general exam questions

Question 1

what are the important similarities between analysis of variance and regression?

Answer 1

• both models take into account the random error association with each observation
• total variance is partitioned into different components

Question 2

What is R equivalent to?

Answer 2

the zero-order correlation of observed scores with predicted scores

Question 3

multicollinearity is associated with —- in multiple regression

Answer 3

increased type 1 and type 2 error

Question 4

What is a covariate in ANCOVA

Answer 4

A covariate is a continuous variable that is not the main independent variable of interest but has the potential to influence the dependent variable. Covariates are included in ANCOVA to control for their effects and reduce the potential confounding or variability they introduce. It helps to isolate the effects of the main independent variable on the dependent variable.

Question 5

What is one important benefit of using eta-squared rather than partial eta-squared as an effect size measure in ANOVA?

Answer 5

Eta squared is the proportion of the total variance in the dependant variable that is accounted for by an independent variable, so the impact of different effects within a study can be meaningfully compared

Question 6

A developmental psychologist sought to explore the effects of type of observer and type of task on performance in children. The children were classified into one of two different observer conditions (father versus stranger) with 32 children
in each condition. All children were then tested on each of two educational tasks (reading and math). Were participants crossed with type of observer or type of task?

Answer 6

Participants is crossed with type of observer.

Here is why:

a. Type of observer is a fixed factor: True. The type of observer (father versus stranger) is deliberately manipulated by the researcher.

b. Participants is crossed with type of task: True. Participants (children) experience both types of tasks (reading and math).

c. Participants is crossed with type of observer: False. Participants are not crossed with the type of observer; they are assigned to one specific observer condition.

d. Type of task is a fixed factor: True. The type of task (reading and math) is deliberately manipulated by the researcher.

Question 7

A social psychologist sought to explore the effects of type of observer and type
of task on expressed frustration in children given a pair of two problem-solving tasks. The children were classified into one of two different observer conditions (mother versus stranger) with 10 children in each condition. All children were then tested on each of three problem-solving tasks (easy, moderate, and difficult), and expressed frustration was assessed.
To avoid problems that might arise from violations of assumptions, if the Type of Observer x Type of Task interaction were significant, the researcher would usually test the simple effects of type of task using:
a. the data for all groups combined
b. the pooled error term
c. a separate error term for each simple effect
d. none of the above

Answer 7

C. a seperate error term for each simple effect. This is because you need to account for specific variance

Question 8

If variable A is predicted from variable B, and variable C is predicted separately from variable B, and the residual variance in A is correlated with the residual variance in C, this correlation is:

a. a partial correlation
b. a semi-partial (part) correlation
c. a multiple correlation
d. a zero-order correlation

Answer 8

A partial correlation

A partial correlation measures the relationship between two variables while controlling for the influence of one or more other variables. In the given scenario, the residual variance in variable A is correlated with the residual variance in variable C, indicating a relationship between the two variables after accounting for their shared prediction from variable B. Therefore, it is a partial correlation.

A semi partial correlation measures the unique association between 2 variables while controlling for the shared variance explained by a third variable. In this case, variable B serves as the common predictor for A and C

Question 9

What is one advantage of using standardised regression coefficients (Beta), rather than unstandardised coefficients?

Answer 9

Standardised coefficients all use the same metric scale, so we can compare the coefficients associated with each predictor within one regression equation

Question 10

In moderated multiple regression, what is required to demonstrate that an interaction is present?

Answer 10

The product term between the key predictor and moderator is related to the criterion at the final step of the model

Question 11

In step 3 of a hierarchal regression analysis, what is the difference between R2 and R2 change?

Answer 11

R2 at step 3 reflects total variance explained by all the predictors entered at steps 1, 2 and 3, whereas R2 change at step 3 reflects the variance explained by the step 3 predictors (beyond those of steps 1 and 2).

Question 12

What is one key difference between analysis of an interaction using MMR analyses, and analysis or an interaction using ANOVA?
(hint: about continuous and categorical variables)

Answer 12

Interactions in MMR analyses can include continuous as well as categorical variables, whereas interactions in ANOVA can only include categorical variables

Question 13

What is the best way to describe the distinction between experimental designs and correlational designs?

Answer 13

experiments involve random assignment, whereas correlational studies so not

Question 14

if you have one primary focal IV, the advantages of using a three-way factorial design (compared to a one-way ANOVA) include:

Answer 14

information about generalisability across levels of moderators

In a three-way factorial design, you can examine the interaction effects between the primary focal IV and two other variables (moderators). This allows you to understand how the relationship between the primary focal IV and the dependent variable may differ or be influenced by different levels of the moderators. It provides valuable insights into the generalizability and robustness of the findings across different conditions.

Question 15

what is a key advantage of repeated-measures designs compared to between-group designs?

Answer 15

fewer statistical assumptions - same participants doing each testing variable

Question 16

In a fully-crossed factorial design investigating the impact of connectedness and impulsivity on substance abuse, a significant two-way interaction was observed. What does this suggests about the relationship between connectedness and substance abuse?

Answer 16

That the effect of connectedness on substance abuse depends on impulsivity.

Question 17

Which statement about the implications of a significant 2-way interaction in a between-groups ANOVA involving stress and drug dosage is FALSE?

A significant interaction tells us that the cell means will provide a more accurate
account of the treatment effects than the marginal means

A significant interaction tells us that the effects of stress differ, depending on which level of drug dosage we consider

A significant interaction tells us that the main effects may portray a more accurate picture of the IVs’ effects than the simple effects

A significant interaction tells us that the simple effects for drug dosage will differ, depending on which level of stress we consider

Answer 17

FALSE: A significant interaction tells us that the main effects may portray a more accurate picture of the IVs’ effects than the simple effects.

Question 18

what is the cell mean equation?

Answer 18

cell mean = grand mean + treatment effect for A + treatment effect for B

Question 19

describe what moderated and mediated mean

Answer 19

moderated = when the relationship of the IV and DV changes depending on a 3rd variable (a moderator).

mediated = when one variable (a mediator) explains the relationship between two other variables (IV and DV)

Rule of thumb –> moderator focusses on the 3rd variable, mediator focusses on understanding

Question 20

explain a between groups design, within-participants design, and a mixed design:

Answer 20

A between-groups design, also known as an independent groups design, involves separating participants into different groups or conditions. Each group experiences only one level of the independent variable, and the groups are compared to assess the effects of the independent variable on the dependent variable.

A within-participants design, also called a repeated measures design involves using the same participants across different conditions or levels of the independent variable. Each participant experiences all levels of the independent variable, and their performance or response is measured and compared within the same individual.

A mixed design combines elements of both the between-groups design and the within-participants design. In a mixed design, one or more independent variables are manipulated between groups, while at least one independent variable is manipulated within participants. This allows researchers to examine both between-group and within-group effects simultaneously.

Question 21

A researcher seeks to increase the power of their design by blocking participants on baseline depression and then randomly assigning them within each depression block to treatment conditions. If baseline depression affects the DV, then the inclusion of baseline depression
as a factor in the study design, and analysis as a factorial design rather than as a one-way design, will:

Answer 21

• reduce the error term used to evaluate the treatment main effect, allow the researcher to see if the effects of treatment generalise across levels and increase error df

Question 22

If a researcher uses a blocking variable that is strongly associated with the dependent measure, they can expect an increase in…?

Answer 22

An increase in power

Question 23

WG variance describes the deviations of ———- scores from the ——– means, whereas BG variance describes the deviations of ——– means around the ——– mean

Answer 23

WG variance describes the deviations of INDIVIDUAL scores from the GROUP means, whereas BG variance describes the deviations of GROUP means around the GRAND mean

Question 24

In ANCOVA, when we compare the adjusted means for the levels of the focal IV instead of the observed means, to test the main effect of focal IV, we _______ .

assume that in the population, the covariate means are the same across levels of the focal IV

increase Type I error

assume that in the population, the covariate and the focal IV are significantly correlated

All of the other answers are true.

Answer 24

assume that in the population, the covariate means are the same across levels of the focal IV

Question 25

if we observe a high multiple correlation (R) in a regression model, it is likely that the predictors have high/low validities and high/low collinearities

Answer 25

high, high

Question 26

What is the difference between a simple simple effect and a simple simple comparison?

Answer 26

A simple simple effect is used to follow up significant simple interactions, whereas a simple simple comparison is used to follow up significant simple simple effects that have more than 2 levels

Question 27

In ANCOVA, when we compare the adjusted means for levels of the focal IV instead of the observed means, to test the main effect of the focal IV, we assume —-?

Answer 27

Assume that in the population, the covariate means are the same across the levels of the focal IV

Question 28

In a regression analysis the observed variability in the criterion, Y , is partitioned into two components, SSregression and SSresidual. What does SSregression (the variability due to the regression) represent?

Answer 28

The sum of the squared deviations of predicted scores (y hat) from the mean of Y

Question 29

A relationship researcher hypothesises that receiving unexpected gifts increases
satisfaction with a romantic relationship, especially in a long relationship, and especially
for young participants. In the study, young versus old participants read scenarios in which
they imagined they do or do not receive an unexpected gift, in a relationship which is
either 6 weeks or 6 years old. The dependent measure is imagined satisfaction with the
relationship. There is a significant three-way interaction. Which of the following results
for follow-up tests would support the hypotheses?

a) Simple interaction of gift receiving x duration of relationship significant for young
participants but not for old.
b) Simple simple effects of age significant at each level of gift receiving and duration of
relationship.
c) Simple simple effects of gift receiving significant at each level of age and duration of
relationship, but effect sizes larger for young participants and longer relationships.
d) Simple simple effects of duration of relationship only significant for young
participants in the receiving gift condition.

Answer 29

c) Simple simple effects of gift receiving significant at each level of age and duration of relationship, but effect sizes larger for young participants and longer relationships.

The researcher hypothesized that receiving unexpected gifts would increase satisfaction with a romantic relationship, especially in a long relationship and particularly for young participants. The significant three-way interaction suggests that the effects of gift receiving, age, and duration of relationship together influence imagined satisfaction with the relationship.

In this case, the significant simple simple effects of gift receiving at each level of age and duration of relationship support the hypothesis. Additionally, the larger effect sizes observed for young participants and longer relationships further support the notion that receiving unexpected gifts has a stronger impact on satisfaction in these specific conditions.

Question 30

In a 3-way factorial ANOVA, a significant omnibus interaction effect of Factor 2 x Factor 3 means that?

The effect of Factor 3 depends on the effect of Factor 2, but differently depending on the level of Factor 1.

Disregarding Factor 1, the effect of Factor 2 changes at different levels of Factor 3, and/or the effect of Factor 3 changes at different levels of Factor 2.

The effect of Factor 3 is the same at different levels of Factor 2 if the interaction is ordinal.

The effect of Factor 2 is the same at each level of Factor 3, and the effect of Factor 3 is the same at each level of Factor 1.

Answer 30

Disregarding Factor 1, the effect of Factor 2 changes at different levels of Factor 3, and/or the effect of Factor 3 changes at different levels of Factor 2.

Question 31

what is a two-way interaction

Answer 31

a two-way interaction refers to the simultaneous influence of two independent variables on a dependent variable. It occurs when the effect of one independent variable on the dependent variable differs across the levels of another independent variable. In other words, the relationship between the independent variables and the dependent variable is not the same for all combinations of the levels of the variables.

Question 32

in a two-way S—- P— ANOVA design, the number of error term is —

Answer 32

Split-plot, 2

Question 33

Compared to a between-groups ANOVA design with the same N, a within-groups ANOVA design should deliver higher/lower type 1/type 2 error?

Answer 33

Lower type 2 error

(type 2 error occurs in hypothesis testing when we fail to reject a null hypothesis that is actually false. In other words, it is the error of not detecting a true effect or relationship when it exists.)

Question 34

A researcher conducts a 3 x 2 repeated-measures ANOVA, with Factor A having 3 levels and Factor B having 2 levels. What is the correct error term for Factor B?

Answer 34

The interaction of the participant factor with Factor B.

In a repeated-measures ANOVA, the error term for Factor B is the interaction of the participant factor (P) with Factor B. It captures within-participants variability and accounts for individual differences. It helps determine if the observed differences in Factor B are statistically significant or due to random variation.

Question 35

A researcher conducts a natural groups study comparing three groups’ performance on an academic test. The observed means are nearly identical. She covaries out IQ, and discovers that the adjusted means are very different and the effect is significant. What can she can conclude?

Answer 35

There are group differences on the covariate

Question 36

In a 2 x 4 between-subjects design, based on 120 subjects, df for the interaction term will be:

Answer 36

3, 112

Question 37

The significant main effect of A in a 3-way between-subjects design makes you certain about which of the following?
- The B x C interaction is irrelevant
- Disregarding B and C, the means of the levels of A are not equal
- There is no effect of B or C on A
- The effect of A is the same at all combinations of B and C

Answer 37

disregarding B and C, the means of the levels of A are not equal

Question 38

Which of the following is NOT expected in a blocking design where A is the
blocking factor and B is a treatment factor?
- A main effect of A
- A main effect of B
- An interaction of A and B
- None of the above (all are expected)

Answer 38

An interaction of A and B

Question 39

In a 3 x 3 within-subjects design, a researcher calculates epsilon (ε) for A, for B, and for the AxB interaction to be .60, .40, and .20 respectively. In which case will the use of the epsilon adjustment make the greatest difference to the numerator
and denominator df for the critical F?

• In the test of the A main effect
• In the test of the B main effect
• In the test of the AxB interaction
• It is impossible to say based on this information

Answer 39

In the test of AxB interaction

Question 40

In a factorial ANOVA, if group differences differ in magnitude rather than in direction, what kind of interaction are you most likely to have?

Answer 40

Ordinal

Question 41

When considering a within-subjects design, which of the following problems would be the hardest to solve?

Answer 41

Methodological issues (sequencing effects)

Question 42

list 2 things about omega squared:
it is a measure of —–
it describes —–

Answer 42

it is a measure of effect size
it describes the proportion of the variance in the DV in the population that is accounted for by the effect

Question 43

in a fully within-participants designs, the error term used for any effect is?

Answer 43

equal to the interaction between treatment effect and the random factor participants

Question 44

what is the differences between a main effect and a simple effect?

Answer 44

A main effect is an omnibus test whereas a simple effect is not

A main effect compares marginal means whereas a simple effect compares cell means

Question 45

Simple effects may m—— or q—— a main effect

Answer 45

moderate or qualify

Question 46

when testing a simple effect to follow up a two-way interaction in between-groups ANOVA, it is as if you:

Answer 46

Run a one-way ANOVA with the same error term as the original factorial ANOVA

Question 47

To estimate the correlation between variables X and Y with the effect of variables Z removed from both X and Y, one would use?

Answer 47

Part (semi partial) correlation