 34 how is homogeneity of variance measured? the Fmax test. 6 of 34 How would you report an F ratio formally? F(df effect, df error) = F, p level, MSE 7 of 34 Why is the repeated measures ANOVA more sensitive to treatment effects? because it removes individual difference, which reduces the SSerror 8 of 34 what are the assumptions of sphericity, necessary for a RM ANOVA? correlations between treatment conditions should be equal, homogeneity of covariance 9 of 34 what does homogeneity of covariance mean and how is it calculated? the differences between pairs of scores should be roughly equal, and a significant Mauchley's test means this is NOT the case. 10 of 34 What can you do if the assumption of sphericity is violated? Use the Greenhouse-Geisser correction, which makes the F test more conservative. 11 of 34 What is the difference between pairwise comparisons and complex comparisons? complex comparisons use averages to compare 12 of 34 which post-hoc test only looks at pairwise comparisons? Tukey 13 of 34 which post hoc test looks at all possible comparisons? Scheffe 14 of 34 which post hoc test looks at only a sub set of comparisons? Dunnett 15 of 34 which post hoc test adjusts probability level rather than critical F value? Bonferroni 16 of 34 which post hoc test would you typically use for a repeated measures design? Bonferroni 17 of 34 why not use a t test to compare means instead of post-hoc testing? it would be too liberal, because it doesn't use the family-wise type 1 error rate 18 of 34 what are 2 advantages of using 2-way factorial ANOVA? more realistic and more economical 19 of 34 what is a mixed-model ANOVA? involves between subject and within subject factors 20 of 34 what is the rule for error terms in mixed-model ANOVA? any time there is a within-subject component, use the WS error term. 21 of 34 what is the follow up strategy recommended by Keppel? test interaction. Significant? test simple effects. Significant? test simple comparisons. 22 of 34 what is within cell error? an average of all the variances from all the cells in the design, must be calculated. 23 of 34 why is statistical significance (p value) not a good measure of treatment effects alone? F statistics are dependent on sample size 24 of 34 what do effect sizes measure? the strength of an effect 25 of 34 What is the difference between eta and omega effect sizes? use eta when only referring to your own sample, use omega when you want to generalise 26 of 34 what is the difference between original and partial effect sizes? use partial when there is a multi-factorial design 27 of 34 what is a type 1 error? false positive 28 of 34 what is a type 2 error? false negative 29 of 34 what does alpha indicate? the chance of making a type 1 error 30 of 34 what does beta indicate? the chance of making a type 2 error 31 of 34 what is the relationship between power and beta? power = 1-beta 32 of 34 what are 3 dangers of underpowered studies? statistically significant results found less, non-significant results can't be interpreted, and inconsistent results within a field. 33 of 34 to ensure power in a study, what is the recommendation? calculate necessary sample size in advance, using alpha, power, and effect size as input. 34 of 34
Comments
No comments have yet been made