# SOCR EduMaterials AnalysisActivities TwoPairedT

(Difference between revisions)
Jump to: navigation, search
 Revision as of 20:11, 4 August 2007 (view source)Annie (Talk | contribs) (→Two Paired Sample T-Test Example)← Older edit Current revision as of 23:35, 9 August 2009 (view source)IvoDinov (Talk | contribs) (7 intermediate revisions not shown) Line 1: Line 1: - ==Two Paired Sample T-Test Example== + ==[[SOCR_EduMaterials_AnalysesActivities | SOCR Analysis]] - Two Paired Sample T-Test Example== - We will demonstrate Two Paired T-Test with some SOCR built-in examples.  This example is based on the data taken from the statistical program "R." For more information of the R program, please see [http://cran.r-project.org/ CRAN Home Page]. The dataset used here is "shoes" under R's "MASS" library. In the dataset, ten boys are given to the wear of shoes of materials A and B for one foot. We'd like to find if one material is better than the other. + + This is a test for the hypothesis that two groups are with the same mean, while the data are actually paired. That is, Each value of group 1 has a corresponding value in group 2. + + ===Two Paired Sample T-Test Example=== + + Go to the [http://www.socr.ucla.edu/htmls/ana/TwoPairedSampleTTest_Analysis.html SOCR Two Paired Sample T-Test Applet]. We will demonstrate Two Paired T-Test with a SOCR built-in example.  This example is based on a dataset from the statistical program "R." For more information of the R program, please see [http://cran.r-project.org/ CRAN Home Page]. The dataset used here is "'''shoes'''" under R's "'''MASS"''' library. In the dataset, ten boys are given to the wear of shoes of materials A and B for one foot. We'd like to find if one material is better than the other. Line 17: Line 22: - '''2.''' Click on the "'''Mapping'''" button to get to the "'''Mapping'''" panel. Click on '''"ADD"''' under '''VARIABLE 1''' and to '''VARIABLE 2''' add the variables for Two Paired Sample T-Test. We care about the difference, so which on goes to '''VARIABLE 1''' is arbitrary. You can just add column '''A''' to '''VARIABLE 1''' and '''B''' to '''VARIABLE 2'''. + '''2.''' Click on the "'''Mapping'''" button to get to the "'''Mapping'''" panel. Click on '''"ADD"''' under '''VARIABLE 1''' to add '''A''' and click on '''ADD''' under '''VARIABLE 2''' to add '''B'''. We care about the difference, so which on goes to '''VARIABLE 1''' or '''VARIABLE 2''' is arbitrary. Here, the computer will set difference = Variable 2 - Variable 1. If you like the subtraction to go the other way around, simple assign '''VARIABLE 1''' and '''VARIABLE 2''' in the opposite way. - Here, the computer will set difference = Variable 2 - Variable 1. If you like the substraction to go the other way around, simple assign '''VARIABLE 1''' and '''VARIABLE 2''' in a different way. + Line 31: Line 35: '''Questions for students:''' By looking at the test results, what do you think about the data? What can you say about the two different shoe materials? '''Questions for students:''' By looking at the test results, what do you think about the data? What can you say about the two different shoe materials? + + ===Example 2=== + In many experiments the same variable is measured under two different conditions.  For example, in clinical trials the participants may be evaluated at baseline and then evaluated again at the end of the treatment.  For example, the blood pressure is measured for several patients before and after administration of certain drug.  The difference between the value at baseline and the value at the end is computed for each participant as follows: + +
+ {| class="wikitable" + |- + ! || Value at baseline || Value at the end  || Difference + |- + | Subject 1  || $x_{1b}$  ||$x_{1a}$  || $d_1=x_{1b}-x_{1a}$ + |- + | Subject 2 || $x_{2b}$  ||$x_{2a}$    ||$d_2=x_{2b}-x_{2a}$ + |- + | Subject 3  || $x_{3b}$  ||$x_{3a}$  || $d_3=x_{3b}-x_{3a}$ + |- + | ... || ...|| ... || ... + |- + | ... || ...|| ... || ... + |- + | Subject ''n'' || $x_{nb}$ ||$x_{na}$ || $d_n=x_{nb}-x_{na}$ + |}
+ + We then compute the sample mean and sample standard deviation of the differences. + : $\bar d=\frac{\sum_{i=1}^n d_i}{n}, \ \ \ \ s_d^2=\frac{\sum_{i=n}^n(d_i-\bar d)^2}{n-1}$ + + The hypothesis we want to test is:  $H_0: \mu_d=d_0$ vs. $H_a: \mu_dd_0$ or $\mu_d \ne d_0$. + + If we choose $d_0=0$ we are testing whether the before and after treatment are the same. + + * Test statistic: + : $t=\frac{\bar d - d_0}{\frac{s_d}{\sqrt{n}}}$ + + * Assumption: + : The differences are treated as a random sample from a normal distribution. + + We reject $H_0$ if the $t$ value falls in the rejection region which is based on the significance level $\alpha$ and $n-1$ degrees of freedom. + + To study the effect of cigarette smoking on platelet aggregation researchers drew blood samples from 11 individuals before and after they smoked a cigarette and measured the percentage of blood platelet  aggregation (from ''Mathematical Statistics and Data Analysis'', John Rice, Third Edition, Duxbury, 2007).  Platelets are involved in the formation of blood clots, and it is known that smokers suffer more from disorders involving blood clots than do nonsmokers.  This study can be found in [http://circ.ahajournals.org/cgi/reprint/48/3/619.pdf Levine, P. H. (1973). An acute effect of cigarette smoking on platelet function, Circulation, 48, 619-623]. + +
+ {| class="wikitable" + |- + ! Before    || After  || Difference + |- + | 25 || 27 || 2 + |- + | 25 || 29 || 4 + |- + | 27  || 37  || 10 + |- + | 44  || 56 || 12 + |- + | 30  || 46 || 16 + |- + | 67  || 82 || 15 + |- + | 53  || 57 || 4 + |- + | 53  || 80 || 27 + |- + | 52  || 61 || 9 + |- + | 60  || 59 || -1 + |- + | 28  || 43 || 15 + |}
+ + * Test the null hypothesis that the means before and after are the same.  Use $\alpha=0.05$. + + ===References=== + * [http://circ.ahajournals.org/cgi/reprint/48/3/619.pdf Levine, P. H. (1973). An acute effect of cigarette smoking on platelet function, Circulation, 48, 619-623]. + +
+ + {{translate|pageName=http://wiki.stat.ucla.edu/socr/index.php?title=SOCR_EduMaterials_AnalysisActivities_TwoPairedT}}

## SOCR Analysis - Two Paired Sample T-Test Example

This is a test for the hypothesis that two groups are with the same mean, while the data are actually paired. That is, Each value of group 1 has a corresponding value in group 2.

### Two Paired Sample T-Test Example

Go to the SOCR Two Paired Sample T-Test Applet. We will demonstrate Two Paired T-Test with a SOCR built-in example. This example is based on a dataset from the statistical program "R." For more information of the R program, please see CRAN Home Page. The dataset used here is "shoes" under R's "MASS" library. In the dataset, ten boys are given to the wear of shoes of materials A and B for one foot. We'd like to find if one material is better than the other.

Here's the steps of the activity:

1. Click on Two Paired Sample T-Test at the left panel's combo box.

On the right panel, first click on EXAMPLE 3 and next click on "Data" to retreive the demonstrated example. This is the example we'll be looking at.Note that the data are difference of B - A.

2. Click on the "Mapping" button to get to the "Mapping" panel. Click on "ADD" under VARIABLE 1 to add A and click on ADD under VARIABLE 2 to add B. We care about the difference, so which on goes to VARIABLE 1 or VARIABLE 2 is arbitrary. Here, the computer will set difference = Variable 2 - Variable 1. If you like the subtraction to go the other way around, simple assign VARIABLE 1 and VARIABLE 2 in the opposite way.

3. Click on "Calculate" then "Result" to see the results.

Questions for students: By looking at the test results, what do you think about the data? What can you say about the two different shoe materials?

### Example 2

In many experiments the same variable is measured under two different conditions. For example, in clinical trials the participants may be evaluated at baseline and then evaluated again at the end of the treatment. For example, the blood pressure is measured for several patients before and after administration of certain drug. The difference between the value at baseline and the value at the end is computed for each participant as follows:

Value at baseline Value at the end Difference
Subject 1 x1b x1a d1 = x1bx1a
Subject 2 x2b x2a d2 = x2bx2a
Subject 3 x3b x3a d3 = x3bx3a
... ... ... ...
... ... ... ...
Subject n xnb xna dn = xnbxna

We then compute the sample mean and sample standard deviation of the differences.

$\bar d=\frac{\sum_{i=1}^n d_i}{n}, \ \ \ \ s_d^2=\frac{\sum_{i=n}^n(d_i-\bar d)^2}{n-1}$

The hypothesis we want to test is: H0d = d0 vs. Had < d0 or μd > d0 or $\mu_d \ne d_0$.

If we choose d0 = 0 we are testing whether the before and after treatment are the same.

• Test statistic:
$t=\frac{\bar d - d_0}{\frac{s_d}{\sqrt{n}}}$
• Assumption:
The differences are treated as a random sample from a normal distribution.

We reject H0 if the t value falls in the rejection region which is based on the significance level α and n − 1 degrees of freedom.

To study the effect of cigarette smoking on platelet aggregation researchers drew blood samples from 11 individuals before and after they smoked a cigarette and measured the percentage of blood platelet aggregation (from Mathematical Statistics and Data Analysis, John Rice, Third Edition, Duxbury, 2007). Platelets are involved in the formation of blood clots, and it is known that smokers suffer more from disorders involving blood clots than do nonsmokers. This study can be found in Levine, P. H. (1973). An acute effect of cigarette smoking on platelet function, Circulation, 48, 619-623.

Before After Difference
25 27 2
25 29 4
27 37 10
44 56 12
30 46 16
67 82 15
53 57 4
53 80 27
52 61 9
60 59 -1
28 43 15
• Test the null hypothesis that the means before and after are the same. Use α = 0.05.

### References

Translate this page: