Chapter 9 Multivariate Correlational Research

Question 1

multivariate designs involve what?

Answer 1

3(+) variables

Question 2

multiple regression

Answer 2

statistical analysis that measures the relationship b/w predicTor and criterion variables while CONTROLLING FOR OTHER PREDICTORS

Question 3

how is a third variable controlled for?

Answer 3

holding it constant while investigating the relationship b/w 2 other variables

Question 4

multiple regression can help address which validity?

Answer 4

internal

Question 5

does multiple regression establish temporal precedence?

Answer 5

no

Question 6

the overlap b/w 2 circles in a Venn diagram indicates what?

Answer 6

the degree of association b/w 2 variables

Question 7

1 circle in a Venn diagram represents what?

Answer 7

the variability of 1 variable

Question 8

what does it mean when we say we want a unique overlap b/w criterion and the predictor of interest variable?

Answer 8

we want a section where only these 2 variables are overlapping (w/o the 3rd variable)

Question 9

3 possible outcomes of multiple regression

Answer 9

1. predictor has no independent effect on the criterion variable
2. both the predictor and 3rd variable have an independent effect on the criterion variable while controlling for the other (each have a unique overlap w/ the DV)
3. the third variable has no independent effect on the criterion variable when controlling for the predictor

Question 10

does a p value or a beta give significance?

Answer 10

p value

Question 11

finding the unique overlap b/w IV and DV is a method of doing what?

Answer 11

controlling for the 3rd variable

Question 12

beta in regression analysis

Answer 12

“standardized regression coefficient”, the unique overlap b/w IV and DV w/o overlap from 3rd variables

Question 13

example of how beta can indicate strength and direction of relationship

Answer 13

negative beta = negative correlation

Question 14

out of a group of predictor variables, the one with the largest beta value will be the what?

Answer 14

strongest predictor

Question 15

multiple regression can help answer which 2 questions?

Answer 15

1. helps control for several 3rd variables at once
2. often examines the betas for all other predictors to understand which factors most strongly predict the DV

Question 16

b

Answer 16

unstandardized regression coefficient

Question 17

main differences b/w beta and b

Answer 17

beta: standardized, can compare within the same regression analysis (result table) but not across different ones
b: unstandardized, cannot compare within the same regression analysis

Question 18

b template

Answer 18

for every 1 unit change in X, there is a b units change in Y, holding third variables C, D and E constant

Question 19

beta template

Answer 19

this beta means that X is/is not associated with Y such that high/low scores on X go w/ high/low scores on Y, controlling for other predictors C, D and E.

Question 20

mediator

Answer 20

a variable that helps explain the relationship b/w IV and DV; an outcome of the IV responsible for the DV

Question 21

triangular diagram is an indicator of what?

Answer 21

mediation

Question 22

total effect (of mediation) =

Answer 22

direct effect + indirect effect

Question 23

what types of studies can we find mediation?

Answer 23

correlational and experimental

Question 24

can mediation imply causation?

Answer 24

yes

Question 25

similarities b/w mediators, moderators and 3rd variables

Answer 25

-involve multivariate research designs
-can be detected using multiple regression techniques

Question 26

longitudinal designs

Answer 26

study in which the same variables are measured in the same people at dif. points in time

Question 27

can longitudinal designs establish temporal precedence?

Answer 27

yes!

Question 28

can longitudinal designs establish causation?

Answer 28

not on their own!

Question 29

3 types of correlations in longitudinal studies

Answer 29

cross-sectional
autocorrelations
cross-lag

Question 30

cross-sectional correlation

Answer 30

whether 2 variables measured at the same point in time are correlated

Question 31

autocorrelations

Answer 31

the correlation of each variable w/ itself across different time points

Question 32

cross-lag correlations

Answer 32

correlation b/w an earlier measure of 1 variable q/ a later measure of the other variable (X time 1 w/ Y time 2)

Question 33

which correlations of longitudinal studies can establish temporal precedence if significant?

Answer 33

cross-lag correlations

Question 34

4 possible outcomes of longitudinal studies

Answer 34

1. significant correlation b/w earlier X and later Y
2. sig corr b/w earlier Y and later X
3. sig. corr. including both 1 and 2 (mutually reinforcing)
4. zero correlation

Question 35

which possible outcome of longitudinal studies is rare and why?

Answer 35

zero correlation, because longitudinal studies usually look at multiple variables so 99.9% of the time there’s some significant findings

Question 36

why do longitudinal studies lack internal validity?

Answer 36

they cannot automatically rule out 3rd variables

Question 37

advantages of longitudinal studies

Answer 37

may establish temporal precedence
may help study causation when its impossible/unethical to manipulate the IV (ex: parenting style/neglect)

Question 38

drawbacks of longitudinal studies

Answer 38

1. difficult to recruit people who agree to participate across a long time period
2. attrition (participants dropping out)
3. money and time consuming

Question 39

parsimony

Answer 39

extent to which a theory provides the simplest explanation

Question 40

pattern

Answer 40

same results are seen in similar research/theories

Question 41

methodological pluralism

Answer 41

using many methods and designs to test a theory

Question 42

pattern and parsimony are referred to as…

Answer 42

the producers of science