Chapter 8: Multivariate Correlational Research

Question 1

What are three techniques that researchers use to get closer to establishing a casual relationship?

Answer 1

longitudinal designs, multiple regression designs and pattern & parsimony approach.

Question 2

Longitudinal designs allow us to?

Answer 2

evaluate temporal precedence in data

Question 3

multiple regression designs help us to rule out what?

Answer 3

help us to rule out certain third variable explanations

Question 4

pattern & parsimony approach?

Answer 4

results of a variety of correlational studies all support a single causal theory

Question 5

Multivariate designs??

Answer 5

involve more than two measured variables

Question 6

A longitudinal design can provide evidence for temporal precedence by?

Answer 6

measuring the same variables in the same people at different points in time.

Question 7

Cross-sectional correlations?

Answer 7

test to see whether two variables, measured at the same point in time are correlated

Question 8

Autocorrelations?

Answer 8

evaluate the associations of each variable with itself across time….they determine the correlation of one variable with itself, measured on two different occasions.

Question 9

Cross-lag correlations?

Answer 9

show whether the earlier measure of one variable is associated with the later measure of the other variable….they address the directionality problem and help establish temporal precedence.

Question 10

What are three ways to interpret longitudinal designs?

Answer 10

cross-sectional correlations, autocorrelations and cross-lag correlations.

Question 11

By inspecting via cross-lag correlations we can investigate what?

Answer 11

change over time and therefore establish temporal precedence

Question 12

Do longitudinal designs help rule out third variables?

Answer 12

NO

Question 13

Significant relationships in longitudinal designs can help establish what?

Answer 13

covariance

Question 14

A longitudinal design can help researchers make inferences about temporal precedence by comparing what?

Answer 14

the relative strength of the two cross lath correlations, the researchers can sees which path is stronger.

Question 15

Why do we not just do an experiment?(2)

Answer 15

-commonly people cannot be randomly assigned to a variable
- could be unethical to assign some people to a condition
L> when an experiment is not practical or ethical a longitudinal correlational design is a good option

Question 16

Multiple regression can help rule out _____. They can address questions of_____.

Answer 16

• third variables

- internal validity

Question 17

Researchers are able to evaluate whether a relationship between two key variables still holds when they do what?

Answer 17

control for another variable.

Question 18

When researchers use regression they are testing what?

Answer 18

whether some key relationship holds true even when a suspected third variable is statistically controlled for.

Question 19

When researchers are using multiple regression they are studying how many variables?

Answer 19

3 or more

Question 20

What is the first step in multiple regression?

Answer 20

• choose a dependent variable (criterion variable)

L> this is the variable one is most interested in understanding or predicting

Question 21

In multiple regression the dependent variable is almost always where in the regression table?

Answer 21

specified in the top row or in the title of a regression table

Question 22

Aside from the dependent variable what are the other measured variables called in regression analysis?

Answer 22

predictor variables aka independent variables

Question 23

What does beta (B …standardized) show in regression analysis?

Answer 23

it reveals more than r does
L> positive: indicates a positive relationship between the predictor variable and the DV..when the other predictor variables are statistically controlled for.
L> negative: indicates a negative relationship between two variables when the other predictor variables are controlled for.
L> zero: not sig different from zero means that there is no relationship when other predictor variables are controlled for

Question 24

Betas in regression analysis denote ___ and ____ of some relationship. The higher beta is, the ___ the relationship is between that predictor variable and the DV.

Answer 24

direction and strength

stronger

Question 25

Betas change depending on?

Answer 25

other predictor variables are being used-bent controlled for-in the regression

Question 26

What does the b unstandardized coefficient show?

Answer 26

- sign +/- denotes a positive or negative association | - cannot compare two b coefficients within the same table ...you can compare betas.

Question 27

Why can you not compare two b coefficients within the same regression table?

Answer 27

they are computed from the original measurements of the predictor variables

Question 28

Why can you compare two beta coefficients within the same regression table?

Answer 28

- they are computed rom predictor variables that have been changed to standardized units.

Question 29

If a predictor variable shows a large b does it actually denote a stronger relationship to the dependent variable than a predictor variable with a smaller b ?

Answer 29

no

Question 30

Are there beta values for each predicator variable in a multiple regression table?

Answer 30

YES one for each

Question 31

Each beta value is given a ___ value

Answer 31

p

Question 32

P values indicate what about beta?

Answer 32

if it is statistically sig different from zero

Question 33

If p is less than 0.05 what does this mean to beta?

Answer 33

the relationship between the predictor and the DV is stat sig

Question 34

If betas are found to be statistically sig what does this mean about the relationship found in regression analysis?

Answer 34

that it is a replicable relationship

Question 35

Adding several predictors to a regression analysis can help with what?

Answer 35

answering two kinds of questions L> it helps control for several third variables at the same time L> by looking at the betas for all of the other predictor variables we can get a sense of which factors most strongly affecting the DV

Question 36

Popular press sources rarely discuss __,___ or ___ variables. However if you read carefully you can detect what?

Answer 36

betas, p values or predictor variables | - the tell tale signs that a multiple regression has been used.

Question 37

The term controlled for is a clue that the study used what ?

Answer 37

multiple regression

Question 38

The phrase taking into account means that researchers conducted what?

Answer 38

regression analysis

Question 39

The phrase after correcting for indicates that the researchers used ?

Answer 39

multiple regression

Question 40

The fact that researchers cannot control for variables they have not measured is why what kind of design is better than correlational designs?

Answer 40

- experimental designs L> random assignment would make the two groups likely to be equal on any possible third variable even the third variables the researchers did not happen to measure.

Question 41

Pattern and parsimony approach?

Answer 41

- a variety of correlational studies that all point in a single causal direction

Question 42

Parsimony?

Answer 42

the simplest explanation of a pattern of data the best explanation that requires making the fewest exceptions or qualifications

Question 43

If all predictions are tied back to one central principle this is a strong case for what?

Answer 43

parsimony?

Question 44

Mediating variable?(mediator)

Answer 44

- researchers propose a mediating step between two of the variables....mediating analyses often rely on multivariate tools such as regression analyses

Question 45

How does testing for a mediating variable work?

Answer 45

1. test for relationship c...associationg between the IV and DV 2. test for relationship a....IV and mediator 3. test for relationship b....DV and mediator 4. run a regression test using both the mediator and proposed IV to predict the DV to seed if relationship c goes away.....if the proposed mediator is truly the mediator in the test than when it is controlled for the relationship between the proposed IV and the DV will drop.

Question 46

Mediation tests often involve testing what?

Answer 46

the opposite pattern as well and comparing how well each pattern fits the data.

Question 47

Using statistic testings such as structural equation modelling with mediation models what can be compared?

Answer 47

two meditational models ...the reverse and the preferred one ...to see which fits the data best

Question 48

Whats the main difference between third variables and mediators?

Answer 48

- third variables are external to the two variables in the original bivariate relationship L> mediatiors: researchers propose this when they are interested in isolating which aspect of the causal variable is responsible for that relationship L> it is internal to the causal variable and often not considered a nuisance like a third variable

Question 49

Aside from internal validity what other validities should be interrogated in multivariate studies?

Answer 49

external, construct and statistical