Fiona Wong

nd
22 February, 2011.
David Gonzales
Psyc 2317.002
Chapter 5 Homework

1. The SD of the distribution of mean is generally smaller than the SD of the distribution of the
population of individuals b/c there is less variation among means of samples of more than
one score than there are among individual scores. This is b/c the likelihood of two extreme
scores in the same direction randomly ending up in the same sample is less than the
probability of each of those extreme scores being chosen individually.

2. (a) σ2= 102= 100; σ2M = σ2/N= 100/2 = 50; σM = √ σ2M = √50 = 7.07
(b) σ2= 102= 100; σ2M = σ2/N= 100/3 = 33.33; σM = √ σ2M = √33.33= 5.77
(c) σ2= 102= 100; σ2M = σ2/N= 100/4 = 25; σM = √ σ2M = √25= 5

(d) σ2= 102= 100; σ2M = σ2/N= 100/9; σM = √ σ2M = √100/9 = 3.33

3. (a) σ2= 202= 400; σ2M = σ2/N= 400/2 = 200; σM = √ σ2M = √200 = 14.14
(b) σ2= 202= 400; σ2M = σ2/N= 400/3 =133.33; σM = √ σ2M = √133.33 = 11.55
(c) σ2= 202= 400; σ2M = σ2/N= 400/4 = 100; σM = √ σ2M = √100 = 10

(d) σ2= 202= 400; σ2M = σ2/N= 400/9; σM = √ σ2M = √400/9 = 6.67

4. (a) The best estimate of the population mean is the sample mean of 100. From question 2 (a),
the standard deviation of the distribution of means (σM) is 7.07. For the 95% confidence
limits, the Z scores you want are -1.96 and +1.96. Lower limit = (-1.96) (7.07) + 100 = 86.14;
upper limit = (1.96)(7.07) + 100 = 113.86
(b) The best estimate of the population mean is the sample mean of 100. From question 2 (b),
the standard deviation of the distribution of means (σM) is 5.77. For the 95% confidence
limits, the Z scores you want are -1.96 and +1.96. Lower limit = (-1.96) (5.77) + 100 = 88.69;
upper limit = (1.96) (5.77) + 100 = 111.31
(c) The best estimate of the population mean is the sample mean of 100. From question 2 (c),
the standard deviation of the distribution of means (σM) is 5. For the 95% confidence limits,
the Z scores you want are -1.96 and +1.96. Lower limit = (-1.96) (5) + 100 = 90.2; upper
limit = (1.96) (5) + 100 = 109.8

(d) The best estimate of the population mean is the sample mean of 100. From question 2 (d),
the standard deviation of the distribution of means (σM) is 3.33. For the 95% confidence
limits, the Z scores you want are -1.96 and +1.96. Lower limit = (-1.96) (3.33) + 100 = 93.47;
upper limit = (1.96) (3.33) + 100 = 106.53

5. (a) The best estimate of the population mean is the sample mean of 10. From question 3 (a),
the standard deviation of the distribution of means (σM) is 14.14. For the 99% confidence
limits, the Z scores you want are -2.57 and +2.57. Lower limit = (-2.57) (14.14) + 10 =
-26.34; upper limit = (2.57) (14.14) + 10 = 46.34
 Fiona Wong
nd
22 February, 2011.
David Gonzales
Psyc 2317.002
(b) The best estimate of the population mean is the sample mean of 10. From question 3 (b),
the standard deviation of the distribution of means (σM) is 11.55. For the 99% confidence
limits, the Z scores you want are -2.57 and +2.57. Lower limit = (-2.57) (11.55) + 10 = -
19.68; upper limit = (2.57) (11.55) + 10 = 111.31
(c) The best estimate of the population mean is the sample mean of 10. From question 3 (c),
the standard deviation of the distribution of means (σM) is 10. For the 99% confidence limits,
the Z scores you want are -2.57 and +2.57. Lower limit = (-2.57) (10) + 10 = -15.70; Upper
limit = (2.57) (10) + 10 = 35.70

(d) The best estimate of the population mean is the sample mean of 10. From question 3 (d),
the standard deviation of the distribution of means (σM) is 6.67. For the 99% confidence
limits, the Z scores you want are -2.57 and +2.57. Lower limit = (-2.57) (6.67) + 10 = -7.14;
Upper limit = (2.57) (6.67) + 10 = 27.14

6. (a) i. Population 1: People give the experimental treatment

Population 2: People in general (who do not get the experimental treatment)
The research hypothesis is that the population given the experimental treatment
(Population 1) has a different mean score than people in general (Population 2). The
null hypothesis is that Population 1’s mean is the same as Population 2’s.
ii. Comparison distribution is a distribution of means of samples of 10 taken from the
distribution of Population 2. µM = µ = 40; σ2M = σ2/N = 62/10 = 3.6; σM = √ σ2M = √3.6
= 1.90. Because the population is normal, the distribution of means is normal.
iii. For a two-tailed test at the .05 level, the cutoffs are -1.96 and 1.96.
iv. Z= (44-10) / 1.90 = 2.11
v. 2.11 is more extreme than 1.96. Thus, you can reject the null hypothesis. The research
hypothesis is supported; those who receive the experimental treatment score
differently from the general population. The distributions involved are shown on the
bottom.

(b) i. Population 1: People give the experimental treatment

Population 2: People in general (who do not get the experimental treatment)
The research hypothesis is that the population given the experimental treatment
(Population 1) has a different mean score than people in general (Population 2). The
null hypothesis is that Population 1’s mean is the same as Population 2’s.
 Fiona Wong
nd
22 February, 2011.
David Gonzales
Psyc 2317.002
ii. Comparison distribution is a distribution of means of samples of 1 taken from the
distribution of Population 2. µM = µ= 40; σ2M = σ2/N = 62/10 = 3.6; σM = √ σ2M = √3.6 =
1.90. Because the population is normal, the distribution of means is normal.
iii. For a two-tailed test at the .05 level, the cutoffs are -1.96 and 1.96.
iv. Z= (48-10) / 6 = 1.33
v. 1.33 is less extreme than 1.96. Thus, you can’t reject the null hypothesis. The research
hypothesis is not supported. The distributions involved are shown on the bottom.

(c) For part (a), 95% confidence intervals: Lower limit = (-1.96) (1.9) + 44 = 40.28; upper
limit = (1.96) (1.9) + 44 = 47.72. For part (b), 95% confidence interval: Lower limit =
(-1.96) (1.33) + 44 = 36.24; Upper limit = (1.96) (1.33) + 44 = 59.76.

7. (a) i. Population 1: People give the experimental treatment

Population 2: People in general (who do not get the experimental treatment)
The research hypothesis is that the population given the experimental treatment
(Population 1) has a different mean score than people in general (Population 2). The
null hypothesis is that Population 1’s mean is the same as Population 2’s.
ii. Comparison distribution is a distribution of means of samples of 100 taken from the
distribution of Population 2. µM = µ= 82; σ2M = σ2/N = 82/100 = .36; σM = √ σ2M = √.36
= .6. Because the population is normal, the distribution of means is normal.
iii. For a two-tailed test at the .05 level, the cutoffs are -1.96 and 1.96
iv. Z= (82-81)/.8 = 1.25
v. 1.25 is less extreme than 1.96. Thus, you can’t reject the null hypothesis. The research
hypothesis is not supported. The distributions involved are shown on the bottom.

(b) i. Population 1: People give the experimental treatment

Population 2: People in general (who do not get the experimental treatment)
 Fiona Wong
nd
22 February, 2011.
David Gonzales
Psyc 2317.002
The research hypothesis is that the population given the experimental treatment
(Population 1) has a different mean score than people in general (Population 2). The
null hypothesis is that Population 1’s mean is the same as Population 2’s.
ii. Comparison distribution is a distribution of means of samples of 10 taken from the
distribution of Population 2. µM = µ=84; σ2M = σ2/N = 82/100 = .36; σM = √ σ2M = √.36 =
.6. Because the population is normal, the distribution of means is normal.
iii. For a two-tailed test at the .01 level, the cutoffs are -+2.57 and -2.57.
iv. Z = (84-81)/2.53 =1.19
v. 1.19 is less extreme than 2.57. Thus, you can’t reject the null hypothesis. The research
hypothesis is not supported. The distributions involved are shown on the bottom.

(c) For part (a), 99% confidence intervals: Lower limit = (-2.57) (1.19) + 82 = 79.94; Upper
limit = (2.57) (1.19) + 82 = 90.50.

8. (a) i. Population 1: women 70-80 who took part in a special program.

Population 2: women 70-80 who didn’t take part in the special program
The research hypothesis is that the population given the special program (Population 1)
has a different mean score than the population didn’t get the special program (Population
2). The null hypothesis is that the population special program (Population 1) does not
have a different mean score than the population didn’t get the special program
(Population 2).
ii. . Comparison distribution is a distribution of means of samples of 25 taken from the
distribution of Population 2. µM = µ= 1.5; σ2M = σ2/N = .52/25 = .01; σM = √ σ2M = √.01 = .
1
iii. For a one-tailed test at the .01 level the cutoffs are -2.57.
iv. Z= (1.5-1.8)/.1 = -3
v. -3 is more extreme than -2.57. Thus you can reject the null hypothesis. The research
hypothesis is supported. The distributions involved are shown on the bottom.
(b)
 Fiona Wong
nd
22 February, 2011.
David Gonzales
Psyc 2317.002

(c) This is a standard hypothesis-testing problem, with one exception. You can’t compare
directly the reaction time for the group of 25 women tested to a distribution of reaction
times for individual women. The probability of a group of scores having an extreme
mean just by chance is much less than the probability of any single individual having an
extreme chance just by chance. Thus, you need to compare the mean of the group of 25
reaction times to a distribution of what would happen if you were to take many random
groups of 25 reaction time scores and find the mean of each group of 25 scores. Such a
distribution of many means of samples has the same mean as the original distribution of
individual scores. However, it is a narrow curve. This is b/c the chances of extremes are
less. In fact, its variance will be exactly be the variance of the original distribution of
individuals divided by the number of scores in each sample. In this sample, this makes a
distribution of means with a mean of 1.7 and a SD of .1. This will be a normal
distribution b/c a distribution of many means from a normally distributed population is
also normal. The cutoff for significance, using the .01 level and a one tailed test is -2.33.
The mean reaction time of the group of 25 women who received the special program, 1.5
was 3 SD below the mean of distribution of means, making it clearly more extreme than
the cutoff. Thus, you can reject the null hypothesis and conclude that elderly women who
take part in the special program have lower reaction time.
(d) 99% confidence interval: 1.24 to 1.76. The confidence interval is an estimate of the range
of values that is likely to include the true population mean for the group studied. A 99%
confidence interval is the rangel of values you are 99% confident include the true
population mean. The lower limit of this interval is the mean of the lowest distribution of
means that would have a 99% chance of including this sample mean, its upper limit is the
mean of the highest distribution of means that would have a 99% chance of including this
sample mean.
9. (a) i. Population 1: Tested individuals in a condition in which the wording of the question is
changed. Population 2: Tested individual in a condition in which the wording of the
question is not changed.
The research hypothesis is that tested individuals in a condition in which the wording of the
question is changed will give a different mean than Tested individual in a condition in which
the wording of the question is not changed. The null hypothesis is that tested individuals in a
condition in which the wording of the question is changed will not give a different mean than
Tested individual in a condition in which the wording of the question is not changed
ii. . Comparison distribution is a distribution of means of samples of 25 taken from the
distribution of Population 2. µM = µ= 5.5; σM = .8.
 Fiona Wong
nd
22 February, 2011.
David Gonzales
Psyc 2317.002
iii. For a one-tailed test at the .01 level the cutoffs are 2.57.
iv. Z = (5.9 -5.5)/.2 = 2
v. You can reject the null hypothesis. The research hypothesis is supported.
(b)

(c) This is a standard hypothesis-testing problem, with one exception. You can’t compare
directly the reaction time for the group of 16 people tested the difference of reaction due
the the different wordings of the question. The probability of a group of scores having an
extreme mean just by chance is much less than the probability of any single individual
having an extreme chance just by chance. Thus, you need to compare the mean of the
group of 25 reaction times to a distribution of what would happen if you were to take
many random groups of 25 reaction time scores and find the mean of each group of
16scores. Such a distribution of many means of samples has the same mean as the
original distribution of individual scores. However, it is a narrow curve. This is b/c the
chances of extremes are less. In fact, its variance will be exactly be the variance of the
original distribution of individuals divided by the number of scores in each sample. In
this sample, this makes a distribution of means with a mean of 1.7 and a SD of .1. This
will be a normal distribution b/c a distribution of many means from a normally
distributed population is also normal. The cutoff for significance, using the .01 level and
a one tailed test is -2.33. The mean reaction time of the group of 16 people who received
the special program, 2 was 3 SD below the mean of distribution of means, making it
clearly more extreme than the cutoff. Thus, you can reject the null hypothesis and
conclude that elderly women who take part in the special program have lower reaction
time.
(d) 95% confidential interval: 5.51 to 6.29
10. The error bars are the lines that go above and below the top of each bar. The error bars
show, for each particular group, the SD of the distribution of means for people like thos
in this group.
11. 99% confidence interval: 38.8% to 40.5%. The confidence interval is an estimate of the
range of values that is likely to include the true population mean for the group studied. A
99% confidence interval is the rangel of values you are 99% confident include the true
population mean. The lower limit of this interval is the mean of the lowest distribution of
means that would have a 99% chance of including this sample mean, its upper limit is the
 Fiona Wong
nd
22 February, 2011.
David Gonzales
Psyc 2317.002
mean of the highest distribution of means that would have a 99% chance of including this
sample mean.