Exam Flashcards
What is the difference between MS treatment and MS error?
MS treatment assumes the null hypothesis is true
MS error is __________ condition variance (error) while MS treatment is __________ condition variance (error + treatment)
Within, between
If there is an effect of the treatment, the MS treatment will be _______ than MS error
Bigger
MS treatment also includes the effect of the ________
Treatment
If the null hypothesis is true then MS error and MS treatment will be ______
Equal
What are MS error and MS treatment?
Estimates of population variance
What does oX (sigma, subscript X Bar) represent?
Standard error of the mean
What are the basic assumptions shared by all three t-tests?
Data is normally distributed
Interval or ratio data
Assumptions of a one sample t-test?
Interval or ratio
Normally distributed
Assumptions of a paired samples t-test?
Interval or ratio
Normally distributed
Differences between pairs of scores are normally distributed
Assumptions of an independent samples t-test?
Interval or ratio
Normally distributed
Independence of scores
Homogeneity of variance
What is the test of homogeneity of variance?
Levene’s test
If Levene’s test is significant we report the _____ line
Bottom
What is alpha?
The chance of incorrectly rejecting a true null hypothesis
The chance of making a Type I error
The cutoff point against which the p-value is judged
What is power?
The chance of rejecting an incorrect null hypothesis (finding an effect when it is really there)
Describe the central limit theorem
Taking multiple random samples from a population, the more samples we take the more our data will approach the normal distribution
How many degrees of freedom do you report in an ANOVA? E.g. F (df)
2
What does ij refer to in a conceptual model?
Which person in which group
One way between groups ANOVA, conceptual model is made up of: (use formula sheet, use ij)
The overall mean, the effect of which group they’re in, plus a little bit of error due to difference between that persons score and the group mean
The one way of one way between groups ANOVA means…
One IV
What is the main issue with follow up tests for ANOVAs?
A high Type 1 error rate (family wise error)
Roughly how is family wise error rate for ANOVAs calculated?
Per comparison error rate x number of comparisons
When running follow up tests for a planned comparison ANOVAS, we use l________ c_______ which assign weights to each group to control for family wise error rate.
Linear contrasts
BMW
Between groups (Iv 2xlevels) Mann-Whitney or Wilcoxons
WW
Within groups (IV 2 levels) Wilcoxons signed ranks test
BKW
Between group (IV >2 levels) Kruskal Wallis test
WF
Within groups (IV >2 levels)
Fried mans test
Multiple comparisons tests for factorial ANOVAS
Two types
A priori - linear contrasts
Post hoc - SNK Sheffes Test
The standard deviation of the sampling distribution is also called the…
Standard error
Which ANOVAS is uncommon, we don’t need to know about for exam?
Repeated measures factorial ANOVAS
Non parametric tests to not involve estimating ________
Population parameters
Advantages of factorial ANOVAS
Generalisability of results
e.g. Not just whether drug is effective but what level of depression it is effective for
Interaction effects
e.g. CBT + drug
Economy
e.g. Look at several IVs in one study - cheaper!
What are the three types of factorial ANOVA?
Within
Between
Mixed
How many versions of MS treatment are there in a factorial ANOVA?
3 - one for each IV, and one for the interaction
How many MS errors do we calculate for a factorial ANOVA?
One
Which degrees of freedom are reported in an ANOVA? The IV or interaction and _______
Error (within)
When writing up the results of an ANOVAS, we must include what sentence to account for homogeneity of variance.
Levene’s test reveals that the assumption of homogeneity of variance was supported.
When writing up the results of an ANOVAS, we must include what sentence after we check histograms for normally distributed data?
The Shapiro -Wilk test and a visual inspection of the data confirmed the assumption of normality was supported
