Get 1 week of unlimited access
Class Notes (1,031,735)
US (405,217)
STAT (129)
Lecture 7

# STAT 3005 Lecture Notes - Lecture 7: Floor And Ceiling Functions, Confidence Interval, Test StatisticPremium

5 pages103 viewsSpring 2018

Department
Statistics
Course Code
STAT 3005
Professor
H C Tavera
Lecture
7

This preview shows page 1. to view the full 5 pages of the document.
Daniel T. Eisert STAT-3005
1
7.2 Comparing Two Means
Chapter VII: Inference for Means
Comparing Two
Means
Recall ~ the goal given two independent samples with quantitative responses
is to compare two populations where the response is quantitative, examine the
difference between two parameters, and construct confidence intervals or
perform hypothesis tests.
Testing:
Test Statistic: 


Standard Error of :
- First, one should determine whether and are equal or not because
o If they are equal, all of the data can be pooled together to get one
good estimate of the common rather than two less accurate
estimates.
o To do a confidence interval or hypothesis test, the distribution
being dealt with (Z or t, for example), must be known and this
requires knowing whether and are equal or not.
- There are tests for equality of unknown variances, but they are quite
sensitive to departures from normality; therefore, they are NOT robust.
- Generally, consider the variances different if the ratio of larger to
smaller variance is more than 4.




- Case I: Unequal Variance
1. Assume (i.e. 

).
2. Test statistic: 
, where  represents the
degrees of freedom (df).
Two Sample T-Interval
- When the Random, Normal, and Independent conditions are met, a
level C confidence interval for comparing two means is:

where t* is the critical value for confidence level C for the t distribution
with degrees of freedom v.
find more resources at oneclass.com
find more resources at oneclass.com

Unlock to view full version

Subscribers Only

Only half of the first page are available for preview. Some parts have been intentionally blurred.

Subscribers Only
Daniel T. Eisert STAT-3005
2
Comparing Two
Means
Estimating Degrees of Freedom:
- Satterthwaite’s Approximation (Welch-Satterthwaite
Approximation): approximates the degrees of freedom from the t-test
with unequal variance.
- The floor function (L shaped brackets) denotes rounding down to the
next integer.
EXAMPLE: Consider two plant sciences graduate students working on
developing better fertilizers for a particular type of wheat. Joe fertilizes
20 plants and gets heights with  and  while Laura
fertilizes 32 plants and obtains  and . All
measurements are in centimeters.
Conduct a level 0.01 test to see if Laura’s fertilizer is better.
Find a 95% confidence interval for the difference in heights.
- Case II: Equal Variances
1. Assume (i.e. 

).
2. Use the pooled estimate for
2
21
2
22
2
11
2
2
)1()1(
pp
p
ss
nn
snsn
s
=
-+
-+-
=
find more resources at oneclass.com
find more resources at oneclass.com