Flashcards
Experimental psychology
Cards in this deck (34)
- 1Mann-Whitney U testindependent samples t test
- 2Wilcoxon Signed Rank testpaired samples t test
- 3Kruskal-Wallis testone way between subjects anova
- 4Friedman testone way repeated measures anova
- 53 advantages of non parametric testsdon't need normal distribution
don't need interval or ratio data
simpler computations - 63 disadvantages of non parametric testsnot as powerful
loss of precision by changing into ranks
unable to detect interactions - 73 assumptions of non parametric testsindependence
data at least ordinal
distributions approx. the same - 8Powerprobability of finding something in the study when it is there
1 - beta - 9type I errorincorrectly rejecting null hypothesis
alpha - 10type II errorincorrectly retaining null hypothesis
beta - 113 ways to increase powerincrease alpha
increase difference between H0 and H1
increase standard deviation (increase n or decrease variance) - 123 values needed to estimate required sample sizealpha
power wanted
effect size wanted - 13sampling errordifference between sample mean and population mean
- 14degrees of freedom needed for z or t testt test
- 15parameter vs. statisticdata about population vs. data about sample
- 16MS errordoesnt depend on null being true
- 17MS treatmentassumes null is true
so uses CLT
contains between condition variance - 18x(ij)= m + t(j) + e(ij)one way anova
t = treatment (effect of group)
e = error (special) - 19x(ij)= m + π(i) + t(j) + e(ij)repeated measures anova
t = treatment (effect of group)
e = error (special)
π = constant associated with that person - 20x(ijk)= m + a(i) + B(ij) + aB(ij) + e(ijk)between groups factorial anova
a = effect of IV a
B = effect of IV b
aB = interaction - 21x(ijk)= m + a(i) + π(j/i) + B(k) + aB(ik) + πB(jk/i) + e(ijk)mixed anova
a = effect of IV a (between)
B = effect of IV b (within)
π = group specific, person specific variance - 223 low constraint research methodscase study
naturaliastic observation
archival - 232 assumptions of one sample t testscale
normal distribution - 24unique assumption for paired samples t testdifference scores are normally distributed
- 252 unique assumptions for independent samples t testindependence
homogeneity of variance - 26FMS treatment / MS error
- 27familywise error rateper comparison error rate (a) x number of comparisons (c)
- 28planned comparisons (a priori)linear constrasts
- 29post hoc testsscheffe, SNK
- 30shapiro wilk testtest of normality
- 31levene's testtest for homogeneity of variance
- 32simple effects and how to follow upa significant interaction
look at the effect of one IV on the DV at each level of the other IV separately - 33main effects and following upsignificant effect of either of the IVs in a factorial anova
if 2 levels: look at graph
if more levels: post hoc tests for anova - 34error in one war repeated measures anova is ___ than between groupssmaller
Information
- Description
- 3rd year Abnormal Psychology Flashcards on Experimental psychology, created by a.gar291 on 18/05/2016.
- Created by
- a.gar291
- Views
- 1
- Rating
More on Statistical Inference and Hypothesis Testing