1
Chapter 10: Comparing Two Groups
Section 10.1: Categorical Response: How Can We Compare Two Proportions?
2
Learning Objectives
1. Bivariate Analyses2. Independent Samples and Dependent Samples3. Categorical Response Variable4. Example5. Standard Error for Comparing Two Proportions6. Confidence Interval for the Difference Between
Two Population Proportions7. Interpreting a Confidence Interval for a
Difference of Proportions
3
Learning Objectives
9. Significance Tests Comparing Population Proportions
10.Examples
11.Class Exercises
4
Learning Objective 1:Bivariate Analyses
Methods for comparing two groups are special cases of bivariate statistical methods: there are two variables The outcome variable on which comparisons are
made is the response variable
The binary variable that specifies the groups is the explanatory variable
Statistical methods analyze how the outcome on the response variable depends on or is explained by the value of the explanatory variable
5
Learning Objective 2:Independent Samples
Most comparisons of groups use independent samples from the groups:
The observations in one sample are independent of those in the other sample Example: Randomized experiments that
randomly allocate subjects to two treatments
Example: An observational study that separates subjects into groups according to their value for an explanatory variable
6
Learning Objective 2:Dependent Samples
Dependent samples result when the data are matched pairs – each subject in one sample is matched with a subject in the other sample Example: set of married couples, the men
being in one sample and the women in the other.
Example: Each subject is observed at two times, so the two samples have the same subject
7
Learning Objective 3:Categorical Response Variable
For a categorical response variable Inferences compare groups in terms of
their population proportions in a particular category
We can compare the groups by the difference in their population proportions:
(p1 – p2)
8
Experiment: Subjects were 22,071 male physicians
Every other day for five years, study participants took either an aspirin or a placebo
The physicians were randomly assigned to the aspirin or to the placebo group
The study was double-blind: the physicians did not know which pill they were taking, nor did those who evaluated the results
Learning Objective 4:Example: Aspirin, the Wonder Drug
9
Results displayed in a contingency table:
Learning Objective 4:Example: Aspirin, the Wonder Drug
10
What is the response variable? The response variable is whether the
subject had a heart attack, with categories ‘yes’ or ‘no’
What are the groups to compare? The groups to compare are:
Group 1: Physicians who took a placebo Group 2: Physicians who took aspirin
Learning Objective 4:Example: Aspirin, the Wonder Drug
11
Estimate the difference between the two population parameters of interest p1: the proportion of the population who would
have a heart attack if they participated in this experiment and took the placebo
p2: the proportion of the population who would have a heart attack if they participated in this experiment and took the aspirin
Learning Objective 4:Example: Aspirin, the Wonder Drug
12
008.0009.0017.0)ˆˆ(
009.011037/104ˆ
017.011034/189ˆ
21
2
1
pp
p
p
Sample Statistics:
Learning Objective 4:Example: Aspirin, the Wonder Drug
13
To make an inference about the difference of population proportions, (p1 – p2), we need to learn about the variability of the sampling distribution of: )ˆˆ(
21pp
Learning Objective 4:Example: Aspirin, the Wonder Drug
14
Learning Objective 5:Standard Error for Comparing Two Proportions
The difference, , is obtained from sample data
It will vary from sample to sample
This variation is the standard error of the sampling distribution of :
)ˆˆ(21
pp
)ˆˆ(21
pp
2
22
1
11)ˆ1(ˆ)ˆ1(ˆ
n
pp
n
ppse
15
Learning Objective 6:Confidence Interval for the Difference Between Two Population Proportions
The z-score depends on the confidence level This method requires:
Categorical response variable for two groups Independent random samples for the two
groups Large enough sample sizes so that there are at
least 10 “successes” and at least 10 “failures” in each group
2
22
1
11
21
)ˆ1(ˆ)ˆ1(ˆ)ˆˆ(
n
pp
n
ppzpp
16
Learning Objective 6:Confidence Interval Comparing Heart Attack Rates for Aspirin and Placebo 95% CI:
0.011) (0.005,or ,003.0008.011037
)009.1(009.
11034
)017.1(017.96.1)009.017(.
17
Learning Objective 6:Confidence Interval Comparing Heart Attack Rates for Aspirin and Placebo Since both endpoints of the confidence interval
(0.005, 0.011) for (p1- p2) are positive, we infer that (p1- p2) is positive
Conclusion: The population proportion of heart attacks is larger when subjects take the placebo than when they take aspirin
18
Learning Objective 6:Confidence Interval Comparing Heart Attack Rates for Aspirin and Placebo The population difference (0.005, 0.011) is small Even though it is a small difference, it may be
important in public health terms For example, a decrease of 0.01 over a 5 year
period in the proportion of people suffering heart attacks would mean 2 million fewer people having heart attacks
19
Learning Objective 6:Confidence Interval Comparing Heart Attack Rates for Aspirin and Placebo The study used male doctors in the U.S
The inference applies to the U.S. population of male doctors
Before concluding that aspirin benefits a larger population, we’d want to see results of studies with more diverse groups
20
Learning Objective 7:Interpreting a Confidence Interval for a Difference of Proportions Check whether 0 falls in the CI If so, it is plausible that the population
proportions are equal If all values in the CI for (p1- p2) are positive, you
can infer that (p1- p2) >0 If all values in the CI for (p1- p2) are negative,
you can infer that (p1- p2) <0 Which group is labeled ‘1’ and which is labeled
‘2’ is arbitrary
21
Learning Objective 7:Interpreting a Confidence Interval for a Difference of Proportions The magnitude of values in the confidence
interval tells you how large any true difference is
If all values in the confidence interval are near 0, the true difference may be relatively small in practical terms
22
Learning Objective 8:Significance Tests Comparing Population Proportions
1. Assumptions:
Categorical response variable for two groups
Independent random samples
23
Assumptions (continued):
Significance tests comparing proportions use the sample size guideline from confidence intervals: Each sample should have at least about 10 “successes” and 10 “failures”
Two–sided tests are robust against violations of this condition At least 5 “successes” and 5 “failures” is adequate
Learning Objective 8:Significance Tests Comparing Population Proportions
24
Learning Objective 8:Significance Tests Comparing Population Proportions2. Hypotheses: The null hypothesis is the hypothesis of no
difference or no effect:
H0: p1=p2 The alternative hypothesis is the hypothesis of
interest to the investigator
Ha: p1≠p2 (two-sided test)
Ha: p1<p2 (one-sided test)
Ha: p1>p2 (one-sided test)
25
Learning Objective 8:Significance Tests Comparing Population ProportionsPooled Estimate Under the presumption that p1= p2, we
estimate the common value of p1 and p2 by the proportion of the total sample in the category of interest
• This pooled estimate is calculated by combining the number of successes in the two groups and dividing by the combined sample size (n1+n2)
26
Learning Objective 8:Significance Tests Comparing Population Proportions
3. The test statistic is:
where is the pooled estimate
z ( ˆ p 1 ˆ p 2) 0
ˆ p (1 ˆ p )1
n1
1
n2
ˆ p
27
Learning Objective 8:Significance Tests Comparing Population Proportions
4. P-value: Probability obtained from the standard normal table of values even more extreme than observed z test statistic
5. Conclusion: Smaller P-values give stronger evidence against H0 and supporting Ha
28
Learning Objective 9:Example: Is TV Watching Associated with Aggressive Behavior?
Various studies have examined a link between TV violence and aggressive behavior by those who watch a lot of TV
A study sampled 707 families in two counties in New York state and made follow-up observations over 17 years
The data shows levels of TV watching along with incidents of aggressive acts
29
Learning Objective 9:Example: Is TV Watching Associated with Aggressive Behavior?
30
Learning Objective 9:Example: Is TV Watching Associated with Aggressive Behavior? Define Group 1 as those who watched less than
1 hour of TV per day, on the average, as teenagers
Define Group 2 as those who averaged at least 1 hour of TV per day, as teenagers
p1 = population proportion committing aggressive acts for the lower level of TV watching
p2 = population proportion committing aggressive acts for the higher level of TV watching
31
Test the Hypotheses:
H0: (p1- p2) = 0
Ha: (p1- p2) ≠ 0
using a significance level of 0.05 Test statistic:
Learning Objective 9:Example: Is TV Watching Associated with Aggressive Behavior?
ˆ p 5 154
88 6190.225
se0 ˆ p 1 ˆ p 1
n1
1
n2
0.225(0.775)
1
88
1
619
0.0476
z ˆ p 1 ˆ p 2
se0
0.057 0.2490.0476
0.1920.0476
4.04
32
Learning Objective 9:Example: Is TV Watching Associated with Aggressive Behavior?
33
Conclusion: Since the P-value is less than 0.05, we reject H0
We conclude that the population proportions of aggressive acts differ for the two groups
The sample values suggest that the population proportion is higher for the higher level of TV watching
Learning Objective 9:Example: Is TV Watching Associated with Aggressive Behavior?
34
A university financial aid office polled a simple random sample of undergraduate students to study their summer employment.
Not all students were employed the previous summer. Here are the results:
Is there evidence that the proportion of male students who had summer jobs differs from the proportion of female students who had summer jobs?
Summer Status Men Women
Employed 718 593
Not Employed 79 139
Total 797 732
Learning Objective 9:Test of Significance: Two Proportions Summer Jobs Example
35
Null: The proportion of male students who had summer jobs is the same as the proportion of female students who had summer jobs. [H0: p1 = p2]
Alt: The proportion of male students who had summer jobs differs from the proportion of female students who had summer jobs. [Ha: p1 ≠ p2]
Hypotheses:
Learning Objective 9:Test of Significance: Two Proportions Summer Jobs Example
36
n1 = 797 and n2 = 732 (both large, so test statistic follows a Normal distribution) Pooled sample proportion:
Test statistic:
Test Statistic:
Learning Objective 9:Test of Significance: Two Proportions Summer Jobs Example
37
Hypotheses: H0: p1 = p2
Ha: p1 ≠ p2
Test Statistic:
z = 5.07 P-value:
P-value = 2P(Z > 5.07) = 0.000000396 (using a computer)
Conclusion:
Since the P-value is quite small, there is very strong evidence that the proportion of male students who had summer jobs differs from that of female students.
Learning Objective 9:Test of Significance: Two Proportions Summer Jobs Example
38
Learning Objective 9:Test of Significance: Two Proportions Drinking and unplanned sex
In a study of binge drinking, the percent who said they had engaged in unplanned sex because of drinking was 19.2% out of 12708 in 1993 and 21.3% out of 8783 in 2001
Is this change statistically significant at the 0.05 significance level?
The P-value is 0.0002 < .05. The results are statistically significant. But are they practically significant?
39
Learning Objective 10:Test of Significance: Two ProportionsClass Exercise 1 A survey of one hundred male and one hundred
female high school seniors showed that thirty-five percent of the males and twenty-nine percent of the females had used marijuana previously. Does this survey indicate a difference in proportions for the population of high school seniors? Test at α=5%,
40
Learning Objective 10:Test of Significance: Two ProportionsClass Exercise 2 A random sample of 500
persons were questioned regarding political affiliation and attitude toward government sponsored mandatory testing of AIDS. The results were as follows:
favor Undecided Opposed Total
Dem 135 80 65 200
Rep 95 60 65 220
Total 230 140 130
Is there a difference in the proportions of Democrats and Republicans who are undecided regarding mandatory testing for AIDS? Test at α=5%
41
Chapter 10: Comparing Two Groups
Section10.2: Quantitative Response: How Can We Compare Two Means?
42
Learning Objectives
1. Comparing Means2. Standard Error for Comparing Two Means3. Confidence Interval for the Difference
between Two Population Means4. Example: Nicotine – How Much More
Addicted Are Smokers than Ex-Smokers?5. How Can We Interpret a Confidence Interval
for a Difference of Means?6. Significance Tests Comparing Population
Means
43
Learning Objective 1:Comparing Means
We can compare two groups on a quantitative response variable by comparing their means
44
Learning Objective 1:Example: Teenagers Hooked on Nicotine
A 30-month study: Evaluated the degree of addiction that
teenagers form to nicotine 332 students who had used nicotine
were evaluated The response variable was constructed
using a questionnaire called the Hooked on Nicotine Checklist (HONC)
45
The HONC score is the total number of questions to which a student answered “yes” during the study
The higher the score, the more hooked on nicotine a student is judged to be
Learning Objective 1:Example: Teenagers Hooked on Nicotine
46
The study considered explanatory variables, such as gender, that might be associated with the HONC score
Learning Objective 1:Example: Teenagers Hooked on Nicotine
47
How can we compare the sample HONC scores for females and males?
We estimate (µ1 - µ2) by ( ):
2.8 – 1.6 = 1.2
On average, females answered “yes” to about one more question on the HONC scale than males did
Learning Objective 1:Example: Teenagers Hooked on Nicotine
x 1 x 2
48
To make an inference about the difference between population means, (µ1 – µ2), we need to learn about the variability of the sampling distribution of:
)(21
xx
Learning Objective 1:Example: Teenagers Hooked on Nicotine
49
Learning Objective 2:Standard Error for Comparing Two Means
The difference, , is obtained from sample data. It will vary from sample to sample.
This variation is the standard error of the sampling distribution of :
)xx(21
)xx(21
2
2
2
1
2
1
n
s
n
sse
50
Learning Objective 3:Confidence Interval for the Difference Between Two Population Means
A confidence interval for 1 – 2 is:
t.025 is the critical value for a 95% confidence level from the t distribution The degrees of freedom are calculated using software. If you are not using
software, you can take df to be the smaller of (n1-1) and (n2-1) as a “safe” estimate
x 1 x 2 t.025
s12
n1
s2
2
n2
51
Learning Objective 3:Confidence Interval for the Difference between Two Population Means This method assumes:
Independent random samples from the two groups
An approximately normal population distribution for each group
this is mainly important for small sample sizes, and even then the method is robust to violations of this assumption
52
Learning Objective 4:Example: Nicotine – How Much More Addicted Are Smokers than Ex-Smokers?
Data as summarized by HONC scores for the two groups:
Smokers: = 5.9, s1 = 3.3, n1 = 75
Ex-smokers: = 1.0, s2 = 2.3, n2 = 257
x 1
x 2
53
Were the sample data for the two groups approximately normal? Most likely not for Group 2 (based on the
sample statistics: = 1.0, s2 = 2.3) Since the sample sizes are large, this lack
of normality is not a problem
Learning Objective 4:Example: Nicotine – How Much More Addicted Are Smokers than Ex-Smokers?
x 2
54
95% CI for (µ1- µ2):
We can infer that the population mean for the smokers is between 4.1 higher and 5.7 higher than for the ex-smokers
)7.5 ,1.4( ,8.09.4257
3.2
75
3.3985.1)19.5(
22
or
Learning Objective 4:Example: Nicotine – How Much More Addicted Are Smokers than Ex-Smokers?
55
Learning Objective 4:Example: Exercise and Pulse Rates
A study is performed to compare the mean resting pulse rate of adult subjects who exercise regularly to the mean resting pulse rate of those who do not exercise regularly.
This is an example of when to use the two-sample t procedures.
n mean std. dev.
Exercisers 29 66 8.6
Non-exercisers 31 75 9.0
56
Learning Objective 4:Example: Exercise and Pulse Rates
Find a 95% confidence interval for the difference in population means (non-exercisers minus exercisers).
Note: we use the “safe” estimate of 29-1=28 for our degrees of freedom in this calculation
“We are 95% confident that the difference in mean resting pulse rates (non-exercisers minus exercisers) is between 4.35 and 13.65 beats per minute.”
2
22
1
21
21 ns
ns
txx
75 66 2.048(9.0)2
31
(8.6)2
29
9 4.65
4.35 to 13.65
57
Learning Objective 4:Class Exercise 1
Attitude toward mathematics was measured for two different groups. The attitude scores range from 0 to 80 with the higher scores indicating a more positive attitude. The first group consisted of Elementary education majors and the other group consisted of majors from several other areas. The results were as follows: N mean SD
Elementary Ed 75 42.7 15.5 Non Elem. Ed 110 49.3 17.0
Find a 95% confidence interval for µ1 - µ2
58
Learning Objective 4:Class Exercise 2
Are girls less inclined to enroll in science courses than boys? One recent study of fourth, fifth, and sixth graders asked how many science courses they intended to take. The resulting data were used to compute the following summary statistics:
Calculate a 99% confidence interval for the difference between males and females in mean number of science courses planned
n Mean SD
Males 203 3.42 1.49
Females 224 2.42 1.35
59
Learning Objective 5:How Can We Interpret a Confidence Interval for a Difference of Means? Check whether 0 falls in the interval When it does, 0 is a plausible value for (µ1 – µ2),
meaning that it is possible that µ1 = µ2
A confidence interval for (µ1 – µ2) that contains only positive numbers suggests that (µ1 – µ2) is positive We then infer that µ1 is larger than µ2
60
Learning Objective 5:How Can We Interpret a Confidence Interval for a Difference of Means?
A confidence interval for (µ1 – µ2) that contains only negative numbers suggests that (µ1 – µ2) is negative We then infer that µ1 is smaller than µ2
Which group is labeled ‘1’ and which is labeled ‘2’ is arbitrary
61
Learning Objective 6:Significance Tests Comparing Population Means
1. Assumptions:
Quantitative response variable for two groups
Independent random samples
62
Learning Objective 6:Significance Tests Comparing Population Means
Assumptions (continued):
Approximately normal population distributions for each group This is mainly important for small sample sizes,
and even then the two-sided t test is robust to violations of this assumption
63
Learning Objective 6:Significance Tests Comparing Population Means
2. Hypotheses:
The null hypothesis is the hypothesis of no difference or no effect:
H0: (µ1- µ2) =0
64
Learning Objective 6:Significance Tests Comparing Population Proportions
2. Hypotheses (continued):
The alternative hypothesis:
Ha: (µ1- µ2) ≠ 0 (two-sided test)
Ha: (µ1- µ2) < 0 (one-sided test)
Ha: (µ1- µ2) > 0 (one-sided test)
65
Learning Objective 6:Significance Tests Comparing Population Means
3. The test statistic is:
t (x 1 x 2) 0
s12
n1
s22
n2
Note change from “z” to “t” in formula
66
Learning Objective 6:Significance Tests Comparing Population Means
4. P-value: Probability obtained from the standard normal table
5. Conclusion: Smaller P-values give stronger evidence against H0 and supporting Ha
67
Learning Objective 6:Example: Does Cell Phone Use While Driving Impair Reaction Times? Experiment:
64 college students
32 were randomly assigned to the cell phone group
32 to the control group
68
Experiment (continued): Students used a machine that simulated
driving situations At irregular periods a target flashed red or
green Participants were instructed to press a
“brake button” as soon as possible when they detected a red light
Learning Objective 6:Example: Does Cell Phone Use While Driving Impair Reaction Times?
69
For each subject, the experiment analyzed their mean response time over all the trials
Averaged over all trials and subjects, the mean response time for the cell-phone group was 585.2 milliseconds
The mean response time for the control group was 533.7 milliseconds
Learning Objective 6:Example: Does Cell Phone Use While Driving Impair Reaction Times?
70
Boxplots of data:
Learning Objective 6:Example: Does Cell Phone Use While Driving Impair Reaction Times?
71
Test the hypotheses:
H0: (µ1- µ2) =0
vs.
Ha: (µ1- µ2) ≠ 0
using a significance level of 0.05
Learning Objective 6:Example: Does Cell Phone Use While Driving Impair Reaction Times?
72
Learning Objective 6:Example: Does Cell Phone Use While Driving Impair Reaction Times?
P-Value
73
Conclusion: The P-value is less than 0.05, so we can
reject H0
There is enough evidence to conclude that the population mean response times differ between the cell phone and control groups
The sample means suggest that the population mean is higher for the cell phone group
Learning Objective 6:Example: Does Cell Phone Use While Driving Impair Reaction Times?
74
What do the box plots tell us? There is an extreme outlier for the cell
phone group It is a good idea to make sure the results of
the analysis aren’t affected too strongly by that single observation
Delete the extreme outlier and redo the analysis
In this example, the t-statistic changes only slightly
Learning Objective 6:Example: Does Cell Phone Use While Driving Impair Reaction Times?
75
Insight: In practice, you should not delete outliers
from a data set without sufficient cause (i.e., if it seems the observation was incorrectly recorded)
It is however, a good idea to check for sensitivity of an analysis to an outlier
If the results change much, it means that the inference including the outlier is on shaky ground
Learning Objective 6:Example: Does Cell Phone Use While Driving Impair Reaction Times?
76
Learning Objective 6:Example: Females or males more nicotine dependent Test the claim that there is a difference
between males and females and their level of dependence on nicotine with a level of significance of 1%
Mean S NFemale 2.8 3.6 150
Male 1.6 2.9 182
We would reject the claim
at a 1% level of significance
77
Learning Objective 6:Class exercise 1
Many people take ginkgo supplements advertised to improve memory. Are these over the counter supplements effective?
Based on the study results below, is there evidence that taking 40 mg of ginkgo 3 times a day is effective in increasing mean performance?
Test the relevant hypothesis using α=5%n Mean S
Ginkgo 104 5.6 0.6
Placebo 115 5.5 0.6
78
Learning Objective 6:Class Exercise 2
Attitude toward mathematics was measured for two different groups. The attitude scores range from 0 to 80 with the higher scores indicating a more positive attitude. One group consisted of Elementary education majors and the other group consisted of majors from several other areas. The results were as follows: N mean SD
Elementary Ed 75 42.7 15.5 Non Elem. Ed 110 49.3 17.0
Calculate the P-value, and give your conclusion for testing
H0: µ1 - µ2 = 0, Ha: µ1 - µ2 < 0 at a level of significance equal to 0.05.
79
Chapter 10: Comparing Two Groups
Section 10.3: Other Ways of Comparing Means and Comparing Proportions
80
Learning Objectives
1. Alternative Method for Comparing Means: the Pooled Standard Deviation
2. Comparing Population Means, Assuming Equal Population Standard Deviations
3. Examples
4. The Ratio of Proportions: The Relative Risk
81
Learning Objective 1:Alternative Method for Comparing Means
An alternative t- method can be used when, under the null hypothesis, it is reasonable to expect the variability as well as the mean to be the same
This method requires the assumption that the population standard deviations be equal
82
Learning Objective 1:The Pooled Standard Deviation
This alternative method estimates the common value σ of σ1 and σ1 by:
2
)1()1(
21
2
22
2
11
nn
snsns
83
Learning Objective 2:Comparing Population Means, Assuming Equal Population Standard Deviations Using the pooled standard deviation estimate, a
95% CI for (µ1 - µ2) is:
This method has df =n1+ n2- 2
21
025.21
11)(
nnstxx
84
Learning Objective 2:Comparing Population Means, Assuming Equal Population Standard Deviations The test statistic for H0: µ1=µ2 is:
This method has df =n1+ n2- 2
21
21
11)(
nns
xxt
85
Learning Objective 2:Comparing Population Means, Assuming Equal Population Standard Deviations These methods assume:
Independent random samples from the two groups
An approximately normal population distribution for each group
This is mainly important for small sample sizes, and even then, the CI and the two-sided test are usually robust to violations of this assumption
σ1=σ2
86
Learning Objective 3:Example: Is Arthroscopic Surgery better than Placebo? Calculate the P-Value and determine if there is a
statistical difference between Arthroscopic surgery and Placebo at 5% level of significance.
With a P-value of 0.63, we should not reject the null that there is no difference between placebo and Arthroscopic surgery
87
Learning Objective 3:Example: Is Arthroscopic Surgery better than Placebo? Calculate a 95% Confidence Interval
We are 95% Confident that the difference between the placebo and surgery is in this range -10.6 to 6.4.
Notice that 0 is within this range. Thus, we should not reject the null hypothesis at the 5% significance level that there is no difference between the two treatment groups
88
Learning Objective 3:Example: Are Vegetarians More Liberal?
Respondents were rated on a scale of 1-7 with 1 being liberal and 7 being the most conservative. Is there a significant difference between Non-vegetarian and vegetarians? Assume equal variances.
H0: μ(nveg)= μ(veg) vs. Ha: μ(nveg)≠ μ(veg)
Mean S NNonvegetarian 3.18 1.72 51
Vegetarian 2.22 0.67 9
89
Learning Objective 3:Example: Are Vegetarians More Liberal?
Depending on your assumption on whether the variance of both groups are equal or not impacts the conclusion of statistical significance.
Without assumption of equal variances:
90
Learning Objective 3:Example: Are Vegetarians More Liberal?
Calculate a 95% confidence interval
Assuming Equal Variances
91
Assuming unequal variances, what is the 95% Confidence Interval?
Learning Objective 3:Example: Are Vegetarians More Liberal?
92
Learning Objective 4:The Ratio of Proportions: The Relative Risk
The ratio of proportions for two groups is:
In medical applications for which the proportion refers to a category that is an undesirable outcome, such as death or having a heart attack, this ratio is called the relative risk
The ratio describes the sizes of the proportions relative to each other
2
1
ˆˆ
pp
93
Learning Objective 4:The Ratio of Proportions: The Relative Risk
Recall Physician’s Health Study:
The proportion of the placebo group who had a heart attack was 1.82 times the proportion of the aspirin group who had a heart attack.
ˆ p 1 189/11034 0.0171
ˆ p 2 104 /11037 0.0094
sample relative risk = ˆ p 1 ˆ p 2 0.0171 0.0094 1.82
94
Chapter 10: Comparing Two Groups
Section 10.4: How Can We Analyze Dependent Samples?
95
Learning Objectives
1. Dependent Samples
2. Example: Matched Pairs Design for Cell Phones and Driving Study
3. To Compare Means with Matched Pairs, Use Paired Differences
4. Confidence Interval For Dependent Samples
5. Paired Difference Inferences
96
Learning Objectives
6. Comparing Proportions with Dependent Samples
7. Confidence Interval Comparing Proportions with Matched-Pairs Data
8. McNemar’s Test
97
Learning Objective 1:Dependent Samples
Each observation in one sample has a matched observation in the other sample
The observations are called matched pairs
98
Learning Objective 2:Example: Matched Pairs Design for Cell Phones and Driving Study
The cell phone analysis presented earlier in this text used independent samples:
One group used cell phones
A separate control group did not use cell phones
99
An alternative design used the same subjects for both groups
Reaction times are measured when subjects performed the driving task without using cell phones and then again while using cell phones
Learning Objective 2:Example: Matched Pairs Design for Cell Phones and Driving Study
100
Data:
Learning Objective 2:Example: Matched Pairs Design for Cell Phones and Driving Study
101
Benefits of using dependent samples (matched pairs): Many sources of potential bias are
controlled so we can make a more accurate comparison
Using matched pairs keeps many other factors fixed that could affect the analysis
Often this results in the benefit of smaller standard errors
Learning Objective 2:Example: Matched Pairs Design for Cell Phones and Driving Study
102
To Compare Means with Matched Pairs, Use Paired Differences:
For each matched pair, construct a difference score
d = (reaction time using cell phone) – (reaction time without cell phone)
Calculate the sample mean of these differences:
Learning Objective 3:To Compare Means with Matched Pairs, Use Paired Differences
x d
103
Learning Objective 3:To Compare Means with Matched Pairs, Use Paired Differences The difference ( – ) between the means
of the two samples equals the mean of the difference scores for the matched pairs
The difference (µ1 – µ2) between the population means is identical to the parameter µd that is the population mean of the difference scores
x 1
x 2
x d
104
Learning Objective 4:Confidence Interval For Dependent Samples
Let n denote the number of observations in each sample
This equals the number of difference scores The 95 % CI for the population mean
difference is:
deviation standard their is s
sdifference theofmean sample theis
d
025.
d
dd
xn
stx
105
Learning Objective 5:Paired Difference Inferences
These paired-difference inferences are special cases of single-sample inferences about a population mean so they make the same assumptions
106
Learning Objective 5:Paired Difference Inferences
To test the hypothesis H0: µ1 = µ2 of equal means, we can conduct the single-sample test of H0: µd = 0 with the difference scores
The test statistic is:
1 with 0 ndf
ns
xt
d
d
107
Learning Objective 5:Paired Difference Inferences
Assumptions: The sample of difference scores is a
random sample from a population of such difference scores
The difference scores have a population distribution that is approximately normal
This is mainly important for small samples (less than about 30) and for one-sided inferences
108
Learning Objective 5:Paired Difference Inferences
Confidence intervals and two-sided tests are robust: They work quite well even if the normality assumption is violated
One-sided tests do not work well when the sample size is small and the distribution of differences is highly skewed
109
Learning Objective 5:Example: Cell Phones and Driving Study
The box plot shows skew to the right for the difference scores Two-sided inference is robust to violations
of the assumption of normality The box plot does not show any severe
outliers
110
Significance test: H0: µd = 0 (and hence equal population
means for the two conditions) Ha: µd ≠ 0
Test statistic:
46.5
325.52
6.50 t
Learning Objective 5:Example: Cell Phones and Driving Study
111
Learning Objective 5:Example: Cell Phones and Driving Study
112
The P-value displayed in the output is approximately 0
There is extremely strong evidence that the population mean reaction times are different
Learning Objective 5:Example: Cell Phones and Driving Study
113
95% CI for µd =(µ1 - µ2):
69.5) (31.7,or
18.950.6 )32
5.52(040.26.50
Learning Objective 5:Example: Cell Phones and Driving Study
114
We infer that the population mean when using cell phones is between about 32 and 70 milliseconds higher than when not using cell phones
The confidence interval is more informative than the significance test, since it predicts possible values for the difference
Learning Objective 5:Example: Cell Phones and Driving Study
115
Learning Objective 6:Comparing Proportions with Dependent Samples
A recent GSS asked subjects whether they believed in Heaven and whether they believed in Hell:
Belief in Hell
Belief in Heaven Yes No Total
Yes 833 125 958
No 2 160 162
Total 835 285 1120
116
Learning Objective 6:Comparing Proportions with Dependent Samples
We can estimate p1 - p2 as:
Note that the data consist of matched pairs. Recode the data so that for belief in heaven or
hell, 1=yes and 0=no
ˆ p 1 ˆ p 2 958 1120 835 1120 0.11
Heaven Hell Interpretation Difference, d Frequency
1 1 believe in Heaven and Hell 1-1=0 833
1 0 believe in Heaven, not Hell 1-0=1 125
0 1 believe in Hell, not Heaven 0-1=-1 2
0 0 do not believe in Heaven or Hell 0-0=0 160
117
Learning Objective 6:Comparing Proportions with Dependent Samples
Sample mean of the 1120 difference scores is
[0(833)+1(125)-1(2)+0(160)]/1120=0.11 Note that this equals the difference in
proportions We have converted the two samples of binary
observations into a single sample of 1120 difference scores. We can now use single-sample methods with the differences as we did for the matched-pairs analysis of means.
ˆ p 1 ˆ p 2
118
Learning Objective 7:Confidence Interval Comparing Proportions with Matched-Pairs Data Use the fact that the sample difference is
the mean of difference scores of the re-coded data
We can then find a confidence interval for the population mean of difference scores using single sample methods
ˆ p 1 ˆ p 2
119
Learning Objective 7:Confidence Interval Comparing Proportions with Matched-Pairs Data
n 1120
x d 0.1098
sd 0.3185
95% CI = 0.1098 1.96 0.3185 1120 0.1098 0.0187
(0.091, 0.128)
120
Learning Objective 8:McNemar Test for Comparing Proportions with Matched-Pairs Data Hypotheses: H0: p1=p2, Ha can be one or two
sided Test Statistic: For the two counts for the
frequency of “yes” on one response and “no” on the other, the z test statistic equals their difference divided by the square root of their sum.
P-value: The probability of observing a sample even more extreme than the observed sample
121
Learning Objective 8:McNemar Test for Comparing Proportions with Matched-Pairs Data Assumptions:
The sum of the counts used in the test should be at least 30, but in practice, the two-sided test works well even if this is not true.
122
Learning Objective 8:Example: McNemar’s Test
Recall GSS example about belief in Heaven and Hell:
Belief in Hell
Belief in Heaven Yes No Total
Yes 833 125 958
No 2 160 162
Total 835 285 1120
123
Learning Objective 8:Example: McNemar’s Test
McNemar’s Test:
P-value is approximately 0. Note that this result agrees with the
confidence interval for p1-p2 calculated earlier
z 125 2
125 210.9
124
Chapter 10: Comparing Two Groups
Section 10.5: How Can We Adjust for Effects of Other Variables?
125
Learning Objectives
1. A Practically Significant Difference
2. Control Variable
3. Can An Association Be Explained by a Third Variable?
126
Learning Objective 1:A Practically Significant Difference
When we find a practically significant difference between two groups, can we identify a reason for the difference?
Warning: An association may be due to a lurking variable not measured in the study
127
Learning Objective 2:Control Variable
In a previous example, we saw that teenagers who watch more TV have a tendency later in life to commit more aggressive acts
Could there be a lurking variable that influences this association?
128
Perhaps teenagers who watch more TV tend to attain lower educational levels and perhaps lower education tends to be associated with higher levels of aggression
Learning Objective 2:Control Variable
129
We need to measure potential lurking variables and use them in the statistical analysis
If we thought that education was a potential lurking variable we would want to measure it
Including a potential lurking variable in the study changes it from a bivariate study to a multivariate study
A variable that is held constant in a multivariate analysis is called a control variable
Learning Objective 2:Control Variable
130
Learning Objective 2:Control Variable
131
This analysis uses three variables: Response variable: Whether the
subject has committed aggressive acts
Explanatory variable: Level of TV watching
Control variable: Educational level
Learning Objective 2:Control Variable
132
Learning Objective 3:Can An Association Be Explained by a Third Variable? Treat the third variable as a control
variable Conduct the ordinary bivariate analysis
while holding that control variable constant at fixed values (multivariate analysis)
Whatever association occurs cannot be due to the effect of the control variable
133
At each educational level, the percentage committing an aggressive act is higher for those who watched more TV
For this hypothetical data, the association observed between TV watching and aggressive acts was not because of education
Learning Objective 3:Can An Association Be Explained by a Third Variable?