Scribd
Upload
6 views44 pages

Ch10 note

The document discusses statistical inference techniques for small samples, including methods for estimating population means, comparing two means, and analyzing population variances. It covers the use of distributions such as Student's t, chi-square, and F distributions, emphasizing the importance of sample independence and normality. Additionally, it highlights the paired-t test and its application in experimental designs, as well as the implications of pairing on degrees of freedom and variability.

Uploaded by

thunderstorm.mg12
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views44 pages

Ch10 note

The document discusses statistical inference techniques for small samples, including methods for estimating population means, comparing two means, and analyzing population variances. It covers the use of distributions such as Student's t, chi-square, and F distributions, emphasizing the importance of sample independence and normality. Additionally, it highlights the paired-t test and its application in experimental designs, as well as the implications of pairing on degrees of freedom and variability.

Uploaded by

thunderstorm.mg12
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 44

Inference from Small Samples

• The basic concepts of large-sample statistical estimation and


hypothesis testing for population means and proportions were
introduced in Chapters 8 and 9. Because all of these techniques rely on
the Central Limit Theorem to justify the normality of the estimators
and test statistics, they apply only when the samples are large.
Inference from Small Samples
• Small-sample inferences concerning a population mean (10.2)
• Small-sample inferences concerning the difference in two means:
Independent random samples (10.3)
• Paired-difference test: Dependent samples (10.4)
• Inferences concerning a population variance (10.5)
• Comparing two population variances (10.6)
• Student’s t distribution (10.1)
• Chi-square distribution (10.5)
• F distribution (10.6)
Chi-square Distribution
• Let 𝑍 be a random variable following the standard normal distribution
1 −𝑧 2
𝑓𝑍 𝑧 = 𝑒 , 𝑧 ∈ (−∞, ∞).
2𝜋
Then 𝑍 2 follows a chi-square distribution with one degree of freedom.
We use 𝜒12 to denote the chi-square distribution with one degree of
freedom.
• Let 𝑋1 , … , 𝑋𝑟 𝑖.𝑖.𝑑.
~
𝜒1
2
. Then σ 𝑘
𝑋
𝑖=1 𝑖 ~𝜒 2
𝑘 , where 𝜒 2
𝑘 denote the chi-
square distribution with 𝑘 degrees of freedom
1
𝑓𝜒2 𝑥 = 𝑘/2 𝑥 𝑘/2−1 𝑒 −𝑥/2 , 𝑥 > 0.
𝑘 2 Γ 𝑘/2
Chi-square Distribution
Chi-square Distribution
2
• 𝑃 𝜒𝑑𝑓 > 𝜒𝑎2 = 𝑎
Student’s t Distribution
𝑍
• Let Z~𝑁(0,1) and 𝑋~𝜒𝑣2 be independent. Then ~𝑡𝑣 where 𝑡𝑣 is a
𝑋/𝑣
Student’s t distribution with degrees of freedom v.
Student’s t Distribution
• 𝑃 𝑡𝑑𝑓 > 𝑡𝑎 = 𝑎
Student’s t Distribution
F distribution
𝑋/𝑑1
• Let𝑋~𝜒𝑑21 𝑌~𝜒𝑑22
and be independent. Then ~𝐹𝑑1 ,𝑑2 where
𝑌/𝑑2
𝐹𝑑1 ,𝑑2 is an F distribution with degrees of freedom 𝑑1 and 𝑑2 .
F distribution
• 𝑃 𝐹𝑑𝑓1 ,𝑑𝑓2 > 𝐹𝑎 = 𝑎
F distribution
Small-Sample Inferences Concerning
a Population Mean
Small-Sample Inferences Concerning
a Population Mean
Small-Sample Inferences Concerning
a Population Mean
Small-Sample Inferences Concerning
a Population Mean
Small-Sample Inferences Concerning
a Population Mean
Small-Sample Inferences Concerning
a Population Mean
Small-Sample Inferences Concerning
a Population Mean
Small-Sample Inferences Concerning
a Population Mean
Difference Between Two
Population Means: Independent
Random Samples
Small-Sample Inferences for the Difference Between
Two Population Means: Independent Random Samples
Small-Sample Inferences for the Difference Between
Two Population Means: Independent Random Samples
Small-Sample Inferences for the Difference Between
Two Population Means: Independent Random Samples
Small-Sample Inferences for the Difference Between
Two Population Means: Independent Random Samples
Small-Sample Inferences for the Difference Between
Two Population Means: Independent Random Samples
• The samples must be randomly selected. Samples not randomly
selected may introduce bias into the experiment and thus alter the
significance levels you are reporting.
• The samples must be independent. If not, this is not the appropriate
statistical procedure.
• The populations from which you sample must be normal. However,
moderate departures from normality do not seriously affect the
distribution of the test statistic, especially if the sample sizes are nearly
the same.
• The population variances should be equal or nearly equal to ensure
that the procedures are valid.
Small-Sample Inferences for the Difference Between
Two Population Means: Independent Random Samples
• When 𝜎12 ≠ 𝜎22 (Welch's t-test)

𝑥1ҧ − 𝑥ҧ2 − 𝐷0
𝑡0 =
𝑠12 𝑠22
+
𝑛1 𝑛2
• 𝑡0 is not distributed exactly as t-dist. However, the distribution of 𝑡0 is well
approximated by 𝑡𝜈 if we use
2 2 2
𝑠1 𝑠2
+
𝑛1 𝑛2
𝑣= 2
𝑠1 /𝑛1 2 𝑠22 /𝑛2 2
+
𝑛2 − 1 𝑛2 − 1
• Welch, B. L. (1947), "The generalization of "student's" problem when several
different population variances are involved.", Biometrika, 34 (1-2): 28–35 26
Small-Sample Inferences for the Difference Between
Two Population Means: Independent Random Samples
Small-Sample Inferences for the Difference
Between Two Means: A Paired-t Test
• To compare the wearing qualities of two
types of automobile tires, A and B, a tire
of type A and one of type B are randomly
assigned and mounted on the rear wheels
of each of five automobiles. The
automobiles are then operated for a
specified number of miles, and the
amount of wear is recorded for each tire.
A Paired-t Test
Small-Sample Inferences for the Difference
Between Two Means: A Paired-t Test
Small-Sample Inferences for the Difference
Between Two Means: A Paired-t Test
Small-Sample Inferences for the Difference
Between Two Means: A Paired-t Test
Small-Sample Inferences for the Difference
Between Two Means: A Paired-t Test
• The paired-t test or matched pairs design used in the tire wear
experiment is a simple example of an experimental design called a
randomized block design.
• When there is a great deal of variability among the experimental units,
the effect of this variability can be minimized by blocking—that is,
comparing the different procedures within groups of relatively similar
experimental units called blocks.
Small-Sample Inferences for the Difference
Between Two Means: A Paired-t Test
• It is important for you to remember that the pairing or blocking occurs
when the experiment is planned, and not after the data are collected.
An experimenter may choose to use pairs of identical twins to
compare two learning methods.
• Once you have used a paired design for an experiment, you no longer
have the option of using the unpaired analysis.
Small-Sample Inferences for the Difference
Between Two Means: A Paired-t Test
• Although pairing was very beneficial in the tire wear experiment, this
may not always be the case. In the paired analysis, the degrees of
freedom for the t-test are cut in half—from 2 𝑛 − 1 to (𝑛 − 1).
• This reduction increases the critical value of t for rejecting 𝐻0 and also
increases the width of the confidence interval for the difference in the
two means.
• If pairing is not effective, this increase is not offset by a decrease in
the variability, and you may in fact lose rather than gain information
by pairing.
Inferences Concerning
a Population Variance
Inferences Concerning a Population Variance
Inferences Concerning a Population Variance
Inferences Concerning a Population Variance
Inferences Concerning a Population Variance
Comparing Two
Population Variances
Comparing Two Population Variances
Comparing Two Population Variances
Comparing Two Population Variances

You might also like