2-way ANOVA & Beyond

Question 1

What 4 things are required for establishing causality?

Answer 1

Theoretical credibility, co-variation, elimination of confounds and temporal precedence.

Question 2

Describe necessary causation.

Answer 2

If you see Effect(B), you must be at level B.

Question 3

Describe sufficient causation.

Answer 3

If you see Effect(B), you may be at level B.

Question 4

Describe Simpson’s paradox.

Answer 4

Where a correlation seems to go one direction, but adding a factor means there are many correlations, possibly in a different direction.

Question 5

What is observed in 2-way ANOVA?

Answer 5

Grand mean + F1 effect + F2 effect + Interaction effect + error.

Question 6

What does adding a factor that matters do?

Answer 6

Decrease SS, df and error term.

Question 7

What does adding a factor that doesn’t matter do?

Answer 7

No decrease in SS, decrease in df, increase in error term and power, possible loss of significance.

Question 8

What is used to control for multiple t tests?

Answer 8

The overall effect size given by R squared.

Question 9

What is eta-squared?

Answer 9

The proportion of all variability explained the effects/interactions.

Question 10

What is partial eta-squared?

Answer 10

The proportion f the variance that could be explained, that the effect/interaction explains.

Question 11

What does including participants as a random factor allow us to do?

Answer 11

Account for variability in a dataset due to individual differences.

Question 12

Why do we calculate the average across participant means rather than the average across trials?

Answer 12

We usually want to talk about ”people on average” rather than the average trial.

Question 13

What makes up a nesting effect?

Answer 13

The mean for one level of nested factor 2 - (grand mean + one level of factor 1 effect).

Question 14

What is observed in a nested ANOVA?

Answer 14

Grand mean + F1 effect + F2(F1) effect + error.

Question 15

How can bivariate data be analysed?

Answer 15

Pearson’s correlation and simple linear regression/1-way ANOVA.

Question 16

How can multivariate data be analysed?

Answer 16

Partial and multiple correlation and multiple linear regression/n-way ANOVA/ANCOVA.

Question 17

What does partial correlation ask about?

Answer 17

The partial variance of X3.

Question 18

What does semi-partial correlation ask about?

Answer 18

The total variance of X3.

Question 19

Give 2 key aspects of partial correlation.

Answer 19

It is a “pure” measure uncontaminated by other variables and its uniqueness of variance makes it theoretically simpler.

Question 20

Give 2 key aspects of semi-partial correlation.

Answer 20

It has a more intuitive baselines and it allows easy comparison of coefficients because X3 is constant.

Question 21

How can ANCOVA be used to control for a variable?

Answer 21

You can add an extra variable, and use a regression line to estimate when X is the same.

Question 22

What is observed in an ANCOVA?

Answer 22

Grand mean + factor effect + participant effect + individual differences + noise.

Question 23

How can a covariant remove unwanted variability?

Answer 23

If some DV variability is attributed to the covariate it is removed from the error sum of squares.

Question 24

When can adding a covariate increase degrees of freedom?

Answer 24

When the covariate is not related to the DV.

Question 25

When can adding a covariate reduce power?

Answer 25

When it is co-linear with the first.

Question 26

What assumptions does ANCOVA share with ANOVA?

Answer 26

Normality of error distributions, homogeneity of error variance and independence of “error term.”

Question 27

What assumptions does ANCOVA share with regression?

Answer 27

Interval/ratio scale data and linear relationship.

Question 28

Give 3 reasons to use ANCOVA compared to ANOVA.

Answer 28

Accounts for more variability, “controls for” other variables and quantifies which subject characteristics are important.

Question 29

Give 3 things a good covariate should be.

Answer 29

Related to the dependent variable, unrelated to the other independent variables and unrelated to the other covariates.