Point Estimate vs Interval

Interval Estimate Estimate

 An interval estimate is an interval

of numbers within which the
parameter value is believed to fall.

Point Estimation: How Do We Make

a Best Guess for a Population
Parameter? Confidence Interval
A confidence interval is an interval
containing the most believable values
 Use an appropriate sample statistic: for a parameter
• For the population mean, use the sample  The probability that this method
mean
produces an interval that contains the
• For the population proportion, use the parameter is called the confidence
sample proportion level
• This is a number chosen to be close to 1,
most commonly 0.95

A 95% Confidence Interval for a

Population Proportion Margin of Error
 The margin of error measures how
 Fact: Approximately 95% of a normal
distribution falls within 1.96 standard accurate the point estimate is likely to
deviations of the mean be in estimating a parameter
• That means: With probability 0.95, the  The distance of 1.96 standard errors
sample proportion falls within about 1.96 in the margin of error for a 95%
standard errors of the population confidence interval
proportion

Finding the 95% Confidence Interval

Confidence Interval for a Population Proportion

A confidence interval is constructed  A 95% confidence interval uses a margin of

by adding and subtracting a margin of error = 1.96(standard errors)
error from a given point estimate  [point estimate ± margin of error] =
 When the sampling distribution is
approximately normal, a 95%
p̂ ± 1.96(standard errors)
confidence interval has margin of
error equal to 1.96 standard errors
Finding the 95% Confidence Interval Finding the 95% Confidence Interval
for a Population Proportion for a Population Proportion
 The exact standard error of a sample proportion
equals:  In practice, we use an estimated standard
p(1 − p) error:
n
ˆ (1 − p
p ˆ)
se =
 This formula depends on the unknown population n
proportion, p
 In practice, we don’t know p, and we need to
estimate the standard error

Finding the 95% Confidence Interval Example: Would You Pay Higher
for a Population Proportion Prices to Protect the Environment?

 A 95% confidence interval for a population  In 2000, the GSS asked: “Are you
proportion p is: willing to pay much higher prices in
order to protect the environment?”
p̂(1 - p̂)
p̂ ± 1.96(se), with se = • Of n = 1154 respondents, 518 were
n willing to do so

Example: Would You Pay Higher Example: Would You Pay Higher
Prices to Protect the Environment? Prices to Protect the Environment?

 Find and interpret a 95% confidence 518

p̂ = = 0.45
interval for the population proportion 1154
of adult Americans willing to do so at (0.45)(0.55)
the time of the survey se = = 0.015
1154
p̂ ± 1.96(se) = 1.96(0.015)
= 0.45 ± 0.03 = (0.42, 0.48)

Sample Size Needed for Large-Sample

Confidence Interval for a Proportion “95% Confidence”
 With probability 0.95, a sample
 For the 95% confidence interval for a
proportion p to be valid, you should have at
proportion value occurs such that the
least 15 successes and 15 failures: confidence interval contains the
population proportion, p
npˆ ≥ 15 and n(1 - p̂) ≥ 15  With probability 0.05, the method
produces a confidence interval that
misses p

What is the Error Probability for

Different Confidence Levels the Confidence Interval Method?
 In using confidence intervals, we
must compromise between the
desired margin of error and the
desired confidence of a correct
inference
• As the desired confidence level
increases, the margin of error gets
larger
 How to Construct a Confidence
Interval for a Population Mean
 Point estimate ± margin of error
How Can We Construct a  The sample mean is the point
Confidence Interval To Estimate a estimate of the population mean
Population Mean?  The exact standard error of the
sample mean is σ/ n
 In practice, we estimate σ by the
sample standard deviation, s

How to Construct a Confidence How to Construct a Confidence

Interval for a Population Mean Interval for a Population Mean
 For large n…  In practice, we don’t know the
• and also population standard deviation
 For small n from an underlying population  Substituting the sample standard
that is normal… deviation s for σ to get se = s/ n
 The confidence interval for the population introduces extra error
mean is:  To account for this increased error,
σ we replace the z-score by a slightly
x ± z( ) larger score, the t-score
n

How to Construct a Confidence

Interval for a Population Mean Properties of the t-distribution
 The t-distribution is bell shaped and
 In practice, we estimate the standard symmetric about 0
error of the sample mean by se = s/ n  The probabilities depend on the

 Then, we multiply se by a t-score from degrees of freedom, df

the t-distribution to get the margin of  The t-distribution has thicker tails and
error for a confidence interval for the is more spread out than the standard
population mean normal distribution

Summary: 95% Confidence Example: eBay Auctions of

Interval for a Population Mean Palm Handheld Computers
 A 95% confidence interval for the
population mean µ is:  Doyou tend to get a higher, or a
s lower, price if you give bidders the
x±t ( ); df = n - 1
.025
n “buy-it-now” option?

 To use this method, you need:

• Data obtained by randomization
• An approximately normal population distribution

Example: eBay Auctions of Example: eBay Auctions of

Palm Handheld Computers Palm Handheld Computers
 Consider some data from sales of the  “Buy-it-now”
option:
Palm M515 PDA (personal digital 235 225 225 240 250 250 210
assistant)
 Bidding
only:
 During the first week of May 2003, 25
of these handheld computers were 250 249 255 200 199 240 228
auctioned off, 7 of which had the 255 232 246 210 178 246 240
“buy-it-now” option 245 225 246 225
Example: eBay Auctions of Example: eBay Auctions of
Palm Handheld Computers Palm Handheld Computers
 Summary of selling prices for the two
types of auctions:  To construct a confidence interval
using the t-distribution, we must
buy_now N Mean StDev Minimum Q1 Median Q3
no 18 231.61 21.94 178.00 221.25 240.00 246.75
assume a random sample from an
yes 7 233.57 14.64 210.00 225.00 235.00 250.00 approximately normal population of
buy_now Maximum
selling prices
no 255.00
yes 250.00

Example: eBay Auctions of Example: eBay Auctions of

Palm Handheld Computers Palm Handheld Computers
 Let µ denote the population mean for
the “buy-it-now” option  The 95% confidence interval for the “buy-it-
now” option is:
 The estimate of µ is the sample mean:

x = $233.57 s 14.64
x ± t.025 ( ) = 233.57 ± 2.44( )
 The sample standard deviation is: n 7
s = $14.64  which is 233.57 ± 13.54 or (220.03, 247. 11)

Example: eBay Auctions of Example: eBay Auctions of

Palm Handheld Computers Palm Handheld Computers
 Notice that the two intervals overlap
 The95% confidence interval for the a great deal:
mean sales price for the bidding only • “Buy-it-now”: (220.03, 247.11)
option is: • Bidding only: (220.70, 242.52)
(220.70, 242.52)  There is not enough information for us to
conclude that one probability distribution
clearly has a higher mean than the other

How Do We Find a t- Confidence

Interval for Other Confidence
Levels?

 The95% confidence interval uses t.025

since 95% of the probability falls
between - t.025 and t.025

 For 99% confidence, the error

probability is 0.01 with 0.005 in each
tail and the appropriate t-score is t.005