ERM 681 MIDTERM EXAM I Fall 2017 Part I (MCQs 21), II (Long Qs 8), III (Very Long Qs 2), HW3 : Q 1 to 5

ERM 681 MIDTERM EXAM I Fall 2017

PART I

Select the BEST answer. Enter your answers in CANVAS “Midterm Section I” under Quizz Each answer is worth 1 point. You will have only one attempt to enter your answer in Canvas.

Chapter 11 Simple Analysis of Variance

1. Which of the following is most accurate?

· When H0 is true, MSTreat underestimates the population variance.

· When H0 is true, MSError underestimates the population variance.

· When H0 is false, MSTreat underestimates the population variance.

· When H0 is false, MSError overestimates the population variance.

2. What does a failure to reject the null hypothesis imply in a one-factor ANOVA?

· There is no difference in any pair of means selected from the k population means.

· The k population means are equal to a specified population mean.

· Each population mean differs significantly from every other population mean.

· Some subset of population means significantly differs from some other subset of population means.

3. What happens to the critical F value if you keep the number of treatments constant and alpha constant, but the overall sample size decreases?

· The critical F value will decrease.

· The critical F value will stay the same.

· The critical F value will increase.

· There is not enough information to answer this question.

4. For a set of 12 test scores, s = 10. Six of the twelve were girls, and for the girls s = 8. For the six boys, s = 6. What was the sum of squares for the total group of 12?

· 480

· 1000

· 1100

· 1200

· none of the above

5. Which of the following is an assumption in one-way ANOVA?

· The design does not always have to be balanced.

· The group(s) with the larger variance(s) also have larger n’s

· The population variances are the same for each group.

· The sample variances are the same for each group.

Use the following incomplete ANOVA summary table to answer the next four questions. You must find the missing values in order to answer the questions.

Source	SS	df	MS	F
Treatment	140.10	3	46.70	6.42
Error	116.32	16	7.27
Total	256.42	19

6. What is the expected experimental error variance that can be attributed to the treatment groups (assume a balanced design) under the null hypothesis?

· 6.12

· 6.42

· 7.27

· 13.50

· 46.70

7. If Ho is true, which significance level is likely to produce the most Type I error when testing the F statistic

· .01

· .025

· .05

· .50

8. What can you determine about the treatment groups?

· There is no significant difference between these groups.

· Each group is significantly different from every other group.

· At least two groups are significantly different from one other group.

· More than two groups are significantly different from one another.

9. How many treatment groups are in this study?

· 3

· 4

· 12

· 16

Stop Using ANOVA Summary Table.

10. If we know that the null hypothesis is true for a one-factor ANOVA and if the sample size of each treatment group is increased, what is the effect on the probability of achieving statistical significance at the specified alpha level? It is

· increased.

· decreased.

· the same.

· Impossible to determine because more information is needed.

11. What does the alternate hypothesis for one-factor analysis of variance address?

· Difference in population means

· Difference in population mean squares

· Difference in population variances

· Difference in population sum of squares

Chapter 12: Multiple Comparisons Among Treatment Means

12. A researcher carried out all possible pairwise comparisons on four group means and for each test set α to .05. What would be the correct family-wise Type I error rate in this case?

· .050

· .185

· .200

· .265

13. Which of the following is always true of a priori orthogonal contrasts?

· The decision to make the contrasts is made before data are collected.

· Comparisons are always made using a t-test.

· A Bonferroni, or similar familywise error rate correction, should be made.

· Comparisons are decided upon after initial data analysis is completed.

14. Which of the following set of contrast coefficients would be used to compare the average of treatments 1 and 6 with the average of treatments 2 and 4 if there were six means in all?

· -1 1 0 -1 0 1

· -1 1 0 1 0 -1

· -1/2 1/2 0 1/2 0 -1/2

· 1/2 -1/2 0 1/2 0 -1/2

15. Which of the following post hoc tests could possibly yield the largest number of significant mean differences?

· Bonferroni

· Neuman-Keuls

· Scheffe

· Tukey

CHAPTER 13: Factorial ANOVA

For next 4 problems, use the following ANOVA table in which α = .01.

Source	df	MS	F	FCRIT
Ethnicity (A)	2	45
Guilty (B)		70
AxB		17
Error	60	10
Total	65

16. For which source is the null hypothesis rejected at the .01 level of significance?

· Ethnicity (A)

· Anxiety (B)

· A x B

· None of the above

· All of the above

17. How many cells were there in the above design?

· 1

· 2

· 3

· 6

18. What is the total number of subjects in the design?

· 60

· 64

· 65

· 66

19. In the ANOVA summary table SS A x B would equal?

· 17

· 34

· 51

· 102

STOP USING TABLE

20. In ANOVA, the interaction of two variables is certainly present when

a) the two variables are positively correlated.

b) the two variables are negatively correlated.

c) the row effects are NOT consistent across the columns.

d) the main effects do NOT account for all of the total observed score variance.

21. When analyzing mean differences among more than two samples, doing independent t tests on all possible means

a) decreases the probability of a Type I error

b) does not change the probability of a Type I error

c) increases the probability of a Type I error

d) cannot be determined from the information provided