Exam 5 - PSY 211
Exam 5 - PSY 211
Course Reference #22021132
Mike Hoerger
Name: _____________________________
Section I. Multiple Choice [90 points: 3 points per question]
1 ANOVA is to _______ as repeated-measures ANOVA is to _______. a) single sample t-test; between-group t-test b) within-subject t-test; single-sample t-test c) between-group t-test; within-subject t-test
2 Chi-square test for goodness of fit usually involves how many categorical variables? a) 0 b) 1 c) 2
3 With regard to ANOVA, how many of the following statements are true?
I. The null hypothesis is false when all groups have the same mean.
II. If F = 0.13, the null hypothesis should always be accepted.
III. If F = 3.56, the alternative hypothesis should always be accepted. a) 0 b) 1 c) 2 or 3
4 Compared to the LSD post hoc test, the Bonferroni test is more _____. a) conservative b) independent c) liberal
5 Attractiveness is likely this type of variable a) polynomial b) continuous c) categorical
6 Using ANOVA, F will always fall within what range? a) between negative infinity and infinity b) between 0 and 1 c) between 0 and infinity
7 The test statistic chi-square ( χ 2 ) is often mispronounced. “Chi” starts with a
______ sound, and rhymes with the word ______. a) ch; he b) ch; eye c) k; eye
8 a) 1 b) 2 c) 3
How many of the following are parametric tests?
I. standard deviation
II. between-group t-test
III. chi-square
IV. repeated-measures ANOVA
9 Which of the following best characterizes the null hypothesis for chi-square tests? a) The frequency of people in each category will be a specific proportion b) The expected frequencies will be significantly lower than obtained frequencies c) The group means expected at the population level will not reliably differ
10 Statistical test most appropriate for comparing depression (diagnosis vs. no diagnosis) to favorite color (blue vs. black vs. other) a) correlation (r) b) ANOVA (F) c) chi-square ( χ 2 )
11 To determine whether an obtained F value is statistically significant, it must be compared to a critical F value. What two pieces of information are needed to determine the critical F value? a) sample size, number of groups b) mean, sample standard deviation c) expected frequency, obtained frequency
12 A CMU task force on diversity examines the relationship between ethnicity and several outcome variables, such as GPA, well-being, and anxiety. Well-being would best be characterized as a) an independent variable b) a confounding variable c) a dependent variable
13 What is the function of a post hoc test? a) Indicate whether any statistically significant group differences have occurred b) Describe which groups reliably differ c) Set the critical value for the F test (or chi-square)
14 The repeated-measures ANOVA is designed to decrease this type of error a) individual differences b) uncontrolled environmental factors c) unreliability of measures
15 Which of the following most represents a significant ANOVA correctly typed in
APA format? a) The ANOVA was statistically significant, F (3, 256) = 12.31, p > .05 b) The ANOVA was statistically significant, F (5, 61) = 1.13, p < .001 c) The ANOVA was statistically significant, F(4, 81) = 7.13, p = .04
16 Based on the following repeated-measures ANOVA data, estimate the value of
F:
Participant
Ellen
Jim
Intake
10
5
Discharge
9
4
Follow-up
10
6
Andrea
Brian
Corey a) 0.92 b) 1.87 c) 3.68
7
9
4
6
10
7
7
9
4
17 Which of the following would lead to a statistically significant chi-square? a) Approximately equal expected and observed frequencies b) An observed frequency of zero in at least one of the groups c) Discrepancy between expected and observed frequencies
18 As variability due to chance decreases, the value of F will a) increase b) stay the same c) decrease
19 If F = 5, the result is statistically significant a) Always b) Sometimes c) Never
20 In a psychiatric hospital, researchers attempt to examine the relationship between psychological diagnosis and number of suicide attempts. The diagnosis of “schizophrenia” would be considered a) an element b) a level c) a factor
21 If F = 1, the result is statistically significant a) Always b) Sometimes c) Never
22 Two variables, gender and ethnicity, are most likely a) independent b) codependent c) dependent
23 Type of test typically used for longitudinal studies a) ANOVA b) repeated-measures ANOVA c) chi-square test for goodness of fit
24 One study (N = 47) uses ANOVA, and another study (N = 47) uses repeatedmeasures ANOVA. Which study design has more power? a) repeated-measures ANOVA b) ANOVA c) they’re equal in power
The remaining multiple choice problems (25-30) involve interpreting SPSS
Output. Be sure to write all answers on the scantron form.
25 The following Output is from what type of test?
ANOV A
87. Delaying Gratification
Between Groups
W ithin Groups
Total
Sum of
Squares
3.038
995.882
998.920
a) between-group t-test b) ANOVA c) repeated-measures ANOVA df
5
320
325
Mean S quare
.608
3.112
F
.195
Sig.
.964
26 There are two important chi-square Output boxes in SPSS. One of them is directly below. Determine which box (A, B, or C) is the correct corresponding second box. Only one answer is correct. The other boxes contain incorrect information.
2.
Gender
Total
Female
Male
2. Gender * 26. Electronic Communication Crosstabulation
Cell
Phone
123
26. Electronic Communication
21
Ins tant
Messaging
42
Text
Messaging
37 Count
Expected
Count
Count
Expected
Count
Count
Expected
Count
118.3
50
54.7
173
173.0
23.9
14
11.1
35
35.0
44.5
23
20.5
65
65.0
53.0
36.3
16
16.7
53
Total
223
223.0
103
103.0
326
326.0
a)
Chi-Square Te sts
Pearson Chi-S quare
Lik elihood Ratio
Linear-by-Linear
As soc iation
N of V alid Cases
Value
2.205
2.161
.325
326 a df
3
3
1
As ymp. Sig.
(2-sided) a. 0 c ells (.0% ) have expected count less than 5. The
.531
.540
.569
b) c)
Pearson Chi-S quare
Lik elihood Ratio
Linear-by-Linear
As soc iation
N of V alid Cases
Value
3.918
3.725
.325
325 a df
3
3
1
As ymp. Sig.
(2-sided) a. 0 c ells (.0% ) have expected count less than 5. The
.321
.325
.468
Pearson Chi-S quare
Lik elihood Ratio
Linear-by-Linear
As soc iation
N of V alid Cases
Value
9.831
9.462
.465
a df
3
3
1
As ymp. Sig.
(2-sided)
326 a. 0 c ells (.0% ) have expected count less than 5. The minimum expected count is 11. 06.
.023
.038
.027
27 A researcher conducted a study examining whether being in a relationship (yes or no) was related to favorite fast food place. How would a researcher correctly summarize the following analyses?
27. Fas t
Food
Choice
Total
27. Fast Food Choice * 19. Relationship Status Crosstabulation
Arby's
Burger King
McDonald's
Taco Bell
Wendy's
Other
Count
Expected Count
Count
Expected Count
Count
Expected Count
Count
Expected Count
Count
Expected Count
Count
Expected Count
Count
Expected Count
56
56.3
27
25.8
10
16.0
153
153.0
19. Relationship
Status
No
23
25.3
Yes
31
28.7
16
12.2
21
17.4
10
13.8
16
19.6
64
63.7
28
29.2
24
18.0
173
173.0
Total
54
54.0
26
26.0
37
37.0
120
120.0
55
55.0
34
34.0
326
326.0
Chi-Square Te sts
Pearson Chi-S quare
Value
8.366
a df
5
As ymp. Sig.
(2-sided)
.137
Lik elihood Ratio
Linear-by-Linear
As soc iation
8.528
1.162
5
1
.129
.281
N of V alid Cases 326 a) The most interesting finding is that people who are not in a relationship are more likely to prefer Burger King or McDonald’s, when compared to those in a relationship. b) Surprisingly, daters are more likely to enjoy Arby’s, Taco Bell, and “Other” fast food places than non-daters. c) Unfortunately, dating status was not related to fast food preference.
28 Based on our data file, an ANOVA was conducted comparing type of vehicle (none, car, or other) to enjoyment of violence shows (high scores = more enjoyment). The
F test was statistically significant, and the post hoc results are below. How many of the following statements are true?
I. People who do not have cars (None) watch significantly more violent shows than people who drive cars (Car).
II. People who drive cars (Car) watch significantly more violent shows than people who drive other types of vehicles (Other).
III. There were no significant differences between people who do not have cars
(None) and people who drive other types of vehicles (Other)
IV. None of the observed differences are likely to occur at the population level.
V. None of the observed differences are due to sampling error.
Multiple Comparisons
Dependent Variable: 82. Violent Shows
LSD a) 1 b) 2 c) 3
(I) 21. Vehicle
None
(J) 21. Vehicle
Car
Mean
Difference
(I-J)
.3173
Std. Error
.3043
Car
Other
None
-.2505
-.3173
.3513
.3043
Other
Other
None
-.5679*
.2505
.2358
.3513
Car .5679*
*. The mean difference is s ignificant at the .05 level.
.2358
Sig.
.298
.476
.298
.017
.476
.017
95% Confidence Interval
Lower Bound Upper Bound
-.281
.916
-.942
-.916
-1.032
-.441
.104
.441
.281
-.104
.942
1.032
29 The following data were used to evaluate the effectiveness of an intervention designed to decrease PTSD symptoms. Ratings of PTSD symptoms were made on a 0-10 scale, with higher scores indicating more symptoms. Ratings were made at the start of the study (Intake), end of the study (Discharge), and at a 1-year follow-up
(Follow-up).
Using the following Output, report the appropriate value for F.
Intake
Discharge
Follow-up
Descriptive Statistics
Mean
8.0645
6.6452
7.6774
Std. Deviation
1.80620
1.42708
N
31
31
2.11955
31
Tests of Within-Subjects Effects
Measure: MEASURE_1
Source factor1
Error(factor1)
Sphericity Assumed
Greenhous e-Geisser
Huynh-Feldt
Lower-bound
Sphericity Assumed
Greenhous e-Geisser
Huynh-Feldt
Lower-bound
Type III Sum of Squares
33.376
33.376
33.376
33.376
237.290
237.290
237.290
237.290
df
2
1.980
2.000
1.000
60
59.388
60.000
30.000
Mean Square
16.688
16.860
16.688
33.376
3.955
3.996
3.955
7.910
F
4.220
4.319
4.319
4.373
Sig.
.019
.020
.019
.049
Pairwise Comparisons
Measure: MEASURE_1
(I) factor1
1
2
3 1
2
1
3
(J) factor1
2
3
Mean
Difference
(I-J)
1.419*
.387
-1.419*
-1.032
-.387
1.032
Std. Error
.503
.483
.503
.528
.483
.528
Sig.
a
.008
.430
.008
.060
.430
.060
95% Confidence Interval for
Difference a
Lower Bound Upper Bound
.392
2.447
-.600
1.374
-2.447
-2.110
-1.374
-.045
-.392
.045
.600
2.110
a. adjustm ents).
30 Is the treatment reliably effective at follow-up? a) Yes b) No c) More information needed
Section II. Short Answer [10 points]
31. [6 points] Describe the differences between the following four analyses:
1) correlation, 2) between-group t-test, 3) ANOVA, 4) chi-square
Correlation: compares two continuous variables
Between-group t-test: compares a categorical variable (with two categories) to a continuous variable
ANOVA: compares a categorical variable (with two or more categories) to a continuous variable
Chi-square: involves categorical variables only
This was all that was required for full credit, but some people provided excellent additional details. Most people earned full credit.
32. [2 points] You design a study to determine which type of several beers is the best.
What type of analysis would you be most appropriate for this type of study?
There are often multiple ways to correctly analyze data, though some are better than others. Any correct choice that was logically supported earned full credit
…
“Beer” is a categorical variable. To judge which is best, some type of quality measure, such as a rating scale would likely be employed, indicating a continuous variable.
ANOVA is the most obvious choice. This is the most obvious answer, but a few other options are also defensible.
A chi-square test for goodness of fit might be appropriate, but one would need to specify exactly how this type of design was used. For example, people could vote for their favorite beer and one could specify that all beers would receive the same number of votes as the null hypothesis. Some people provided no explanation, which earned no points, as this type of alternative analysis is only appropriate if adequately explained.
Some people also said to use chi-square but gave illogical reasons for it.
A chi-square test for independence could be used if each beer was given a simple categorical quality rating, like “good” or “bad” – I don’t think anyone said this.
For full credit and one extra credit point, one could have supported using the repeatedmeasures ANOVA. In this design, each person would make a quality rating for each beer. Because each person would rate all beers, the repeated-measures ANOVA could be used (this is slightly different from some of the examples we used in class, which mainly focus on changes over time). Because each person would act as there own control, this study would have higher power than a simple ANOVA. Only a couple people provided this response – very smart!
33. [2 points] There are numerous political parties in the United States, and their supporters often make foolish claims of a variety of sorts. Ann Coulter is a political commentator who opposes the Democratic Party. A few months ago, she said she wished females did not have the right to vote because being fem ale is “correlated” with voting Democrat. From a purely statistical viewpoint, what is wrong with her word choice?
The question asks “what is wrong with her word choice?” and the word correlated is in quotations. The correct answer would be that because gender and political party are both categorical, chi-square would be the appropriate analysis. Non-statisticians frequently use the word “correlated” any time to variables are related, even if the analysis involves t-tests, ANOVA, or chi-square.
Some people raised very interesting points about causation, p-values, and effect sizes, but these are not related to the wording of her statement
– thus, these people ignored the question.
It is important to be able to spot these types of verbal errors, as they are common among people who have a poor understanding of statistics. If a person does not understand a correlation coefficient, they probably have no business citing any statistics, let alone arguing that the 19 th amendment should be overturned.
PSY 211
Exam #5
Study Guide
Here are the major concepts for the exam. You are responsible for all material presented in lecture, so make sure to ask a friend or refer to the book if you missed class. There are 30 multiple choice questions on the exam (3 points each), and the remaining 10 points will come from short answer.
Continuous vs. categorical variable
*When to use each type of analysis: correlation, between-group t-test, repeated-measures t-test, ANOVA, repeated-measures ANOVA, chi-square
The independent and dependent variable for each type of analysis
Similarities between t-tests and ANOVA
Differences between ANOVA and repeated-measures ANOVA
Factor vs. level
Null and alternative hypotheses for ANOVA
Rationale of the F statistic for ANOVA and repeated-measures ANOVA
Range of F values, approximate F values when null hypothesis is correct vs. when the alternative hypothesis is correct
Shape of the F distribution; differences between the F, t, and z distributions
Critical F value is based on two different reference numbers called degrees of freedom, with one based on the sample size, and one based on the number of groups
Rationale of using post hoc tests
Difference between F-test and post hoc tests
How the LSD test differs from other post hoc tests
*Basic information from Output for ANOVA, repeated-measures ANOVA, and chisquare
Identify appropriate APA style write-ups on multiple choice questions
*Based on a description of a study be able to identify whether ANOVA or repeatedmeasures ANOVA would be used
Two types of error involved in experiments
Repeated-measures ANOVA controls for one of the types of errors
Parametric vs. non-parametric tests
Difference between chi-square test for goodness of fit and chi-square test for independence
Hypothesis testing with chi-square
Distinction between observed vs. expected frequencies in chi-square
Independent vs. dependent
Also look over the review from the final Monday of class.
