Scribd
Upload
64 views4 pages

2 Mean Hypothesis Tests With Sigma Unknown (2009)

This document provides a template for conducting a hypothesis test to compare the means of two populations. It guides the user through entering sample data or summary statistics, specifying the null hypothesis and significance level, and reports the results of the test including test statistics, p-values, confidence intervals, and conclusions.

Uploaded by

Tin Nguyen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as XLS, PDF, TXT or read online on Scribd
64 views4 pages

2 Mean Hypothesis Tests With Sigma Unknown (2009)

This document provides a template for conducting a hypothesis test to compare the means of two populations. It guides the user through entering sample data or summary statistics, specifying the null hypothesis and significance level, and reports the results of the test including test statistics, p-values, confidence intervals, and conclusions.

Uploaded by

Tin Nguyen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as XLS, PDF, TXT or read online on Scribd
You are on page 1/ 4

Sample Sample

Hypothesis test: Difference of two means with sigma(s) unknown


#1 #2
54.5 51 hypothesized mean, difference, m1-2 m2: 0 ← This term is added to the µ2 side of the null
67.8 62 level of significance, a: 0.05
64.5 54 null hypothesis µ1 = µ2
70.4 58 equal variances assumed for two pops? no Population variances NOT assumed equal.
72.5 47 Since you provided data for Delete Sample 1 raw data
64.9 56 ← sample 1, I'll use it to ← if you want to enter
83.3 67 compute x1-bar, s1, and n1. summary data instead.
72.8 81 Since you provided data for Delete Sample 1 raw data
68.8 67 ← sample 2, I'll use it to ← if you want to enter
62 compute x2-bar, s2, and n2. summary data instead.
mean of sample 1, x1-bar: 68.83333333333 = AVERAGE(range1)
standard deviation of sample 1, s1: 7.759510293827 = STDEV(range1)
size of sample 1, n1: 9 = COUNT(range1)
mean of sample 2, x2-bar: 60.5 = AVERAGE(range2)
standard deviation of sample 2, s2: 9.721111047612 = STDEV(range2)
size of sample 2, n2: 10 = COUNT(range2)
difference in sample means, x1-bar - x2-bar: 8.333333333333 =x1-bar - x2-bar
estimated pooled standard deviation, sp N/A
standard error (unpooled): 4.01746188532 =sp * SQRT(1/n1 + 1/n2)
deg. of freedom for samp. dist., df: 16 See book equation 10.5
t*, critical t value for nonrejection: 2.119905299221 =TINV(2a/# of tails, df)
margin of error, MOE: 8.51663874011 =std err x crit t
lower limit for nonrejection region*: -8.51663874011 =mu1 - mu2 - MOE*
upper limit for nonrejection region*: 8.51663874011 =mu1 -mu2 + MOE*
lower limit of 1-a confidence interval: -0.18330540678 =x1-bar - x2-bar - MOE
upper limit of 1-a confidence interval: 16.84997207344 =x1-bar - x2-bar + MOE
tsam: t score of difference of sample means: 2.074278131619 =((x1-bar -x2-bar) - (mu1 - mu2))/std err
P value of difference of sample means: 0.054550320891 =TDIST(ABS(sample t),df,# tails)
Conclusion: Do not reject null hypothesis
Check
Normal approximation holds if the two sampled populations are similar and not too assymmetric

Template copyright 2009 by Scott P Stevens


Hypothesis Test for the Difference of Two Population Means
Template by Scott Stevens

Instructions
1. Enter the data.
a. If you have the observations for sample 1, enter them in the yellow cells in Column A of the spreadsheet.
If you do not have the sample 1 observations, leave the first column blank. Enter
sample 1's mean, sample size, and sample standard deviation in the appropriate yellow boxes.
b. If you have the observations for sample 2, enter them in the yellow cells in Column B of the spreadsheet.
If you do not have the sample 2 observations, leave the second column blank. Enter
sample 2's mean, sample size, and sample standard deviation in the appropriate yellow boxes.

2. a. In the appropriate yellow box, enter the hypothesized difference (usually 0) between the means of populations 1 and 2.
b. Enter the desired level of significance. Commonly used values are 0.01, 0.05, and 0.1.
(The smaller the number used, the stronger the implication that follows from a rejected null hypothesis.)
c. Choose from the pulldown menu whether you wish a <, = or > symbol in your null hypothesis.

3. Use the pulldown menu to indicate whether you wish to assume that the two populations have the same variance.
Unless you have reason to believe that they do, it's generally better NOT to assume that the variances are equal.
(Note: This spreadsheet does not currently do hypothesis tests including a finite population multiplier.)
(Note: This spreadsheet does not currently do hypothesis tests in which population variance is known. In practice, it
is almost always estimated from the sample.)

Results
1. The template goes through the work of conducting the hypothesis test. If equal population variances are assumed,
a pooled estimate of population variance is used. Formulas used in the calculation are shown in the blue boxes.
Words in CAP ITALICS (except for MOE) are entered into Excel exactly as typed. Lower case words refer to quantities
already computed. The formula for the degrees of freedom when population variances are not assumed equal is too
complicated to appear on the sheet, but it can be found as equation 10.5 in the text.

2. The template reports:


a. The lower and upper limits of the nonrejection region.
b. The t value and p-value for the observed difference of sample means. The p-value is 1- or 2-tailed, as appropriate to th
c. The critical t value, t*
d. The conclusion drawn from the hypothesis test.
e. The (two tailed) confidence interval for the difference of population means at the 1 - α level of confidence.

3. The template evaluates the reasonableness of the assumption that the underlying sampling distribution is normal.

Note: Column H is hidden on this spreadsheet. Some calculations necessary for Excel are done in this column.
s of populations 1 and 2.

e same variance.
nces are equal.

nown. In practice, it

ces are assumed,


he blue boxes.
ords refer to quantities
sumed equal is too

ailed, as appropriate to the problem.

bution is normal.

You might also like