Chapter 1: Introduction to Statistics

Answers to Even-Numbered Problems

2. The population is the entire group of individuals (or scores) of interest for a particular
research study. A sample is a group selected from a population that usually is used to represent
the population in a research study. A parameter is a characteristic, usually a numerical value, that
describes a population. A statistic is a characteristic, usually numerical, that describes a sample.

4. Sampling error is the naturally occurring difference between a sample and the population
from which the sample is obtained. Specifically, the statistics obtained for a sample will be
different from the corresponding parameters for the population and the statistics will differ from
one sample to another. This is a problem for inferential statistics because any difference found
between two treatment conditions may be explained by the treatments but it also may be
explained by sampling error.

6. The goal of an experiment is to demonstrate the existence of a cause-and-effect relationship

between two variables. To accomplish the goal, an experiment must manipulate an independent
variable and control other, extraneous variables.

8. This is not an experiment because no independent variable is manipulated. The researchers

are comparing two preexisting groups of individuals.
10. a. The dependent variable is comprehension of the passage, which is measured by the test
scores.
b. Knowledge or comprehension is continuous.
c. ratio scale (zero means no correct answers)

12. a. The independent variable is taking the Tai Chi course versus not taking the course.
b. Nominal scale
c. The dependent variable is the amount of arthritis pain experienced.
d. The amount of pain probably is measured with an interval or a ratio scale.

14. a. An ordinal scale provides information about the direction of difference (greater or less)
between two measurements.
b. An interval scale provides information about the magnitude of the difference between
two measurements.
c. A ratio scale provides information about the ratio of two measurements.

16. Honesty is an attribute or personality characteristic that cannot be observed or measured

directly. Shyness could be operationally defined by identifying and observing external behaviors
associated with being shy. Or, participants could be given a questionnaire asking how they
behave or feel in situations for which shyness might have an influence.

18. a. X = 10
b. X2 = 38
c. X + 1 = 11
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 d. (X + 1) = 14

20. a. X = 0
b. X2 = 50
c. (X + 3) = 15

22. a. (ΣX)2
b. X2
c. Σ(X – 2)
d. Σ(X – 1)2

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 2: Frequency Distributions

Answers to Even-Numbered Problems

2. X f p %
────────────
9 1 0.05 5%
8 0 0.00 0%
7 1 0.05 5%
6 2 0.10 10%
5 4 0.20 20%
4 2 0.10 10%
3 3 0.15 15%
2 5 0.25 25%
1 2 0.10 10%
────────────

4. a. n = 14
b. ΣX = 44
c. ΣX2 = 168

6. X f
60-64 1
55-59 2
50-54 2
45-49 1 Younger drivers, especially those 20 to 29
40-44 2 years old, tend to get more tickets.
35-39 3
30-34 3
25-29 5
20-24 8
15-19 3

8. A regular table reports the exact frequency for each category on the scale of measurement.
After the categories have been grouped into class intervals, the table reports only the overall
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
frequency for the interval but does not indicate how many scores are in each of the individual
categories.

10. a. X f
10 2
9 4
8 4
7 4
6 3
5 2
4 1

b.

5│
│
4│ ┌───┬───┬───┐
│ │ │ │ │
f 3│ ┌───┤ │ │ │
│ │ │ │ │ │
2│ ┌───┤ │ │ │ ├───┐
│ │ │ │ │ │ │ │
1 │ ┌───┤ │ │ │ │ │ │
│ │ │ │ │ │ │ │ │
└─ ──┴───┴───┴───┴───┴───┴───┴───┴───── X
4 5 6 7 8 9 10 11

12.
17│ ┌────┐
16│ │ │
15│ │ │
14│ ┌────┐ │ │
13│ │ │ │ │
12│ │ │ ┌────┐ │ │
11│ │ │ │ │ │ │
10│ │ │ │ │ │ │
f 9│ │ │ │ │ │ │ ┌────┐
8│ │ │ │ │ │ │ │ │
7│ │ │ │ │ │ │ │ │ ┌────┐
6│ │ │ │ │ │ │ │ │ ┌────┐ │ │
5│ │ │ │ │ │ │ │ │ │ │ │ │
4│ │ │ │ │ │ │ │ │ │ │ │ │
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 3│ │ │ │ │ │ │ │ │ │ │ │ │
2│ │ │ │ │ │ │ │ │ │ │ │ │
1│ │ │ │ │ │ │ │ │ │ │ │ │
└ ─┴────┴───┴────┴───┴────┴───┴────┴───┴────┴───┴────┴─
Plumb. Elect. Secur. Book. Nurse Educat.
Job Advertisement Categories

14. X f
15 5
14 6
13 4 The distribution is negatively skewed.
12 2
11 2
10 1

16. X flow fhigh


5 1 5
4 2 6
3 4 3
2 5 1
1 3 0
The scores for children from the high number-talk parents are noticeably higher.

18.

X f cf c%

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 25–29 1 25 100

20–24 4 24 96

15–19 8 20 80

10–14 7 12 48

5–9 3 5 20

0–4 2 2 8

a. 20%
b. 80%
c. X = 14.5
d. X = 24.5

20. a. 32%
b. 59%
c. X = 14.25
d. X = 15.10

22. a. X = 22
b. X = 31.5
c. 93%
d. 28%

24. a. 6
b. 75, 79, 72, 73, 77, 74
c. 1
d. 48

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 3: Central Tendency
Answers to Even-Numbered Problems

2. The mean is 51/9 = 5.67, the median is 6, and the mode is 3.

4. The mean is 66/13 = 5.08, the median is 5, and the mode is 5.

5.  = 120/15 = 8

8. N = 5.

10. a. The combined sample mean is M = 12.

b. The combined sample mean is (24 + 80)/8 = 13.
c. The combined sample mean is (40 + 48)/8 = 11.

12. The original population has N = 15 and ΣX = 120. The new population has N = 15 and ΣX =
105. The new mean is  = 7.

14. The original sample has n = 9 and ΣX = 180. The new sample has n = 9 and ΣX = 198. The
total (ΣX) increased by 18 points. If the score was X = 7, it was increased to X = 25.

16. The original sample has n = 6 and ΣX = 78. The new sample has n = 5 and ΣX = 75. The
new mean is M = 15.

18. The original sample has n = 10 and ΣX = 90. The new sample has n = 9 and the total is still
ΣX = 90. The new mean is M = 10.

20. The original population has N = 8 and ΣX = 128. The new population has N = 7 and ΣX =
105. The removed score must be X = 23.

22. With a skewed distribution, the extreme scores in the tail can displace the mean out toward
the tail. The result is that the mean is often not a very representative value.

24. The participants recalled an average of M = 68/16 = 4.25 humorous sentences compared to
M = 49/16 = 3.06 nonhumorous sentences. Humor does appear to improve memory
performance.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 4: Variability
Answers to Even-Numbered Problems

2. Variance and standard deviation cannot be less than zero because they are computed by
adding squared deviations. Squared deviations are always greater than or equal to zero.

4. a. The definitional formula works well when the mean is a whole number and there is a
relatively small number of scores.
b. The computational formula is better when the mean is a fraction or decimal value and
usually easier with a large number of scores.

6. SS = 80, variance is 16, and the standard deviation is 4.

8. a. μ = 3
│
3 │ ┌───┐
│ │ │
2 │ │ ├───┐
f │ │ │ │
1 │┌───┐ │ │ │
││ │ │ │ │
└┴───┴───┴───┴───┴─── X
1 2 3 4 5

b. The mean is μ = 18/6 = 3. The three scores of X = 3 are exactly equal to the mean (zero
distance) and the greatest distance is 2 points for X = 1. The standard deviation appears to
be about 1 point.
c. SS = 6, σ2 = 1, σ = 1

10. Without some correction, the sample variance underestimates the variance for the
population. Changing the formula for sample variance (using n – 1 instead of N) is the necessary
correction.

12. SS = 14, s2 = 2, and s = 2 = 1.41,

14. a. The mean is μ = 53 and the standard deviation is still σ = 10.

b. The new mean is μ = 100 and the new standard deviation is σ = 20.

16. After subtracting 80 points, the mean is M = 4.5 and the standard deviation is s = 3. Change
back to the original scores by adding 80 points and the mean is 84.5 and s is still 3.

18. a. The range is 7 or 8 points and with SS = 36 the standard deviation is  = 6 = 2.45.
b. After adding 2 points, the range is still 7 or 8 and the standard deviation is still  = 2.45.

20. a. The mean is M = 7 and with SS = 16, s = 2.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 b. After changing X = 10 to X = 0, the new mean is M = 5 and, with SS = 36, s = 3.
c. The mean is displaced in the direction of the extreme score and the standard deviation is
increased.

22. a. For the younger woman, the variance is s2 = 0.786. For the older woman, the variance is
s2 = 1.696.
b. The variance for the younger woman is only half as large as for the older woman. The
younger woman’s scores are much more consistent.

24. a. X = 84 has a much higher location in the distribution with s = 2. It is above the mean by
three times the standard deviation.
b. X = 72 is closer to average with s = 10. It is below the mean but less than one standard
deviation below.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 5: z-Scores: Location of Scores and Standardized
Distributions
Answers to Even-Numbered Problems

2. a. z = 0.50
b. z = 0.20
c. z = –2.00
d. z = –1.50

4. a. X z X z X z
75 1.25 48 –1.00 84 +2.00
54 –0.50 78 1.50 51 –0.75

b. X z X z X z
72 1.00 63 0.25 78 1.50
54 –0.50 45 –1.25 30 –2.50

6. a. X z X z X z
108 0.40 115 0.75 130 1.50
90 –0.50 88 –0.60 95 –0.25

b. X z X z X z
92 –0.40 90 –0.50 136 1.80
115 0.75 130 1.50 75 –1.25

8. a. z = +0.25
b. z = +0.50
c. z = +1.00
d. z = +2.50

10. σ = 3

12. μ = 50

14. μ = 40 and σ = 12. The distance between the two scores is 6 points which is equal to 0.5
standard deviations.

16. a. σ = 8
b. σ = 4

18. a. X = 52 (z = 0.20)
b. X = 60 (z = 1.00)
c. X = 44 (z = –0.60)
d. X = 30 (z = –2.00)

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
20. a. μ = 5 and σ = 4
b. & c. Original X z-score Transformed X
0 –1.25 75
6 0.25 105
4 –0.25 95
3 –0.50 90
12 1.75 135

22. X z X z X z
27 0.40 31 1.20 35 2.00
21 –0.80 22 –0.60 18 –1.40

24. s = 4

26. a. M = 3 and s = 2
b. & c. Original X z-score Transformed X
5 1.00 60
0 –1.50 35
4 0.50 55
5 1.00 60
1 –1.00 40
2 –0.50 45
4 0.50 55

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
Chapter 6: Probability
Answers to Even-Numbered Problems

2. Sampling with replacement means that each individual selected for a sample is returned to
the population before the next individual is selected. This sampling method is used to ensure
that the probability of selecting any specific individual remains constant.

4. a. p = 52/80 = 0.65
b. p = 28/80 = 0.35 (Remember, a random sample requires replacement.)

6. a. body to the left, p = 0.9938

b. body to the left, p = 0.7881
c. body to the right, p = 0.6915
d. body to the right, p = 0.7794

8. a. p = 0.1974
b. p = 0.4972
c. p = 0.7698

10. a. z = –1.64 or –1.65 c. z = 0.39

b. z = 0.52 d. z = –0.84

12. a. tail to the right, p = 0.4013

b. tail to the right, p = 0.2266
c. tail to the left, p = 0.3085
d. tail to the left, p = 0.1056

14. a. p(z > 0.50) = 0.3085

b. p(z < 0.80) = 0.7881
c. p(–1.00 < z < 1.00) = 0.6826

16. a. z = 2.67, p = 0.0038

b. p(1.33 < z < 2.67) = 0.0880
c. p(–0.67 < z < 0.60) = 0.4743

18. a. p(z > 1.00) = 0.1587

b. p(z > –0.67) = 0.7486

20. p(X > 36) = p(z > 2.17) = 0.0150 or 1.50%

22. a. μ = 10
b. σ = √5 = 2.24 and for X = 15.5, z = 2.46, and p = 0.0069
c. For X = 6.5, z = –1.56, and p = 0.0594

24. a. With five options, p = 1/5 for each trial. For n = 100, μ = 20 and σ = 4 For X = 20, z =
±0.13 and p = 0.1034.
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 b. For n = 36, μ = 7.2 and σ = 2.4 For X = 9.5, z = 0.96 and p = 0.1685.
c. For n = 64, μ = 12.8 and σ = 3.2 For X = 16.5, z = 1.16 and p = 0.1230.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 7: Probability and Samples: The Distribution of Sample
Means
Answers to Even-Numbered Problems

2. a. The expected value is μ = 40 and σM = 8/√4 = 4.

b. The expected value is μ = 40 and σM = 8/√16 = 2.

4. The distribution of sample means will not be normal when it is based on small samples (n <
30) selected from a population that is not normal.

6. a. 10/4 = 5 points
b. 10/25 = 2 points

8. a. n > 4
b. n > 25
c. n > 100

10. a. σ = 8 points and z = 0.50

b. σM = 4 points and z = 1.00
c. σM = 2 points and z = 2.00

12. a. With a standard error of 6, M = 58 corresponds to z = 0.50.

b. With a standard error of 3, M = 58 corresponds to z = 1.00.
c. With a standard error of 2, M = 58 corresponds to z = 1.50.

14. a. σM = 5, z = 0.60, and p = 0.2743

b. σM = 3, z = 1.00, and p = 0.1587
c. σM = 1.5, z = 2.00, and p = 0.0228

16. a. σM = 4, z = 0.50 and p = 0.3085

b. Cannot answer because the distribution of sample means is not normal with n = 4.
c. σM = 1, z = 2.00 and p = 0.0228
d. With n = 64, the distribution of sample means is normal. σM = 1, z = 2.00 and p = 0.0228

18. a. Cannot answer because the distribution of sample means is not normal.
b. σM = 0.5, z = 3 and p = 0.0013
c. σM = 0.5, z = ±1.00 and p = 0.6826

20. a.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 b. Even considering the standard error for each mean, there is no overlap between the two
groups. The Tai Chi course does seem to have decreased self-reported pain even
allowing for the standard error.

22. a. With a standard error of 10, M = 86 corresponds to z = 0.90, which is not extreme.
b. With a standard error of 4, M = 86 corresponds to z = 2.25, which is extreme.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 8: Introduction to Hypothesis Testing
Answers to Even-Numbered Problems

2. The alpha level is a small probability value that defines the concept of “very unlikely.” The
critical region consists of outcomes that are very unlikely to occur if the null hypothesis is true,
where “very unlikely” is defined by the alpha level.

4. a. Lowering the alpha level causes the boundaries of the critical region to move farther out
into the tails of the distribution.
b. Lowering α reduces the probability of a Type I error.

6. a. The null hypothesis states that the herb has no effect on memory scores.
b. H0: μ = 50 (even with the herbs, the mean is still 50). H1: μ  50 (the mean has
changed) The critical region consists of z-scores beyond 1.96. For these data, the
standard error is 3 and z = 4/3 = 1.33. Fail to reject the null hypothesis. The herbal
supplements do not have a significant effect on memory scores.

8. a. The null hypothesis states that participation in sports, cultural groups, and youth groups
has no effect on self-esteem. H0: µ = 50, even with participation. With n = 100, the
standard error is 1.5 points and z = 3.8/1.5 = 2.53. This is beyond the critical value of
1.96, so we conclude that there is a significant effect.
b. Cohen’s d = 3.8/15 = 0.253.
c. The results indicate that group participation has a significant effect on self-esteem, z =
2.53, p < .05, d = 0.253.

10. a. With n = 16, the standard error is 2, and z = 3/2 = 1.50. Fail to reject H0.
b. With n = 64, the standard error is 1, and z = 3/1 = 3.00. Reject H0.
c. A larger sample increases the likelihood of rejecting the null hypothesis.

12. a. The null hypothesis states that the overweight students are no different from the overall
population, μ = 4.22. The standard error is 0.10 and the z-score for this sample is z =
2.60. Reject the null hypothesis. The number of snacks eaten by overweight students is
significantly different from the number for the general population.
b. The null hypothesis states that the healthy-weight students do not eat fewer snacks
than the overall population, H0: μ ≥ 4.22. The standard error is 0.12 and the z-score
for this sample is z = –1.75. For a one-tailed test, the critical value is z = –1.65. Reject
the null hypothesis. The number of snacks eaten by healthy-weight students is
significantly lower than the number for the general population.

14. a. The null hypothesis states that there is no effect on reaction time, µ = 400. The critical
region consists of z-scores beyond z = ±2.58. For these data, the standard error is 8 and
z = 22/8 = 2.75. Reject H0. There is a significant change in reaction time.
b. The null hypothesis states that there is no increase in reaction time, µ ≤ 400. The critical
region consists of z-scores beyond z = 2.33. For these data, the standard error is 8 and
z = 22/8 = 2.75. Reject H0. There is a significant increase in reaction time.
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 b. Cohen’s d = 22/40 = 0.55.

16. With n = 4, the standard error is 0.95 and the sample mean corresponds to z = 2.79. This is
well beyond the critical boundary of 1.65. Reject the null hypothesis and conclude that the
number of 90-degree days in the past four years is significantly higher than the overall mean
of µ = 9.6.

18. a. H0: μ = 45 (the supplement has no effect). The standard error is 2.25 and z = 2.31,
which is beyond the critical boundary of 1.96. Reject the null hypothesis and conclude
that the supplement has a significant effect on cognitive performance.
b. Cohen’s d = 5.2/9 = 0.578
c. The blueberry supplement had a significant effect on cognitive skill scores, z = 2.31, p <
.05, d = 0.578.

20. a. With no treatment effect the distribution of sample means is centered at  = 240 with a
standard error of 10 points, and the critical boundary of z = 1.96 corresponds to a
sample mean of M = 220.4. With a 30-point treatment effect, the distribution of sample
means is centered at  = 210. In this distribution a mean of M = 220.4 corresponds to z =
1.04. The power for the test is the probability of obtaining a z-score less than 1.04, which
is p = 0.8508.
b. With a sample of n = 25, the standard error is 6 points. In this case, the critical boundary
of z = 1.96 corresponds to a sample mean of M = 228.24. With a 30-point treatment
effect, the distribution of sample means is centered at  = 210. In this distribution a mean
of M = 228.24 corresponds to z = 3.04. The power for the test is the probability of
obtaining a z-score less than 3.04, which is p = 0.9988.

22. a. The z-score increases (farther from zero).

b. Cohen’s d is not influenced by sample size.
c. Power increases.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 9: Introduction to the t Statistic
Answers to Even-Numbered Problems

2. a. The sample standard deviation describes the variability of the scores in the sample. In this
case, the standard distance between a score and the sample mean is 12 points.
b. The estimated standard error is 3 points. The standard error provides a measure of the
standard distance between a sample mean and the population mean.

4. The sample variance (s2) in the t formula changes from one sample to another and contributes
to the variability of the t statistics. A z-score uses the population variance which is constant from
one sample to another.

6. a. t = 2.821
b. t = 2.539
c. t = 2.462

8. a. M = 7 and s = √24 = 4.90

b. sM = 2.

10. a. 2.2 points

b. The sample variance is 16 and the estimated standard error is sM = 0.8.
c. For these data, t = 2.75. With df = 24 the critical value is t = ±2.064. Reject H0 and
conclude that there is a significant effect.

12. a. With s2 = 32, sM = 2 and t = –5/2 = –2.50. This is beyond the critical boundaries of
±2.365, so we reject the null hypothesis and conclude that there is a significant
treatment effect.
b. With s = 72, sM = 3 and t = –5/3 = –1.67. This value is not beyond the critical
boundaries, so there is no significant effect.
c. As the sample variability increases, the likelihood of rejecting the null hypothesis
decreases.

14. a. With df = 14, the critical values are ±2.145. For these data, the sample variance is
15, the estimated standard error is 1, and t = 4.5/1 = 4.50. Reject the null hypothesis and
conclude that the amount of time spent in the plain chamber is significantly different
from chance.
b. d = 4.5/15 = 1.16.
c. With df = 14, the t values for 90% confidence are 1.761, and the interval extends from
32.739 min to 36.261 minutes.

16. The estimated standard error is 2.1 and t = 4.9/2.1 = 2.33. For a one tailed test, the critical
value is 2.602. Fail to reject the null hypothesis, answering questions did not significantly
improve exam scores.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
18. a. Cohen’s d = 4.2/8 = 0.525. With s = 8, the estimated standard error is 2 and t = 4.2/2 =
2.10. r2 = 4.41/19.41 = 0.227.
b. Cohen’s d = 4.2/20 = 0.21. With s = 20, the estimated standard error is 5 and t = 4.2/5 =
0.84. r2 = 0.706/15.706 = 0.045.
c. Measures of effect size tend to decrease as sample variance increases.

20. a. The estimated standard error is 0.20 and t = 2.2/0.2 = 11.00. The t value is well beyond
the critical value of 2.492. Reject the null hypothesis.
b. Cohen’s d = 2.2/1 = 2.20 and r2 = 121/145 = 0.8345
c. For 95% confidence, use t = 2.064. The interval is 12.2  (2.064)0.2 and extends from
11.787 to 12.613.

22. a. H0: µ = 40. With df = 11 the critical values are t = ±2.201. For these data, M = 37, SS =
118, s2 = 10.73, the standard error is 0.95, and t = 3.16. Reject H0 and conclude that
well-being scores for people with a history of frequent moves as children are significantly
different from scores for other people.
b. Cohen’s d = 3/3.28 = 0.915.
c. The results indicate that well-being scores for people with a history of frequent moves as
children are significantly different from scores for people without frequent moves, t(11) =
3.16, p < .05, d = 0.915.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 10: The t Test for Two Independent Samples
Answers to Even-Numbered Problems

2. The standard error for the independent measures t provides an estimate of the standard
distance between a sample mean difference (M1 – M2) and the population mean difference
(μ1 – μ2). When the two samples come from populations with the same mean (when H0 is true),
the standard error indicates the standard amount of error (distance) between two sample means.

4. a. The first sample has s2 = 10 and the second has s2 = 8. The pooled variance is 108/12 = 9
(halfway between).
b. The first sample has s2 = 10 and the second has s2 = 12. The pooled variance is 108/10 =
10.8 (closer to the variance for the larger sample).

6. a. The pooled variance is 63.

b. The estimated standard error is 4.
c. A mean difference of 10 points produces t = 2.50. With df = 14 the critical boundaries
are ±2.145. Reject H0

8. a. The pooled variance is 24, the estimated standard error is 1.68, and t = 6.6/1.68 = 3.93.
Because df = 32 is not listed, use df = 30 and obtain critical values of 2.750. Reject the
null hypothesis and conclude that experiencing moderate adversity does influence well-
being.
b. Cohen’s d = 6.6/4.90 = 1.347.
c. People who experienced moderate adversity had significantly higher well-being scores
than those with little or no history of adversity, t(32) = 3.93, p < .01, d = 1.347.
10. a. The research prediction is that participants who hear the verb “smashed into” will
estimate higher speeds than those who hear the verb “hit.” For these data, the pooled
variance is 33, the estimated standard error is 2.10, and t(28) = 3.24. With df = 28 and α
= .01, the critical value is t = 2.467. The sample mean difference is in the right direction
and is large enough to be significant. Reject H0.
b. The estimated Cohen’s d = 6.8/√33 = 1.18.
c. The results show that participants who heard the verb “smashed into” estimated
significantly higher speeds than those who heard the verb “hit,” t(28) = 3.24, p < .01, d =
1.18.

12. The exercise group has a mean of M = 23 with SS = 104, and the control group had M = 20
with SS = 160. The pooled variance is 8.8, the estimated standard error is 1.05, and t = 2.86.
With df = 30, the critical value is 2.042. Reject the null hypothesis and conclude that the
exercise program has a significant effect on stride length.

14. a. The null hypothesis states that counting money versus counting paper does not affect the
perception of pain. For the money group the mean is M = 7.56 with SS = 42.22. For the
paper group, M = 11.33 with SS = 38. The pooled variance is 5.01, the standard error is
1.06, and t(16) = 3.57. With df = 16 the critical values are 2.921. Reject the null
hypothesis and conclude that there is a significant difference in the amount of pain
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 experienced after counting money versus counting paper.
b. d = 3.77/2.24 = 1.68

16. a. The two samples combined have a total of 27 participants.

b. With df = 25 and α = .05, the critical region consists of t values beyond 2.060. The t
statistic is in the critical region. Reject H0 and conclude that there is a significant
difference.
c. r2 = 4.29/29.29 = 0.146 or 14.6%

18. The homogeneity of variance assumption specifies that the variances are equal for the two
populations from which the samples are obtained. If this assumption is violated, the t statistic
can cause misleading conclusions for a hypothesis test.

20. a. The pooled variance is 15 and the estimated standard error is 2.

b. The pooled variance is 60 and the estimated standard error is 4.
c. Larger variability produces a larger standard error.

22. a. The estimated standard error for the sample mean difference is 6 points.
b. The estimated standard error for the sample mean difference is 3 points.
c. Larger samples produce a smaller standard error.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 11: The t Test for Two Related Samples
Answers to Even-Numbered Problems

2. A repeated-measures design uses the same group of participants in all of the treatment
conditions.

4. a. An independent-measures design would require two separate samples, each with 20

participants, for a total of 40 participants.
b. A repeated-measures design would use the same sample of n = 20 participants in both
treatment conditions.
c. A matched-subjects design would require two separate samples with n = 20 in each, for a
total of 40 participants.

6. The null hypothesis states that the population mean difference is zero. The sample variance is
3, the estimated standard error is 0.50, and t(11) = 2.40. With df = 11, the critical boundaries are
±2.201. Reject the null hypothesis. There is a significant difference.

8. a. The sample variance is 9, the estimated standard error is 0.75, and t(15) = 2.6/0.75 =
3.47. With critical boundaries of ±2.131, reject H0.
b. Cohen’s d = 2.6/3 = 0.867

10. a. The estimated standard error is 2.5 and t(19) = 1.92. With a critical value of 1.729,
reject the null hypothesis and conclude that attention span is significantly longer
with the medication. .
b. The confidence interval is μD = 4.8 ± 1.328(2.5) and extends from 1.48 to 8.12.

12. a. The null hypothesis states that Tai Chi has no effect on pain and stiffness. The estimated
standard error is 3.4, and t = 2.50. With df = 39, the critical boundaries are ±2.021.
Reject the null hypothesis and conclude that there is a significant effect.
b. Cohen’s d = 8.5/21.5 = 0.395.

14. a. The null hypothesis says that changing answers has no effect, H0: μD = 0. With df = 8
and α = .05, the critical values are t = ±2.306. For these data, MD = 7, SS = 288, the
standard error is 2, and t(8) = 3.50. Reject H0 and conclude that changing answers has a
significant effect on exam performance.
b. For 95% confidence use t = ±2.306. The interval extends from 2.388 to 11.612.
c. Changing answers resulted in significantly higher exam scores, t(8) = 3.50, p < .05, 95%
CI [2.388, 11.612].

16. a. For these data, MD = 14, s2 = 128, the standard error is 3.58, and t(9) = –3.91. For a two-
tailed test with α = .05 the critical boundary is t = 2.262. Reject H0 and conclude that
swearing significantly increases pain tolerance.
b. r2 = 15.288/24.288 = 0.629.
c. Swearing significantly increased the amount of time that participants could tolerate the
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 icy water, t(9) = –3.91, p < .05, r2 = 0.629.

18. a. The estimated standard error is 2 points and t(25) = 2.00. With a critical boundary of
±2.131, fail to reject the null hypothesis.
b. With MD = 10, t(15) = 5.00. With a critical boundary of ±2.131, reject the null
hypothesis.
c. The larger the mean difference, the greater the likelihood of finding a significant
difference.
20. One problem with individual differences is that the participants in one group may be
noticeably different (smarter, older, etc.) than those in another group and these differences may
explain why the groups have different means. This problem is eliminated with a repeated-
measures design because the same individuals are in both groups. The second problem is that
individual differences can increase variance. In a repeated-measures design, the individual
differences are subtracted out of the variance.

22. a Because the scores are the same as in problem 21, the results are also the same. The 2-
point mean difference has an estimated standard error of 1.73 and t = 1.16. With df = 14
there is no significant difference.
b. For the repeated-measures study, the mean difference is 2 points, SS = 152, the sample
variance is 21.71, the estimated standard error is 1.65, and t = 1.21. With df = 7 there is
no significant difference.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 12: Introduction to Analysis of Variance
Answers to Even-Numbered Problems

2. Both the F-ratio and the t statistic are comparing the actual mean differences between sample
means (numerator) with the differences that would be expected if there is no treatment effect (the
denominator if H0 is true). If the numerator is sufficiently bigger than the denominator, we
conclude that there is a significant difference between treatments.

4. SStotal = 86; SSbetween = 20; SSwithin = 66

6. a. There were 3 treatment conditions (dfbetween = k – 1 = 2).

b. A total of N = 33 individuals participated (dfwithin = 30 = N – k).

8. a. k = 4 treatment conditions.
b. The study used a total of N = 52 participants.

10.

Source SS df MS

Between treatments 18 3 6 F = 3.00

Within treatments 40 20 2

Total 58 23

12. Source SS df MS
Between Treatments 48 2 24 F = 6.00
Within Treatments 204 51 4
Total 252 53

14. Source SS df MS
Between Treatments 96 2 48 F(2, 15) = 9.01
Within Treatments 80 15 5.33
Total 176 17
With df = 2, 15, the critical value for α = .05 is 3.68. Reject the null hypothesis.

16. a. As the differences between sample means increase, MSbetween also increases, and the F-
ratio increases.
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 b. Increases in sample variability cause MSwithin to increase and, thereby, decrease the F-
ratio.

18. a. The sample variances are 0.67, 1.33, and 2.67.

b. Source SS df MS
Between Treatments 32 2 16 F(2, 9) = 10.26
Within Treatments 14 9 1.56
Total 46 11
With α = .05, the critical value is F = 4.26. Reject the null hypothesis and conclude that
there are significant differences among the three treatments.

20. a. Source SS df MS
Between Treatments 20 2 10 F(2, 27) = 5.00
Within Treatments 54 27 2
Total 74 29
b. η2 = 20/74 = 0.270.

22. a. The three sample variances are 6.00, 4.00, and 14.00
b. SSwithin = 168, dfwithin = 21, and MSwithin = 8. The average of the three sample variances is
also equal to 8.

24. a. The pooled variance is 8, the estimated standard error is 2. and t(8) = 2.00. With df =
8, the critical value is 2.306. Fail to reject the null hypothesis.
b. Source SS df MS
Between Treatments 32 1 32 F(1, 8) = 4.00
Within Treatments 48 6 8
Total 80 9
With df = 1, 6, the critical-value is 5.99. Fail to reject the null hypothesis. Note that F =
t2.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 13: Repeated-Measures Analysis of Variance
Answers to Even-Numbered Problems

2. The second stage of the repeated-measures ANOVA removes the variability due to
individual differences from the error term. The individual differences are automatically
eliminated from the numerator (between treatments) because the same subjects are used in all
treatments. To keep the F-ratio balanced, the individual differences must also be measured and
removed from the denominator.

4. The repeated-measures F-ratio will have df = 2, 28.

6. Source SS df MS
Between Treatments 50 3 16.67 F(3, 12) = 24.88
Within Treatments 32 16
Between Subjects 24 4
Error 8 12 0.67
Total 82 19
With df = 3, 12, the critical value is 3.49. Reject H0. There are significant differences
among the four distances.

8. a. The null hypothesis states that there are no differences among the three treatments.
With df = 2, 8, the critical value is 4.46.
Source SS df MS
Between Treatments 70 2 35 F(2, 8) = 35
Within Treatments 26 12
Between Subjects 18 4
Error 8 8 1
Total 96 14
Reject H0. There are significant differences among the three treatments.
b. For these data, η2 = 70/78 = 0.897.
c. The analysis of variance shows significant mean differences among the three treatments,
F(2, 8) = 35.00, p < .05, η2 = 0.897.

10. The means and standard deviations for the five delay periods are as follows:

1 month 6 months 1 year 2 years 5 years

M = 866.67 M = 816.67 M = 766.67 M = 700.00 M = 583.33
S = 81.65 s = 81.65 s = 81.65 s = 70.71 s = 60.55

Source SS df MS
Between Treatments 291,333.3 4 72,833.3 F(4, 20) = 56.75
Within Treatments 143,333.3 25
Between Subjects 117,666.7 5
Error 25,666.7 20 1,283.3
Total 434,666.7 29
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 With  = .01, the critical value is 4.43. There are significant differences.

12. Source SS df MS
Between Treatments 8 1 8 F(1, 15)) = 2.00
Within Treatments 95 30
Between Subjects 35 15
Error 60 15 4
Total 103 31

14. Source SS df MS
Between Treatments 33 3 11 F(3, 57) = 3.67
Within Treatments 230 76
Between Subjects 59 19
Error 171 57 3
Total 263 79

16. a. 3 treatments
b. 14 participants

18. The null hypothesis states that there are no differences among the three treatments,
H0: μ1 = μ2 = μ3. With df = 2, 6, the critical value is 5.14.
Source SS df MS
Between Treatments 8 2 4 F(2,6) = 6.00
Within Treatments 94 9
Between Subjects 90 3
Error 4 6 0.667
Total 102 11
Reject H0. There are significant differences among the three treatments.

20. a. The null hypothesis states that there are no differences among the three weeks.
With df = 2, 10, the critical value is 4.10.
Source SS df MS
Between Treatments 12 2 6 F(2, 10) = 15
Within Treatments 100 15
Between Subjects 96 5
Error 4 10 0.4
Total 112 17
Reject H0. There are significant differences among the three weeks.
b. For these data, η2 = 12/16 = 0.75.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 c. The analysis of variance shows significant mean differences in the number of steps
among the three weeks that were tested, F(2, 10) = 15.00, p < .05, η2 = 0.75.

22. a. For the independent-measures ANOVA, we obtain:

Source SS df MS
Between Treatments 48 2 24 F(2,15) = 3.46
Within Treatments 104 15 6.93
Total 152 17
With a critical value of 3.68 for  = .05, fail to reject the null hypothesis.
b. For the repeated-measures ANOVA,
Source SS df MS
Between Treatments 48 2 24 F(2,10) = 12.00
Within Treatments 104 15
Between Subjects 84 5
Error 20 10 2
Total 152 17
With a critical value of 4.10 for  = .05, reject the null hypothesis.
c. The repeated-measures ANOVA reduces the error variance by removing individual
differences. This increases the likelihood that the ANOVA will find significant
differences.

24. a. Source SS df MS
Between Treatments 12 1 12 F(1, 5) = 4.00
Within Treatments 50 10
Between Subjects 35 5
Error 15 5 3
Total 62 11
The F-ratio has df = 1, 5 and the critical value is 6.61. Fail to reject H0. There is no
significant difference between the two means.
b. For the t test, the mean difference is MD = 2, SS for the difference scores is 30, the
variance is 6, and the standard error is 1. t = 2/1 = 2. With df = 5, the critical value is
2.571. Fail to reject H0. There is no significant difference between the two means. Note
that F = t2.

26. a. The null hypothesis states that there is no difference between the two treatments,
H0: μD = 0. The critical region consists of t values beyond ±3.182. The mean difference
is MD = 4. The SS for the difference scores is 48, and t(3) = 2.00. Fail to reject H0.
b. The null hypothesis states that there is no difference between treatments, H0: μ1 = μ2. The
critical value is F = 10.13.
Source SS df MS
Between Treatments 32 1 32 F(1, 3) = 4.00

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Within Treatments 36 6
Between Subjects 12 3
Error 24 3 8
Total 68 7
Fail to reject H0. Note that F = t2.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 14: Two-Factor Analysis of Variance (Independent
Measures)
Answers to Even-Numbered Problems

2. One test for main effects evaluates the mean differences between the rows of the matrix and
the second test for main effects evaluates mean differences between the columns. The test for an
interaction evaluates the significance of any mean differences that are not explained by the row
and/or column differences.

4. a. The main effect for treatment is the 6-point difference between the overall column means,
M = 6 and M = 12.
b. The main effect for gender is the 4-point difference between the overall row means, M =
11 and M = 7.
c. There is no interaction. The effect of the treatment does not depend on gender. With the
treatment, scores increases by an average of 6 points for the males and also increase by
an average of 6 points for the females.

6. a. M = 18
b. M = 12

8. a. df = 1, 28
b. df = 1, 28
c. df = 1, 28

10. a. 2
b. 3
c. 2, 24

12. a. Source SS df MS
Between Treatments 100 3
A 10 1 10 F(1,36) = 2
B 90 1 90 F(1,36) = 18.00
AxB 0 1 0 F(1,36) = 0
Within Treatments 180 36 5
Total 280 39
All F-ratios have df = 1, 36 and the critical value is F = 4.11. The main effect for
factor B is significant, but factor A and the interaction are not.
b. For factor A, η2 = 10/190 = 0.053, for factor B, η2 = 90/270 = 0.333, and for the
interaction, η2 = 0.

14. a. Source SS df MS
Between Treatments 340 3

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 A 80 1 80 F(1,76) = 4.00
B 180 1 180 F(1,76) = 9.00
AxB 80 1 80 F(1,76) = 4.00
Within Treatments 1520 76 20
Total 1860 79
The critical value for all three F-ratios is 3.98 (using df = 1, 70). Both main effects and
the interaction are significant.
b. For the contact factor, eta squared is 80/1600 = 0.050. For the time factor, eta squared is
180/1700 = 0.106. For the interaction, eta squared is 80/1600 = 0.050.
c. For the noncontact athletes, there is little or no difference between the beginning of the
first season and the end of the second season, but the contact athletes show noticeably
lower scores after the second season.

16. a. Source SS df MS
Between Treatments 340 5
Pouring 60 1 60 F(1, 54) = 10.00
Temperature 280 2 140 F(2, 54) = 23.33
Interaction 0 2 0 F(2, 42) = 0
Within Treatments 324 54 6
Total 664 59
b. Temperature and pouring method both have significant effects on the bubbles in the
wine. However, the effects are independent, there is no interaction.

18. Source SS df MS
Between Treatments 72 5
A 12 1 12 F(1, 42) = 4.00
B 36 2 18 F(2, 42) = 6.00
AxB 24 2 12 F(2, 42) = 4.00
Within Treatments 126 42 3
Total 198 47

20. a. Source SS df MS
Between Treatments 128 3
Gender 0 1 0 F(1, 12) = 0
Treatments 64 1 64 F(1, 12) = 7.53
Gender x Treat. 64 1 64 F(1, 12) = 7.53
Within Treatments 102 12 8.5
Total 230 15
With df = 1, 12, the critical value for all three F-ratios is 4.75. The main effect for the
treatments and the interaction are significant, but there is no gender difference.
b. The main effect for treatments and the interaction both have η2 = 64/166 =0.3855 and

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 η2 = 0 for gender.

22. a. Source SS df MS
Between Treatments 260 5
Difficulty 120 1 120 F(1, 24) = 24.00
Arousal 80 2 40 F(2, 24) = 8.00
AxB 60 2 30 F(2, 24) = 6.00
Within Treatments 120 24 5
Total 380 29

With df = 1, 24, the critical value for the difficulty factor is 4.26. With df = 2, 24, the
critical value for the arousal factor and for the interaction is 3.40. Both main effects
and the interaction are significant.
b. For the difficulty main effect, η2 = 120/240 = 0.50. For the arousal main effect, η2 =
80/200 = 0.40. For the interaction, η2 = 60/180 = 0.333.
c. For the easy task, increased arousal consistently leads to improved performance. For the
difficulty task, a moderate increase in arousal improves performance but a high level of
arousal results in lowered performance.

24. a. Source SS df MS
Between Treatments 67.7 3
Color 25.6 1 25.6 F(1, 36) = 30.84
Gender 22.5 1 22.5 F(1, 36) = 27.11
ColorGender 19.6 1 19.6 F(1, 36) = 23.61
Within Treatments 30 36 0.83
Total 97.7 39

b. With a critical value of 4.11, all three F-ratios are significant with  = .05. The
significant interaction indicates that the effect of the background color depends on
whether the participants a judging males or females. The data show a large mean
difference for judgments of the females but almost no effect for judgments of males. It
appears that the color red only influences males judgments of females.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 15: Correlation
Answers to Even-Numbered Problems

2. The numerical value of the Pearson correlation indicates how well the data points fit a
straight line. A value of 1.00 (or –1.00) indicates a perfect linear fit and a value of zero indicates
no linear trend.

4. SP = –10

6. a. The scatter plot shows points clustered around a line sloping up to the
right. The correlation is positive and relatively high.
b. For these scores, SSX = 2, SSY = 18, and SP = 5. The correlation is r = 5/6 =
0.83.

8. a. The scatter plot shows points clustered around a line sloping down to the right. The
correlation should be around –0.90.
b. SSX = 6, SSY = 24, and SP = –10. The correlation is r = –10/12 = –0.83.

10. a. SSX = 32, SSY = 8, and SP = 8. The correlation is r = 8/16 = 0.50.

b. After adding 2 points to each score, the correlation is still r = 8/16 = 0.50.
Adding a constant to each score does not change the value of the correlation.
b. After multiplying each score by 2, SSX is multiplied by 4 and SP is doubled, but the
correlation is still r = 16/32 = 0.50.
Multiplying by a constant to does not change the value of the correlation.

12. a. For the weights, SS = 26 and for the incomes, SS = 6006. SP = –361. The
correlation is r = –0.914.
b. With n = 10, df = 8 and the critical value is 0.632 for  = .05 and 0.765 for  = .01. The
correlation is significant at either alpha level.

14. a. For these data, SS7min = 98, SScognitive = 236, and SP = 127. r = 0.835.
b. With df = 7, the critical value is 0.798. The correlation is significant.
c. r2 = 0.697 or 69.7%

16. a. No. With df = 8, the critical value is r = .632.

b. Yes. With df = 23, the critical value is r = .396.

18. The correlation between crime and amount spent is r = 0.765, the correlation between crime
and population is r = 0.933, and the correlation between amount spent and population is 0.933.
The partial correlation between crimes and amount spent is r = –0.808. Controlling population
produces a strong negative relationship between the amount spent for crime prevention and the
number of serious crimes.

20. rS = +0.960

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
22. a. Using the cognitive scores as the X variable and coding the sport as 1 or 0 for the Y
variable, the data produce SSX = 136, SSY = 4, and SP = 12. The point-biserial correlation
is r = 0.515
b. r2 = 0.265. This is the same as the r2 obtained with the t test in Chapter 10.

24. The converted data show eight people with scores of 0, 0; two people with scores of 0, 1;
three people with scores of 1, 0; and seven people with scores of 1, 1. (Note: Any two different
numbers can be used for the X and Y values.) The correlation is 0.503.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 16: Introduction to Regression
Answers to Even-Numbered Problems

2. The best fitting line is determined by the error between the predicted Y values on the line and
the actual Y values in the data. The regression equation is determined by the line with the
smallest total squared error.

4. X Ŷ
0 6
2 10
3 12
–4 –2

6. The closer the correlation is to 1.00 (or –1.00), the smaller the standard error of estimate.

8. a. The regression equation is Ŷ = 2X – 3

b. SSX = 10, SSY = 50, SP = 20, r = +0.894. Using the correlation, SSresidual = (1 - r2)SSY
= (0.201)(50) = 10.05 and the standard error of estimate is 1.59.

10. a. SSX = 12, SP = 12 Ŷ = X + 3

b. X Ŷ
3 6
5 8
2 5
2 5
4 7
1 4
4 7

12. a. b = 84/22,400 = 0.00375 and a = 3.10 – 0.00375(580) = 0.925.

b. r = 0.50 and r2 = 0.25 or 25%.
c. SSregression = 0.315 with df = 1, and SSresidual = 0.945 with df = 13. F = 0.315/0.073 =
4.32. With df = 1, 13 the F-ratio is not significant.

14. SSX = 2798.5, SP = –124.05, MX = 71.5, MY = 8.81, Ŷ = –0.044X + 11.96

16. a. SSresidual = 128 and the standard error of estimate is 4.

b. SSresidual = 168 and the standard error of estimate is 4.58.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
18. a. SSX = 8 SP = 8 Ŷ = X + 1
b. & c. –––––––––––––––––––––––––––––––––––––––––
Squared
X Y Ŷ Residual Residual
–––––––––––––––––––––––––––––––––––––––––
3 2 4 –2 4
5 6 6 0 0
4 9 5 4 16
2 3 3 0 0
5 6 6 0 0
5 4 6 –2 4
–––––––––––––––––––––––––––––––––––––––––
24 = SSresidual
d. SSY = 32
r = 8/16 = 0.50
(1 – r2) = 0.75 0.75(SSY) = 24

20. R2 = [1.1(48) + 3(56)]/280 = 0.789 or 78.9%.

22. Population by itself predicts 87% of the variance. Adding the amount spent on prevention as
a second variable produces a multiple regression R2 = 0.953 or 95.3%. The increase from
87% to 95.3% is equal to 8.3% or 0.083. With SScrime = 230, this amounts to SSadditional =
0.083(230) = 19.09 and F = 19.09/1.2 = 15.91. With df = 1, 9, the F-ratio is significant with
 = .05 or  = .01.

24. dfresidual = 24 = n – 3. n = 27

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 17: The Chi-Square Statistic: Tests for Goodness of Fit
and Independence
Answers to Even-Numbered Problems

2. a. The null hypothesis states that the gender distribution for theater goers is not different from
the distribution for the general population of students. For a sample of 55 students, the
expected frequencies are 33 females (60%) and 22 males (40%), and chi-square = 1.21.
With df = 1, the critical value is 3.84. Fail to reject H0 and conclude that the gender
distribution for theater goers is not significantly different from the distribution for the
population of students.
b. The null hypothesis states that the gender distribution for basketball fans is not different
from the distribution for the general population of students. For a sample of 140 students,
the expected frequencies are 84 females (60%) and 56 males (40%), and chi-square =
20.12. With df = 1, the critical value is 3.84. Reject H0 and conclude that the gender
distribution for basketball fans is significantly different from the distribution for the
population of students.

4. a. The null hypothesis states that the age distribution for people who get speeding tickets is
not different from the distribution for the population of licensed drivers. With df = 1, the
critical value is 3.84. The expected frequencies are 40 over age 25 and 10 under age 25,
and chi-square = 6.125. Reject the null hypothesis and conclude that the age distribution
for people who receive speeding tickets is significantly different from the distribution for
the population of drivers.
b. The null hypothesis states that the age distribution for people who get parking tickets is
not different from the distribution for the population of licensed drivers. With df = 1, the
critical value is 3.84. The expected frequencies are 40 over age 25 and 10 under age 25,
and chi-square = 2.00. Fail to reject the null hypothesis and conclude that the age
distribution for people who receive speeding tickets is not significantly different from the
distribution for the population of drivers.

6. a. The null hypothesis states that couples with the same initial do not occur more often than
would be expected by chance. For a sample of 400, the expected frequencies are 26 with
the same initial and 374 with different initials. With df = 1 the critical value is 3.84, and
the data produce a chi-square of 5.92. Reject the null hypothesis.
b. A larger sample should be more representative of the population. If the sample continues
to be different from the hypothesis as the sample size increases, eventually the difference
will be significant.

8. a. H0 states that the distribution of automobile accidents is the same as the distribution of
registered drivers: 16% under age 20, 28% age 20 to 29, and 56% age 30 or older.
With df = 2, the critical value is 5.99. The expected frequencies for these three categories
are 48, 84, and 168. Chi-square = 13.76. Reject H0 and conclude that the distribution of
automobile accidents is not identical to the distribution of registered drivers.
b. The chi-square test shows that the age distribution for people in automobile accidents is

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 significantly different from the age distribution of licensed drivers, χ2(2, n = 300) =
13.76, p < .05.

10. The null hypothesis states that there are no preferences among the three designs; p = 1/3 for
all categories. With df = 2, the critical value is 5.99. The expected frequencies are fe = 40 for all
categories, and chi-square = 8.60. Reject H0 and conclude that there are significant preferences.

12. The null hypothesis states that there is no relationship between the type of loyalty program
and whether the customers completed their cards. With df = 1, the critical value is 3.84. The
expected frequencies are:
Completed Not completed
Buy 8 get 1 14.5 35.5 50
Buy 10 with 2 free 14.5 35.5 50
29 71
Chi-square = 3.93. Reject H0.

14. a. The null hypothesis states that the distribution of weights for men is the same as the
distribution for women. The expected frequencies are 81.6 desirable and 38.4
overweight for men, and 54.4 desirable and 25.6 overweight for women. With df = 1, the
critical value is 3.84. For these data, chi-square = 5.53. Reject the null hypothesis.
b. The phi-coefficient is 0.166.
c. The chi-square test shows that the proportion of men who are overweight is significantly
greater than the proportion of women, χ2(1, N = 200) = 5.53, p < .05, φ = 0.166.

16. The null hypothesis states that there is no relationship between dream content and gender;
the distribution of aggression content should be the same for males and females. The critical
value is 9.21. The expected frequencies are:
Low Medium High
Female 8.8 8.4 6.8
Male 13.2 12.6 10.2
The chi-square statistic is 25.52. Reject H0 with α = .01 and df = 2.

18. a. The null hypothesis states that there is no relationship between the personalities of the
participants and the personalities of the avatars they create. With df = 1 and α = .05, the
critical value is 3.84. The expected frequencies are:
Participant Personality
Introverted Extroverted
Introverted Avatar 17.1 27.9 45
Extroverted Avatar 20.9 34.1 55
38 62
The chi-square statistic is 4.12. Reject H0.
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 b. The phi-coefficient is 0.203.

20. a. The null hypothesis states that littering is independent of the amount of litter already on
the ground. With df = 2, the critical value is 5.99. The expected frequencies are:
Amount of Litter
None Small Large
Litter 31.33 31.33 31.33
Not Litter 58.67 58.67 58.67

Chi-square = 25.88. Reject H0.

b. V = 0.310 (a medium effect)

22. a. The null hypothesis states that there is no relationship between the type of music and
whether the women give their phone numbers. With df = 1, the critical value is 3.84. The
expected frequencies are:
Phone Number No Number
Romantic Music 17.70 26.30 44
Neutral Music 17.30 25.70 43
35 52
Chi-square = 5.37. Reject H0.
b. The phi-coefficient is 0.248

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 Chapter 18: The Binomial Test
Answers to Even-Numbered Problems

2. H0: p = q = 1/2 (right and left are equally common). The critical boundaries are z = ±2.58.
With X = 104, μ = 72, and σ = 6, we obtain z = 5.33. Reject H0 and conclude that right- and left-
handed rats are not equally common.

4. a. H0: p = 0.065 (just chance). The critical boundaries are z = ±1.96. With X = 19, μ = 13,
and σ = 3.49, we obtain z = 1.72. Fail to reject H0. The initials of spouses are not
significantly different from what would be expected by chance.
b. With X = 38, μ = 26, and σ = 4.93, we obtain z = 2.43. Reject H0. With a larger sample,
the initials of spouses are significantly different from what would be expected by chance.

6. H0: p = .25 (the general population has the same proportion of belief as the psychotherapists).
The critical boundaries are z = ±1.96. With X = 65, μ = 48, and σ = 6, we obtain z = 2.83. Reject
H0 and conclude that the proportion of belief is significantly different for the general population
and for psychotherapists.

8. H0: p(overestimate) = 0.50 (just chance). The critical boundaries are z = ±1.96. With n =
20, X = 15, μ = 10, and σ = 2.24, we obtain z = 2.23. Because the z-score is only slightly in the
critical region, check the critical region. The lower boundary for the X = 15 interval is 14.5,
which corresponds to z = 2.01. Therefore, the entire interval is beyond 1.96. Reject H0 and
conclude that the estimates of being noticed are significantly different from chance.

10. a. With p = q = 1/2, μ = 12.5, and σ = 2.5. To be 1.96 standard deviations above the mean,
requires of score of 17.4 or higher. The next higher score is X = 18, which has real limits
of 17.5 and 18.5. Because these boundaries are both beyond the critical boundary, a
score of X = 18 is entirely in the critical region. It would take at least 18 right answers to
be significantly better than chance.
b. With p = q = 1/2, μ = 25, and σ = 3.54. To be 1.96 standard deviations above the mean,
requires of score of 31.94 or higher. A score of X = 32 has real limits of 31.5 and 32.5.
Because the lower boundary is not beyond the critical boundary, a score of X = 33 is
needed to be in the critical region. It would take at least 33 right answers to be
significantly better than chance.
c. With p = q = 1/2, μ = 37.5, and σ = 4.33. To be 1.96 standard deviations above the
mean, requires of score of 45.98 or higher. However, a score of X = 46 has real limits of
45.5 and 46.5. Because these boundaries overlap with the critical boundary, a score of
X = 46 is not entirely in the critical region. It would take at least 47 right answers to be
significantly better than chance.

12. H0: p = 0.20 (just guessing). The critical boundaries are z = ±1.96. The binomial distribution
has μ = 20 and σ = 4. With X = 27 we obtain z = 1.75. Fail to reject the null hypothesis.
Performance is not significantly better than chance.

14. a. H0: p = 1/3 and q = 2/3 (just guessing). The critical boundaries are z = ±1.96. The
© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.
 binomial distribution has μ = 12 and σ = 2.83. With X = 25 we obtain z = 4.59. Reject
H0 and conclude that the children with autism are performing significantly better than
chance.
b. H0: p = 1/3 and q = 2/3 (just guessing). The critical boundaries are z = ±1.96. The
binomial distribution has μ = 12 and σ = 2.83. With X = 16 we obtain z = 1.41. Fail to
reject H0 and conclude that the children with SLI are not performing significantly better
than chance.

16. a. H0: p = ½ = p(improving). The critical boundaries are z = ±1.96. Discarding the 12
patients showing no change, we have n = 100. The binomial distribution has μ = 50 and
σ = 5. With X = 64 we obtain z = 2.80. Reject H0 and conclude that there is significant
evidence of improving symptoms.
b. H0: p = ½ = p(improving). The critical boundaries are z = ±1.96. Dividing the 12
patients between the two groups, we have n = 112. and X = 70. The binomial distribution
has μ = 56 and σ = 5.29. With X = 70 we obtain z = 2.65. Again, we reject H0 and
conclude that there is significant evidence of improving symptoms.

18. H0: p = ½ = p(improved attendance). The critical boundaries are z = ±1.96.

The binomial distribution has μ = 12 and σ = 2.45. With X = 19 we obtain z = 2.86. Reject the
null hypothesis. There is a significant effect.

20. H0: p = q = 1/2 (higher and lower grades are equally likely). The critical boundaries are z =
±2.58. The binomial distribution has μ = 20 and σ = 3.16. With X = 26 we obtain z = 1.90. Fail
to reject H0 and conclude that there are not significantly more higher grades than would be
expected by chance.

22. H0: p = q = 1/2 (no difference between the two stimulus presentations). With  = .01, the
critical boundaries are z = ±2.58. The binomial distribution has μ = 15 and σ = 2.74. With X =
22 we obtain z = 2.55. Fail to reject H0 and conclude that there is no significant difference
between the first and second presentations.

24. H0: p = q = 1/2 (any change in grade point average is due to chance). The critical
boundaries are z = ±2.58. The binomial distribution has μ = 22.5 and σ = 3.35. With X = 31 we
obtain z = 2.54. Fail to reject H0 and conclude that there is no significant change in grade point
average after the workshop.

© 2017 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part,
except for use as permitted in a license distributed with a certain product or service or otherwise on a password-
protected website for classroom use.