Announcements:

•Discussion today is review for midterm, no credit.

You may attend more than one discussion section. Estimating
•Bring 2 sheets of notes and calculator to midterm.
We will provide Scantron form. Proportions
Chapter 10
with
Homework: (Due Wed) Confidence
Chapter 10: #5, 22, 42

Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated

by Jessica Utts Feb 2010

Confidence interval example from Fri lecture Remember population versus sample:
Gallup poll of n = 1018 adults found 39% • Population proportion: the fraction of the population
believe in evolution. So p̂ = .39 that has a certain trait/characteristic or the probability
of success in a binomial experiment – denoted by p.
A 95% confidence interval or interval estimate for The value of the parameter p is not known.
the proportion (or percent) of all adults who believe in • Sample proportion: the fraction of the sample
evolution is .36 to .42 (or 36% to 42%). that has a certain trait/characteristic – denoted by p̂ .
The statistic p̂ is an estimate of p.
Confidence interval: an interval of estimates
that is likely to capture the population value. The Fundamental Rule for Using Data for Inference:
Available data can be used to make inferences about a
Goal today: Learn to calculate and interpret confidence much larger group if the data can be considered to be
intervals for p and for p1 − p2 and learn general format. representative with regard to the question(s) of interest.

Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 3 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 4
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

Some Definitions: Details for proportions:

• Point estimate: A single number used to estimate a
Sample estimate ± multiplier × standard error
population parameter. For our five situations:
point estimate = sample statistic = sample estimate
= p̂ for one proportion Parameter Sample estimate Standard error
= pˆ 1 − pˆ 2 for difference in two proportions
p̂ pˆ (1 − pˆ )
• Interval estimate: An interval of values used to p s.e.( pˆ ) =
n
estimate a population parameter. Also called a
confidence interval. For our five situations, always:
p1 – p2 pˆ 1 − pˆ 2 See p. 424 for formula
Sample estimate ± multiplier × standard error

Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 5 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 6
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010
Multiplier and Confidence Level More about the Multiplier
• The multiplier is determined by the desired confidence
level.
• The confidence level is the probability that the procedure
used to determine the interval will provide an interval that
includes the population parameter. Most common is .95.
• If we consider all possible randomly selected samples of Note: Increase confidence level => larger multiplier.
same size from a population, the confidence level is the
fraction or percent of those samples for which the Multiplier, denoted as z*, is the
confidence interval includes the population parameter. standardized score such that the
See picture on board. area between −z* and z* under the
• Often express the confidence level as a percent. Common standard normal curve corresponds
levels are 90%, 95%, 98%, and 99%. to the desired confidence level.
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 7 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 8
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

Example of different confidence levels

Formula for C.I. for proportion
Poll on belief in evolution:
n = 1018
Sample estimate ± multiplier × standard error
Sample proportion = .39
For one proportion: A confidence interval Standard error = pˆ (1 − pˆ )
=
.39(1 − .39)
= .0153
for a population proportion p, based on a n 1018
sample of size n from that population, with 90% confidence interval
.39 ± 1.65(.0153) or .39 ± .025 or .365 to .415
sample proportion p̂ is:
95% confidence interval:
.39 ± 2(.0153) or .39 ± .03 or .36 to .42
pˆ (1 − pˆ )
pˆ ± z * 99% confidence interval
n .39 ± 2.58(.0153) or .39 ± .04 or .35 to .43
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 9 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 10
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

Interpretation of the confidence

interval and confidence level: Notes about interval width
• We are 90% confident that the proportion of all adults in
the US who believe in evolution is between .365 and .415.
• Higher confidence <=> wider interval
• We are 95% confident that the proportion of all adults in
the US who believe in evolution is between .36 and .42. • Larger n (sample size) <=> more narrow
• We are 99% confident that the proportion of all adults in interval, because n is in the denominator of
the US who believe in evolution is between .35 and .43.
the standard error.
Interpreting the confidence level of 99%:
The interval .35 to .43 may or may not capture the true • So, if you want a more narrow interval you
proportion of adult Americans who believe in evolution can either reduce your confidence, or
But, in the long run this procedure will produce intervals that
capture the unknown population values about 99% of the increase your sample size.
time. So, we are 99% confident that it worked this time.
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 11 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 12
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010
 Reconciling with Chapter 3 Comparing three versions
formula for 95% confidence interval (Details on board)
Sample estimate ± Margin of error
1 For the evolution example, n = 1018, pˆ = .39
where (conservative) margin of error was
n • Conservative margin of error = .0313 ≈ .03
• Approximate margin of error using z* = 2
pˆ (1 − pˆ )
Now, “margin of error” is 2 2 × .0153 = .0306 ≈ .03
n
• Exact margin of error using z* = 1.96
These are the same when p ˆ = .5 . The new margin of error 1.96 × .0153 = .029988 ≈ .03
is smaller for any other value of p̂ So we say the old
version is conservative. It will give a wider interval.
All very close to .03, and it really doesn’t
make much difference which one we use!
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 13 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 14
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

New example: compare methods Comparing margin of error

Marist Poll in Oct 2009 asked “How often do you text • Conservative margin of error will be okay for
while driving?” n = 1026 sample proportions near .5.
Nine percent answered “Often” or “sometimes” so
• For sample proportions far from .5, closer to
and pˆ = .09 .09(.91)
s.e.( pˆ ) = = .009 0 or 1, don’t use the conservative margin of
1026 error. Resulting interval is wider than needed.
• Conservative margin of error = .0312
• Approximate margin of error = 2 × .009 = .018.
• Note that using a multiplier of 2 is called the
approximate margin of error; the exact one
This time, they are quite different!
uses multiplier of 1.96. It will rarely matter if
The conservative one is too conservative, it’s double
the approximate one! we use 2 instead of 1.96.
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 15 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 16
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

Factors that Determine Margin of Error General Description of the

Approximate 95% CI for a Proportion
1. The sample size, n.
When sample size increases, margin of error decreases. Approximate 95% CI for the population proportion:
2. The sample proportion, p̂. p̂ ± 2 standard errors
If the proportion is close to either 1 or 0 most individuals pˆ (1 − pˆ )
have the same trait or opinion, so there The standard error is s.e.( pˆ ) =
n
is little natural variability and the margin of error
Interpretation: For about 95% of all randomly selected
is smaller than if the proportion is near 0.5.
samples from the population, the confidence interval
3. The “multiplier” 2 or 1.96. computed in this manner captures the population proportion.
Connected to the “95%” aspect of the margin of error. Necessary Conditions: nˆp and n(1− pˆ ) are both greater
Usually the term “margin of error” is used only when the than 10, and the sample is randomly selected.
confidence level is 95%.

Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 17 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 18
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010
Finding the formula for a 95% CI for Same holds for any confidence level;
a Proportion – use Empirical Rule: replace 2 with z*
For 95% of all samples, p̂ is within 2 st.dev. of p pˆ (1 − pˆ )
pˆ ± z ∗
Sampling distribution of p̂ tells us for 95% of all samples: n
where:
−2 standard deviations < p̂ − p < 2 standard deviations
Don’t know true standard deviation, so use standard error. • pˆ is the sample proportion
For approximately 95% of all samples,
−2 standard errors < p̂ − p < 2 standard errors • z* denotes the multiplier.
which implies for approximately 95% of all samples,
pˆ (1 − pˆ )
p̂ – 2 standard errors < p < p̂ + 2 standard errors •. is the standard error of pˆ .
n
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 19 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 20
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

Example 10.3 Intelligent Life Elsewhere? Confidence intervals and “plausible”

values
Poll: Random sample of 935 Americans
Do you think there is intelligent life on other planets? • Remember that a confidence interval is an
Results: 60% of the sample said “yes”, p̂ = .60 interval estimate for a population parameter.
.6(1 − .6 ) • Therefore, any value that is covered by the
s.e.( pˆ ) = = .016
935 confidence interval is a plausible value for
90% Confidence Interval: .60 ± 1.65(.016), or .60 ± .026 the parameter.
.574 to .626 or 57.4% to 62.6% • Values not covered by the interval are still
98% Confidence Interval: .60 ± 2.33(.016), or .60 ± .037
possible, but not very likely (depending on
.563 to .637 or 56.3% to 63.7%
the confidence level).
Note: entire interval is above 50% => high confidence
that a majority believe there is intelligent life.
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 21 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 22
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

Example of plausible values New multiplier: let’s do a confidence level of 50%

Poll: Random sample of 935 Americans
• 98% Confidence interval for proportion who
“Do you think there is intelligent life on other planets?”
believe intelligent life exists elsewhere is:
Results: 60% of the sample said “yes”, p̂ = .60
.563 to .637 or 56.3% to 63.7%
• Therefore, 56% is a plausible value for the We want a 50% confidence interval.
If the area between -z* and z* is .50,
population percent, but 50% is not very then the area to the left of z* is .75.
likely to be the population percent. From Table A.1 we have z* ≈ .67.
• Entire interval is above 50% => high (See next page for Table A.1)
confidence that a majority believe there is 50% Confidence Interval: .60 ± .67(.016), or .60 ± .011
.589 to .611 or 58.9% to 61.1%
intelligent life.
Note: Lower confidence level results in a narrower interval.
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 23 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 24
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010
Remember conditions for using the formula: In Summary: Confidence Interval
for a Population Proportion p
1. Sample is randomly selected from the population.
Note: Available data can be used to make inferences pˆ (1 − pˆ )
about a much larger group if the data can be General CI for p: pˆ ± z ∗
n
considered to be representative with regard to
the question(s) of interest.
Approximate pˆ (1 − pˆ )
2. Normal curve approximation to the distribution pˆ ± 2
of possible sample proportions assumes a 95% CI for p: n
“large” sample size. Both nˆp and n(1− pˆ )
should be at least 10 (although some say these
Conservative 1
need only to be at least 5). pˆ ±
95% CI for p: n
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 25 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 26
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

Section 10.4: Comparing two Case Study 10.3 Comparing proportions

Would you date someone with a great personality
population proportions even though you did not find them attractive?
Women: .611 (61.1%) of 131 answered “yes.”
• Independent samples of size n1 and n2 95% confidence interval is .527 to .694.
• Use the two sample proportions as data. Men: .426 (42.6%) of 61 answered “yes.”
95% confidence interval is .302 to .55.
• Could compute separate confidence intervals Conclusions:
for the two population proportions and see if • Higher proportion of
they overlap. women would say yes.
• Better to find a confidence interval for the CIs slightly overlap
• Women CI narrower
difference in the two population proportions, than men CI due to
larger sample size
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 27 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 28
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

Compare the two proportions by Case Study 10.3 Comparing proportions

Would you date someone with a great personality
finding a CI for the difference even though you did not find them attractive?
Women: .611 of 131 answered “yes.”
C.I. for the difference in two population
95% confidence interval is .527 to .694.
proportions:
Men: .426 of 61 answered “yes.”
Sample estimate ± multiplier × standard error 95% confidence interval is .302 to .55.
Confidence interval for the difference in population
pˆ 1 (1 − pˆ 1 ) pˆ 2 (1 − pˆ 2 ) proportions of women and men who would say yes.
( pˆ 1 − pˆ 2 ) ± z * +
n1 n2 .611(1 − .611) .426(1 − .426)
(.611 − .426) ± z * +
131 61

Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 29 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 30
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010
 95% confidence interval Section 10.5: Using confidence
intervals to guide decisions
• A 95% confidence interval for the difference is
.035 to .334 or 3.5% to 33.4%. • A value not in a confidence interval can be rejected
• We are 95% confident that the population as a likely value for the population parameter.
proportions of men and women who would date • When a confidence interval for p1 − p2 does not cover
someone they didn’t find attractive differ by 0 it is reasonable to conclude that the two population
between .035 and .334, with a lower proportion values differ.
for men than for women. • When confidence intervals for p1 and p2 do not
• We can conclude that the two population overlap it is reasonable to conclude they differ, but if
proportions differ because 0 is not in the interval. they do overlap, no conclusion can be made. In that
case, find a confidence interval for the difference.
Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 31 Copyright ©2004 Brooks/Cole, a division of Thomson Learning, Inc., updated 32
by Jessica Utts Feb 2010 by Jessica Utts Feb 2010

From the Midterm 2 review sheet

for Chapter 10 - you should know these now
1. Understand how to interpret the confidence level
2. Understand how to interpret a confidence interval
3. Understand how the sampling distribution for p̂ leads to
the confidence interval formula (pg. 417-418)
4. Know how to compute a confidence interval for one
proportion, including conditions needed.
5. Know how to compute a confidence interval for the
difference in two proportions, including conditions needed.
6. Understand how to find the multiplier for desired
confidence level.
7. Understand how margin of error from Chapter 3 relates to
the 95% confidence interval formula in Chapter 10
8. Know the general format for a confidence interval for the 5
situations defined in Chapter 9 (see summary on page 483).
33