Answers to Textbook Chapter Questions
Overview
Answers are provided here for questions found at the end of chapters in the textbook. We
have included only questions that have definite, predetermined solutions. Therefore,
answers are not given for every chapter.
ANSWERS
Chapter 6
1. M = 6.0
2. s = 3.23
3. It is the same.
4. Median = 6
6. Stem-and-leaf display
Stem
Leaves
f
6
0248
4
5
01368
5
4
1223566889
10
3
01569
5
2
1578
4
1
38
3
30
7. For confidence intervals, standard error of mean is
From J.R. Thomas, J.K. Nelson, and S.J. Silverman, 2015, Research methods in physical activity instructor guide, 7th ed.
(Champaign, IL: Human Kinetics).
For 95% confidence using the t table, values jump from n = 60 to 100, but we need 99. We
could interpolate, but usually one uses the lower n, in this case 60. Hence, 2.0 is the table
value for 95%. The CI for this mean is 50 ± (0.6 × 2.0), or 50 ± 1.2.
Chapter 8
1. r = .91
2. It is significant (table value of r for 13 df at .01 = .64).
With 30 participants, using 25 df (table does not give 28), one would need a correlation of
.49 to be significant for p < .01. If you interpolate between 25 and 30 df, the actual r needed
for 28 df is .46.
3. b = 0.016, a = 6.95
Y = 6.95 + 0.016X
Note: The value for a (and the entire regression formula) will vary considerably depending on
the number of decimal places one uses. For example, for b = 0.0163, a = 6.10. However, if
you round 0.0163 to 0.02, then a is –4.46! We recommend that you use three or four
decimal places. We rounded b to three places (0.016) for this problem.
4. Predicted VO2max for subject who ran 2,954 m = 54.21
Predicted VO2max for subject who ran 2,688 m = 49.96
Chapter 9
3. a. df: between = 2, within = 15, total = 17; MS: between = 20.05, within = 2.29
b. F = 8.76, p < .05
c. omega square = .46, so 46% of total variance is accounted for by treatment.
d. For Scheffé, the critical difference needed is 2.4; thus circles were significantly more
effective than both clusters and rows. There was no significant difference between clusters
and rows.
Chapter 10
1. The contingency table is shown with the calculated expected frequencies in parentheses.
From J.R. Thomas, J.K. Nelson, and S.J. Silverman, 2015, Research methods in physical activity instructor guide, 7th ed.
(Champaign, IL: Human Kinetics).
Racquetball
Weight
training
Aerobic
dance
Total
Observed
35
45
30
110
Expected
(35)
(32)
(43)
Observed
28
13
49
Expected
(28)
(26)
(36)
Totals
63
58
79
Men
Women
90
200
chi square = 20.40. With 2 df, chi square needed at p < .01 = 9.21.
a. Interpretation: There is a significant difference in the choices of activities by men and
women, chi square (2, N = 200) = 20.40, p < .01. A much larger percentage of men than of
women chose weight training (41% and 14%, respectively), whereas many more women
(54%) chose aerobic dance than did men (27%). There was no difference in racquetball.
b. With five rows and four columns, df = 12, and the chi square needed for significance at the
.05 level is 21.03.
Chapter 11
1. Among the ways that construct validity could be shown are the following:
Known group difference method. For example, varsity athletes’ performance versus
nonathletes’ performance, older children versus younger children, participants who had
been instructed versus those who had not, and so on. If test results separate such groups,
then evidence of construct validity is shown.
Convergent validity. For example, the test correlates moderately with some measure where
performances would be expected to be related, such as a fitness test, a sport skills test, some
other type of manipulative skill test such as juggling, and so on.
From J.R. Thomas, J.K. Nelson, and S.J. Silverman, 2015, Research methods in physical activity instructor guide, 7th ed.
(Champaign, IL: Human Kinetics).
Divergent validity. One would expect very low correlation between the test and some test
that does not measure that ability, such as a test of scholastic aptitude with the motor
dexterity test, a kicking test with the throwing test, or a static balance test with the power
test.
Criterion validity could be shown by correlating one’s test with a test that purportedly
measures that same quality, such as correlating a power test with the Margaria stair-climbing
test, a manipulative test with the pegboard test, or a throwing test with the AAHPERD
softball-throwing test. Judges’ ratings could also be used as a criterion with which to
correlate one’s test results.
3. The standard error is
For 95% confidence, multiply by 2: 3.10 × 2 = 6.2. Therefore, one can say with 95%
confidence that her true score is between 78 ± 6.2, or between approximately 72 and 84.
From J.R. Thomas, J.K. Nelson, and S.J. Silverman, 2015, Research methods in physical activity instructor guide, 7th ed.
(Champaign, IL: Human Kinetics).