PAL PSYU3349 Week 10

Paired and unpaired t-tests involve what type of ratio?

• Variance ratio

• Noise : signal ratio

• Signal : noise ratio

• Odds ratio

What is a purpose of random allocation?

• Ensure generalisation is possible

• Distribute confounding variables between groups

• Minimise carry-over effects

• Minimise participant bias

A larger Cohen's D for an un-paired t-test indicates what?

• Less overlap in scores between groups

• Greater overlap in scores between groups

• Greater likelihood of normally distributed means

• A weaker effect

What can we do to counteract carry-over/order effects in a within-subjects design?

• Random allocation to groups

• Random sampling

• Increase sample size

• Counterbalancing

Which of the following does not increase the signal : noise ratio for an unpaired t-test?

• Decreasing within-group variation

• Increasing SD of differences

• Increasing sample size

• Increasing difference between group means

Why shouldn't we use an unpaired t-test for within-subjects designs?

• Because the assumption of "independence between and within groups" would be violated

• Because the assumption of "normal distribution of within-subjects difference" would be violated

• Because of increased variance due to individual differences

• All of the above

How do we calculate the F-statistic for a main effect in a between-subjects ANOVA?

• F = mean_difference / SE_difference

• F = SS_group / df_group

• F = SS_group/ SS_error

• F = MS_group / MS_error