# Discovering Statistics; Chapter 11 Flashcards

1
Q

What is the most grand variance?

A

The variance of all observations.

2
Q

What does MSm represent?

A

The average amount of variation explained by the model.

3
Q

What does MSr represent?

A

The average amount of variance explained by extraneous variables.

4
Q

How do we test the assumption of homogeneity of variance with ANOVA?

A

Levene’s test; conducted on the absolute differences between the observed scores and the mean/median of the group from which each score came.

If significant (ie. p = .05) = assumption violated.

5
Q

What are the two corrected versions of the F-Ratio?

A

Brown-Forsythe F and Welch’s F (computed in SPSS).

6
Q

Can the F statistic be robust?

A

When all group sizes are equal, it can be robust to violations of normality.

7
Q

What is a planned contrast?

A

Used when you have specific hypothesis to test, to compare b coefficients.

8
Q

What is post-hot testing?

A

Used when you don’t have a specific hypothesis to test, to compare b coefficients.

9
Q

What are the different types of planned contrast?

A

1) Devation; first; compares each category (exc. 1st) to overall experimental effect; last; same but exc. last.
2) Simple; first; each category compared to 1st category; last; same but compared to last.
3) Repeated; each category compared to prev. category
4) Helmert; each catgeroy compared every subsequent category
5) Difference; each category compared to every previous category.

10
Q

What is polynomial contrast?

A

Tests for trends in the data; most basic for is the linear trend (no change in direction), but also have quadratic (1 change), cubic (2 changes) and quartic (3 changes).

11
Q

What are post-hoc tests?

A

Pairwise comparisons designed to compare all different combinations of treatment groups.

12
Q

Are post-hoc tests robust?

A

No, they perform badly under varying group sizes and population variables.

13
Q

In ANOVA testing, what is the R2 value usually referred to as?

A

Eta squared (n2)

14
Q

How do we report ANOVA?

A

F(DFm, DFr) = F value, p = ‘sig.’ value.