# Research Methods and Statistics

What are the assumptions of a dependent ANOVA?
Interval/ratio data, normal distribution, homogeneity of variance
1 of 39
What do you do when just homogeneity of variance is violated when using a dependent ANOVA?
Use sphericity / homogeneity of covariance (MAUCHLY'S TEST) and if this is less than 0.05 then it has been violated and Greenhouse Geisser correction is used instead.
2 of 39
What if other assumptions are violated using dependent ANOVA?
Use Friedman's non-parametric test
3 of 39
What Post-Hoc test do you use with a dependent ANOVA?
Bonferroni
4 of 39
When do you use a dependent ANOVA?
When you have a within subjects design (repeated measures) where the same group of people take part in different conditions in the study.
5 of 39
What Post-Hoc test do you use with an independent ANOVA?
Tukey
6 of 39
When do you use an independent ANOVA?
When you have a between subjects design, where different people take part in different conditions of a study.
7 of 39
What do you do when the assumptions of an independent ANOVA are violated?
Use a non-parametric Kruskal-Wallis test
8 of 39
What are the assumptions of an independent ANOVA?
normal distribution, interval/ratio data, homogeneity of variance
9 of 39
How do you report an ANOVA test?
F=(effect df, error df)=?, MSE=?, p=?
10 of 39
How do you report Friedman's test?
X2(df)=?, p=?
11 of 39
How do you report a Kruskal-Wallis test?
X2(df,n=?)=?, p=?
12 of 39
What are the assumptions of a t test?
normally distributed, DV must be interval/ratio, each data point must be independent, and homogeneity of variance (less than .05 it has been violated)
13 of 39
If assumptions aren't violated and there is a between subjects design, what t test do we use?
Independent t test
14 of 39
If assumptions aren't violated and there is a within subjects design, what type of t test do we use?
Paired sample t test
15 of 39
When assumptions are violated and there is a within samples design, what non-parametric test do we use?
Wilcoxon
16 of 39
When assumptions are violated and there is a between subjects design, what non-parametric test do we use?
Man-Whitney
17 of 39
How do we report t tests?
t(df)=?, p=?
18 of 39
How do we report a Wilcoxon test?
W=?, z=?, p=?
19 of 39
How do we report a Man-Whitney test?
U=?, z=?, p=?
20 of 39
What is the difference between correlations and regression?
Regression also allows the prediction of the DV from the IV. Correlations just examine the relationship between two continuously valued interval/ratio variables.
21 of 39
What is r?
r is the correlation coefficient, and its a measure of strength and direction of the relationship of the variables
22 of 39
What are the assumptions of Pearson's r?
the relationship of the variables must be linear, must be normally distributed, interval/ratio data, and no major outliers
23 of 39
What do you use when the assumptions are violated?
Spearman's rho
24 of 39
How do you report a correlation?
r=?, p=?
25 of 39
What are the assumptions of a regression?
data must interval/ratio, relationship must be linear, normality of residuals, homoscedasticity
26 of 39
How do you report a regression?
F(effect df, error df)=?, MSE=?, p=? Adjusted R2=?
27 of 39
When do you use a Chi-square test?
To find the difference between two variables, when the data is categorical (nominal), and the design of the study is between subjects
28 of 39
What are the assumptions of a chi-square test?
Each participant, item or entity must only contribute to just one cell in the table, and more than 5 frequencies in each cell but 20% of cells are allowed to be less
29 of 39
How do you report a chi-square test?
X2=?, df=?, p=?
30 of 39
What test is used instead of chi-squared, when there are repeated measures?
McNamer (can only be used for 2x2 table)
31 of 39
What is an alternative hypothesis?
The idea that two variables are related in some way
32 of 39
What is a null hypothesis?
The idea that two variables aren't related
33 of 39
What is a type I error?
We think there is a relationship but there isn't (wrongly reject null hypothesis)
34 of 39
What is a type II error?
We think there isn't a relationship between the two variables when there is, but we missed it (wrongly retain the null hypothesis)
35 of 39
What does p have to be in order to be significant and therefore be less than alpha?
0.05
36 of 39
Why is p
In order to minimise our odds of making a mistake, as wrongly rejecting the null hypothesis is BAD
37 of 39
What are z scores?
A way of comparing a data point to the distribution it comes from
38 of 39
How are z scores calculated?
individual score-mean, then divide this by the SD
39 of 39

## Other cards in this set

### Card 2

#### Front

What do you do when just homogeneity of variance is violated when using a dependent ANOVA?

#### Back

Use sphericity / homogeneity of covariance (MAUCHLY'S TEST) and if this is less than 0.05 then it has been violated and Greenhouse Geisser correction is used instead.

### Card 3

#### Front

What if other assumptions are violated using dependent ANOVA?

### Card 4

#### Front

What Post-Hoc test do you use with a dependent ANOVA?

### Card 5

#### Front

When do you use a dependent ANOVA?