Research Methods for the Behavioral Sciences 4th Edition by Charles Stangor – Test Bank

APPENDIX D

Multivariate Statistics

learning objectives

1. 1. Show what simultaneous, hierarchical, and stepwise multiple regression analyses are.
2. 2. Define a loglinear analysis.
3. 3. Show what statistical procedures are used to compare group means.
4. 4. Define multivariate statistics.
5. 5. Determine how factor analyses are used in research.
6. 6. Understand Multivariate Analysis of Variance (MANOVA) and canonical correlation.
7. 7. Define structural equation analyses.
8. 8. Determine what procedures are used to choose the appropriate statistical test for a given research design.

ESSAY QUESTIONS

1. 1. What is multiple regression, and when would it be used to analyze data? Differentiate the uses of simultaneous, hierarchical, and stepwise multiple regression.
2. 2. Outline the basic principles of means comparisons. When and why are means comparisons used, and what different statistical tests are used to compare means under different conditions?
3. 3. What is the general goal of multivariate statistics? What specific multivariate procedures are used to reach this goal?
4. 4. What is factor analysis? How is it used to analyze data? Differentiate exploratory and confirmatory factor analysis.
5. 5. What are structural equation analyses? In what cases might they be used to analyze data, and how are they tested for significance?

MULTIPLE-CHOICE QUESTIONS

Ans: c

Page: 375

Type: F

LO: 1

1. Which of the following is NOT a use of multiple regression analysis?

a. Reducing error by controlling for scores on baseline measures

b. Creating path-analytic diagrams

c. Conducting a covariance structure analysis

d. Predicting scores on a dependent variable given scores on a set of independent variables

Ans: a

Page: 375

Type: F

LO: 1

2. According to the multiple regression equation, the best prediction of the outcome variable (Ŷ) is equal to

a. the intercept plus the sum of all the predictor variables (X’s), each multiplied by their regression coefficients.

b. the sum of all the predictor variables (X’s), each multiplied by their regression coefficients.

c. the intercept minus the sum of all the predictor variables (X’ s), each multiplied by their regression coefficients.

d. the intercept minus the sum of all the predictor variables (X’s), each divided by their regression coefficients.

Ans: b

Page: 375

Type: F

LO: 1

3. In a multiple regression analysis, which of the following represents the linear relationship between each predictor variable and the outcome variable, controlling for each of the other predictor variables?

a. The intercept

b. The regression coefficient

c. The factor score

d. The multiple R

Ans: c

Page: 378

Type: F

LO: 1

4. In a multiple regression analysis, the __________ indicates the extent to which the predictor variables as a group predict the outcome variable.

a. factor scores

b. p-value

c. multiple correlation coefficient

d. regression coefficient

Ans: a

Page: 378

Type: F

LO: 1

5. The effect size statistic for a multiple regression is

a. R.

b. R squared.

c. the F value.

d. the p-value.

Ans: d

Page: 378

Type: F

LO: 1

6. In a multiple regression, R squared indicates

a. the effect size.

b. the test of the significance of R.

c. the residual degrees of freedom.

d. the proportion of the variance in the outcome variable that is accounted for by the predictor variables.

Ans: a

Page: 378

Type: F

LO: 1

7. Which of the following refers to a multiple regression analysis in which all of the predictor variables are entered in at the same time in order to predict the outcome variable?

a. Simultaneous multiple regression

b. Hierarchical multiple regression

c. Stepwise multiple regression

d. Covariance multiple regression

Ans: c

Page: 378

Type: F

8. Which of the following refers to a multiple regression analysis in which the predictor variables that produce the biggest increase in the multiple R are entered first?

a. Simultaneous multiple regression

b. Hierarchical multiple regression

c. Stepwise multiple regression

d. Covariance multiple regression

Ans: b

Page: 378

Type: F

LO: 1

9. A researcher has developed a theory about which predictor variable should be entered first, second, and third to predict the outcome variable. He will want to use which of the following?

a. Simultaneous multiple regression

b. Hierarchical multiple regression

c. Stepwise multiple regression

d. Overweighted multiple regression

Ans: a

Page: 379

Type: C

LO: 1

10. Which of the following can be used to control for the effects of potential confounding variables?

a. Multiple regression

b. Pearson correlation

c. Confirmatory factor analysis

d. Exploratory factor analysis

Ans: a

Page: 380

Type: F

LO: 2

11. A __________ analysis is a statistical procedure that is used when more than one nominal variable is used to predict a nominal outcome variable.

a. loglinear

b. contrast

c. chi-square

d. covariance

Ans: b

Page: 380

Type: F

LO: 3

12. What is the most general method of conducting means comparisons that have been planned a priori?

a. Loglinear analysis

b. Contrast analysis

c. Analysis of covariance

d. Multiple regression

Ans: a

Page: 380

Type: F

LO: 3

13. A means comparison is used

a. to test the differences between and among particular group means.

b. to determine which statistical procedure should be used to compare two or more means.

c. only to test the differences between group means that were planned before data collection.

d. only to test the differences between group means that were planned after data collection.

Ans: c

Page: 380

Type: F

LO: 3

14. A contrast analysis involves

a. computing a correlation coefficient.

b. comparing Pearson correlations.

c. computing an F value.

d. computing beta weights.

Ans: b

Page: 381

Type: F

LO: 3

15. Contrast weights are set by the researchers in order to

a. order the values of the means.

b. indicate how the group means are to be compared.

c. double-check that the F value is significant.

d. turn a correlational design into an experimental design.

Ans: d

Page: 381

Type: F

LO: 3

16. In a contrast analysis, the sum of the contrast weights

a. can vary from -1.00 to +1.00.

b. must equal -l.00.

c. must equal 1.00.

d. must equal zero.

Ans: b

Page: 381

Type: F

LO: 3

17. Means that are not being compared in a contrast analysis are given the contrast weight of

a. +1.00.

b. zero.

c. -1.00.

d. one-half.

Ans: c

Page: 382

Type: F

LO: 3

18. __________ are designed to control the experimentwise alpha level in means comparisons that are made after the data have been collected.

a. A priori means comparisons

b. Contrast analyses

c. Post hoc means comparisons

d. Contrast weights

Ans: c

Page: 382

Type: F

LO: 3

19. Which of the following means comparison analyses involves comparing the F comparison to a critical F value that is adjusted to take into consideration the number of possible comparisons?

a. Contrast analysis

b. Fisher least significant difference test

c. Scheffé means comparison test

d. Tukey honestly significantly difference test

Ans: b

Page: 382

Type: F

LO: 3

20. Which of the following procedures protects against increases in the experimentwise alpha by only allowing post hoc comparisons if the initial ANOVA F value is significant?

a. Contrast analysis

b. Fisher least significant difference test

c. Scheffé means comparison test

d. Tukey honestly significantly difference test

Ans: a

Page: 382

Type: F

LO: 3

21. Which of the following is NOT a post hoc means comparison?

a. A contrast analysis

b. The Fisher least significant difference test

c. The Scheffé means comparison test

d. The Tukey honestly significantly difference test

Ans: c

Page: 382

Type: F

LO: 3

22. Which of the following means comparison analyses is the most conservative, in the sense that it most reduces the likelihood of making a Type 1 error?

a. A contrast analysis

b. The Fisher least significant difference test

c. The Scheffé means comparison test

d. The Tukey honestly significantly difference test

Ans: d

Page: 384-387

Type: F

LO: 4

23. Which of the following is NOT considered a multivariate statistic?

a. Exploratory factor analysis

b. MANOVA

c. Canonical correlation

d. Chi-square analysis

Ans: a

Page: 386

Type: F

LO: 4

24. A(n) ________ is used to analyze the underlying pattern of correlations among a set of measured variables and to reduce the number of variables by extracting a smaller set of factors.

a. exploratory factor analysis

b. confirmatory factor analysis

c. canonical correlation

d. MANOVA

Ans: c

Page: 388

Type: F

LO: 4

25. A researcher wants to use scores on a measure of self-esteem to predict a group of outcome measures, including marital satisfaction and employment satisfaction. Which of the following procedures would be most useful?

a. Loglinear analysis

b. Confirmatory factor analysis

c. Canonical correlation

d. MANOVA

Ans: d

Page: 388

Type: C

LO: 5,7

26. Which of the following statistical tests is used to determine whether the observed relationships among a set of variables conforms to a theoretical prediction about how those variables should be related?

a. Exploratory factor analysis

b. Loglinear analysis

c. Canonical correlation

d. Structural equation analysis

Ans: b

Page: 388-389

Type: F

LO: 5,7

27. Which of the following is a type of structural equation analysis?

a. Exploratory factor analysis

b. Confirmatory factor analysis

c. Canonical correlation

d. MANOVA

Ans: a

Page: 389

Type: F

LO: 5

28. Factors in a factor analysis are equivalent to __________ in a structural equation analysis.

a. latent variables

b. factors

c. factor scores

d. contrasts

Ans: c

Page: 389-390

Type: F

LO: 7

29. A researcher has conducted a structural equation analysis. Which of the following will she use to test how well the collected data fit the hypothesized relationship?

a. The Tukey HSD test

b. The Fisher LSD test

c. The goodness of fit statistic

d. The loglinear statistic

Ans: d

Page: 375

Type: C

LO: 1

30. What statistical analysis would be best used to predict levels of overall aggression from knowledge about blood pressure and eating habits?

a. Chi-square

b. Pearson correlation

c. One-way ANOVA

d. Multiple regression

Ans: d

Page: 378

Type: A

LO: 1

31. Natalie is interested in predicting interpersonal attraction from knowledge about proximity and physical attractiveness. She first enters these two variables in her analysis. Then, after obtaining preliminary results, she adds the predictor variables of income and status. Natalie is using which of the following?

a. Factor analysis

b. Simultaneous multiple regression

c. ANOVA

d. Hierarchical multiple regression

Ans: b

Page: 379-380

Type: C

LO: 2

32. Loglinear analysis is a better statistical test than ANOVA when predicting a nominal dependent variable because

a. p-values are always lower in ANOVA.

b. ANOVA should only be used when the dependent variable has an approximately normal distribution.

c. loglinear analysis is easier to compute.

d. loglinear analysis creates more Type 2 errors.

Ans: a

Page: 382

Type: F

LO: 3

33. After testing the F value in an ANOVA, the researcher may want to follow up with

a. post hoc or a priori contrast analyses

b. a multiple regression analysis

c. a MANOVA

d. a post-experimental correlation

Ans: c

Page: 382

Type: F

LO: 3

34. To decrease the likelihood of making a Type 1 error during the means comparison process, a researcher could

a. compare all means within the study.

b. eliminate all data that supports the null hypothesis.

c. only conduct pairwise comparisons and only if the F value is significant.

d. increase alpha.

Ans: b

Page: 384-386

Type: A

LO: 5

35. Which of the following designs would be most likely to utilize an exploratory factor analysis?

a. A study predicting a measure of conformity during emergency situations

b. A study that has measured a number of personality characteristics related to helping

c. A one-way experimental design with one dependent measure of helping

d. A study assessing a measure of helping behavior across cultures

Ans: c

Page: 388

Type: F

LO: 6

36. Which of the following is a disadvantage of a MANOVA analysis?

a. Independent variables are related to the dependent variables.

b. It requires multiple F tests.

c. The likelihood of a Type 1 error is increased

d. At least three independent variables are needed for reliability.

Ans: b

Page: 388-389

Type: F

LO: 7

37. One of the advantages of a structural equation analysis is that

a. only the relationships between nominal variables need to be tested.

b. relationships among latent variables can be studied.

c. there is a smaller chance of a placebo effect than in correlational studies.

d. p-values do not need to be set at .05.

Ans: a

Page: 387

Type: F

LO: 6

38. In which of the following cases would a MANOVA analysis be most appropriate?

a. There is more than one dependent variable.

b. All of the variables are continuous.

c. The researcher wants to study the relationship between gender and ethnicity.

d. A structural equation analysis has already been performed.

Ans: b

Page: 388-389

Type: F

LO: 7

39. Which of the following best describes the goal of a structural equation analysis?

a. The relationships among nominal variables are tested.

b. The relationships among latent variables are tested.

c. A few dependent variables are expanded into many dependent variables.

d. P-values are set at less than .05.