Research Methods IV

A one way ANOVA examines the difference between four (or more) levels/groups in one variable, whereas a t-test only examines the difference between three levels/groups in one variable.

A one way ANOVA examines the difference between three (or more) levels/groups in one variable, whereas a t-test only examines the difference between two levels/groups in one variable

A one way ANOVA examines the difference between three (you can't examine more than three) levels/groups in one variable, whereas a t-test only examines the difference between two levels/groups in one variable.

A one way ANOVA examines the difference between two (or more) levels/groups in one variable, whereas a t-test only examines the difference between three levels/groups in one variable.

First you have the residual sum of squares (SSR), which is made up of the model sum of squares (SSM) and the total sum of squares (SST). SSM tells us how much variance is explained by our hypothesis or model.

First you have the model sum of squares (SSM), which is made up of the total sum of squares (SST) and the residual sum of squares (SSR). SST tells us how much variance is explained by our hypothesis or model.

First you have the total sum of squares (SST), which is made up of the model sum of squares (SSM) and the residual sum of squares (SSR). SSM tells us how much variance is explained by our hypothesis or model.

First you have the total sum of squares (SST), which is made up of the model sum of squares (SSM) and the residual sum of squares (SSR). The none of the sum of squares values tell us how much of the variance is explained by our hypothesis.

First, total sum of squares (SST) shows the total variation in the data, this is made of the sum of squares for within participant variation (SSW) and the sum of squares for between participant variation (SSB). In the SSB we can further split the variation into the model sum of squares (SSM) and the residual sum of squares (SSR). SSR tells us how much variance is explained by our hypothesis.

First, total sum of squares (SST) shows the total variation in the data, this is made of the model sum of squares (SSM) and the residual sum of squares (SSR). SST tells us how much variance is explained by our hypothesis.

First, total sum of squares (SST) shows the total variation in the data, this is made of the sum of squares for between participant variation (SSB) and the sum of squares for within participant variation (SSW). SSW tells us how much variance is explained by our hypothesis.

First, total sum of squares (SST) shows the total variation in the data, this is made of the sum of squares for between participant variation (SSB) and the sum of squares for within participant variation (SSW). In the SSW we can further split the variation into the model sum of squares (SSM) and the residual sum of squares (SSR). SSM tells us how much variance is explained by our hypothesis.

Because the sum of squares are influenced by the number of scores summed.

Because the sum of squares are influenced by the degrees of freedom but mean squares are not.

This is a trick question - we do use the sum of squares, its the mean squares we don't use.

Because the sum of squares are influenced by the F ratio.

That our MSR is larger than our MSM. This means that there has been an effect of the variable between the groups/within the group - although we don't know if this is significant.

That our MSM is larger than our MSR. This means that there has been an effect of the variable between the groups/within the group - although we don't know if this is significant

That our MSM is larger than our MSR. This means that there has been an effect of the variable between the groups/within the group - and we know that this is significant.

That our MSM is larger than our MSR. This means that there has not been an effect of the variable between the groups/within the group - although we don't know if this is significant.

Sphericity

Interval level data

Normal distribution

Homogeneity of variance

Interval level data

Independence of groups

Normal distribution

Sphericity

Welch

Brown-Forsythe

Tukey

Bonferroni

Mauchly

Levene

Lower-bound

Greenhouse-Geisser

Yes, but only for repeated measures - you should calculate it yourself for independent measures.

Yes, for both

Yes, but only for independent measures - you should calculate it yourself for repeated measures.

No.

That this experiment has 3 levels. Two of them have 3 IVs and one has two.

That this experiment has 3 IVs. Three of them have 3 levels.

That this experiment has 3 IVs. Two of them have 3 levels and one has two.

That this experiment has 2 IVs. Two of them have 3 levels

We can see the main effects of each DV and how they interact.

We can see the main effects of each DV.

We can see the main effects of each IV.

We can see the main effects of each IV and how they interact.

The variance explained by SSM is made up of only the SS for each variable.

The variance explained by SSM is made up of the SS for each variable plus the SS for the interactions.

The variance explained by SSM is made up of the MS for each variable plus the MS for the interactions.

You cannot further split the variance explained by SSM.

We collapse across SSB and look only at the scores split by the stats knowledge variable as compared to the participants' means

We collapse across SSB and look only at the scores split by the stats knowledge variable as compared to the grand mean.

We collapse across SSA and look only at the scores split by the psych knowledge variable as compared to the grand mean.

SSA is the same as SSM in this context.

After calculating SSA and SSB then the remaining variance accounted for by SST is the variance from SSAxB. This is the interaction between the two variables.

After calculating SSA and SSB then the remaining variance accounted for by SSM is the variance from SSAxB. This is the interaction between the two variables

You do not get SSAxB in independent samples factorial ANOVA.

After calculating SSA and SSB then the remaining variance accounted for by SSM is the variance from SSAxB. This is the main effect of the two variables.

When both DVs have a main effect.

When both IVs have a main effect.

When the effect of one IV on the DV is dependent on another IV.

When the effect of one DV on the IV is dependent on another DV.

Sphericity - the Mauchly's test should not be significant

Sphericity - the Mauchly's test should be significant.

Homogeneity of variance - the levene's test should be significant.

Homogeneity of variance - the levene's test should not be significant.

Because this is the best way to explore the differences between the different levels in an interaction.

We don't - multiple comparisons tells us enough.

Because we want to examine the differences between the IVs.

Trick question - this is only used to examine main effects.

Because we need to know how much variance is explained by the SSR output, which is part of the variance explained by SSM.

Because we don't just need to know how much variance is explained by the model but whether each individual variable is a explaining a significant amount of variance.

Because we don't just need to know how much variance is explained by the model but whether each individual variable and their interactions is a explaining a significant amount of variance.

Trick question - we only interpret the SSM line of output in factorial ANOVA.

We split it into SSA and SSB.

We don't split the variance beyond SSM.

We split it into SSA and SSAxB.

We split it into SSA, SSB and SSAxB.

Into SSRA and SSRB.

We don't partition the variance of SSR.

Into SSRA and SSRAxB

Into SSRA, SSRB and SSRAxB.

MSA/MSR

MSA/MSRA

MSA/MSB

MSA/dfA

Homogeneity of Variance - Levene's test should be significant

Sphericity - Mauchly's test should not be significant

Sphericity - Mauchly's test should be significant

Homogeneity of Variance - Levene's test should not be significant

Day 3 TimeMeal 2

TimeDay 6

TimeMeal 2 Day 3

TimeMeal 3 Day 2

For a main effect: IV 1 will have the same effect on the DV in all levels of IV 2
In an interaction: IV 1 will only effect the DV in specific levels of IV 2

Main effects and interactions are the same.

For a main effect: DV 1 will have the same effect on the IV in all levels of DV 2
In an interaction: DV 1 will only effect the IV in specific levels of DV 2

For a main effect: IV 1 will only effect the DV in specific levels of IV 2

A design which has 2+ IVs. Both of which is between subjects.

A design which has 2+ IVs. At least one of which is between subjects and one within subjects

A design which has 3+ IVs. At least one of which is between subjects and one within subjects.

A design which has 2+ IVs. Both of which is within subjects

Only Sphericity which applies when the within subjects IV has 3+ levels as with repeated measures designs that are not mixed.

Only Sphericity which applies when the within subjects IV has 2+ levels as with repeated measures designs that are not mixed.

Both, but Sphericity only applies when the within subjects IV has 3+ levels as with repeated measures designs that are not mixed.

Only Homogeneity of Variance.

The same as for repeated measures ANOVA but we add in the between subjects variable at the 'define' stage.

The same as for repeated measures ANOVA but we add in the between subjects variable at the 'Post Hoc' stage.

Mixed ANOVA uses it's own process which is entirely different from repeated and independent measures ANOVA.

The same as for independent measures ANOVA but we add in the between subjects variable at the 'define' stage.

Are more conservative

Are less likely to accept the alternative hypothesis

Have less statistical power

All of the above

Wilcoxon test

Kruskal-Wallis test

One-way ANOVA

Friedman test

Compare groups using the Mann Whitney test

Compare groups using the Wilcoxon test

Compare groups using a t-test

None of the above. Post-hoc analyses cannot be carried out with non-parametric data

You need a computer to calculate them

They have many assumptions to meet

They are easy to calculate

They are very powerful

Mann-Whitney

Friedman

Wilcoxon

Kruskal-Wallis

Throw out the tied data

Assign any tied values the average of the ranks they would have received had they not been tied

Use a random number table to assign ranks to the tied values

Assign any tied values the lower of the ranks they would have received had they not been tied

3.5

4

4.5

6

Mann-Whitney

Friedman

Wilcoxon

Kruskal-Wallis

Larger

Smaller

More negative

H is not relevant for the Kruskal-Wallis test

Homogeneity of variance

Homoscedasticity

Homogeneity of sample size

Homogeneity of regression slopes

Used to correct for unmeasured variables that could affect the outcome variable

A method of portioning the results of ANOVAs to investigate where differences between independent variables lie

A robust version of ANOVA

An extension of ANOVA that partials out the effect of other measured variables

Attachment style

Secure attachment

Depression

Relationship satisfaction

Does relationship satisfaction have a significant effect on the relationship between attachment and depression?

What would the means of the groups be on relationship satisfaction if their levels of depression were constant?

What would the mean relationship satisfaction be if levels of depression were constant?

What would the mean depression score be for the three groups of attachment styles if their levels of relationship satisfaction were constant?

There could be more than three groups

It is likely that there will be a linear association between depression and relationship satisfaction

It is likely that the regression lines will be parallel

Depression might also be related to attachment style

Variation between groups; variation within groups

Variation within groups; variation between groups

Variation between groups; total variation

Total variation; variation between groups

A variable that is related to the dependent variable

A variable that is directly influenced by the dependent variable

A variable that is related to the independent variable

A variable that is unrelated to all other variables in the study

Caffeine

Score on the first memory test

Score on the second memory test

There was no covariate

To test for multivariate normality

To test for a plausible linear combination of the dependent variables

To test for homogeneity of variance

To test for homogeneity of covariance matrices

You could conduct separate Bonferroni-corrected ANOVAs on each dependent variable

There is no added value in performing discriminant function analysis

You could conclude that all groups differ significantly

None of the answers

The DVs are compiled into a linear combination

The DVs are entered sequentially into a model

The DVs are entered stepwise into an analysis

The DVs are standardized and summed

You have 30 participants per group in your between-participants design

You have equal numbers of participants and it is a large sample size

You have normally distributed dependent variables and all linear combinations of the dependent variables

Box's M has an associated p-value of < 0.05 and you have unequal sample sizes

Several dependent variables

Several independent variables

Nonmetric independent variables

Data with large error terms

Dependent variables are combined into discriminant function variates

It involves the computation of F values

Calculations are performed on matrices which contain cross-products

All of the options

There is a significant multivariate difference between the drug groups

Significant univariate differences are only present for psychiatrist ratings of mood

Self-report ratings of mood are significantly higher than psychiatrist ratings of mood

All of the answers are true

There may have been individual differences in mood at the beginning of the study

There was only one independent variable

The dependent variables were measured on different scales

All of the answers

Pillai’s Trace

Wilks’ Lambda

Roy’s Largest Root

They are all equal

6

5.5

4.5

4

Sleep

Exercise levels

Mood

All of the answers

Partial eta squared (η2p) reflects the proportion of variance that a variable explains that is not explained by other variables in the analysis

Means of two groups

Statistical power of a test and it's ability to find an effect, if one exists

The estimation of the effect size in the population

Does not tell us how important the effect is

Arbitrary, but encourages all or nothing thinking

Greatly influenced by sample size (as sample increases, the critical value decreases)

Is an objective and usually standardised measure of the magnitude of the observed effect

Total cross-product

Cross-product due to the mode

Residual cross-product

Reverse Product