2
$\begingroup$

I am trying to do a few tests in Minitab and I don't understand why they are not working out.

First, my data in minitab has several columns: Class Name Gender Height Age, where

  • Class has two values {A,B};
  • Name is the name of the students;
  • Gender is Female or Not female;
  • Age is the age of the student.

My questions I have to answer are as follows: Is the height of Class A larger than Class B?

My understanding of the question I should be a 2 sample t-test of the differences of the Height of Class A - Class B

However, every time I try to do a two sample t-test it just compares class with height, which is silly and doesn't workout.

How Do I get Minitab to test Class A's height against Class B's height?

Ps. If I am incorrect in the correct hypothesis testing method, please let me know.

$\endgroup$
  • $\begingroup$ Many statistical software allow to perform a two-sample t-test where data is organized in two columns with either (a) scores of group 1 in 1st column, scores of group 2 in 2nd column, or (b) all scores in one column and the classification factor in the other. This should help: ams.sunysb.edu/~kye/ams102/minitab/twottest/twottest.html. $\endgroup$ – chl Dec 9 '12 at 11:24
  • $\begingroup$ if you write that up as an answer, I will accept and upvote. Thanks for the hyperlink. $\endgroup$ – MaoYiyi Dec 10 '12 at 0:03
3
$\begingroup$

I don't have Minitab so I can't show you the exact syntax or GUI menu actions to perform a t-test for independent samples. However, statistical packages (e.g., R, Stata) generally provide commands to perform a t-test when data are organized as either

  1. two series of measurements, one for each group (usually arranged in two columns);
  2. all individual measurements in one column, described by group levels in another column.

With the example from Wikipedia ("weights of screws that were chosen out of a bucket"), that would mean that scheme 1 looks like:

+-----+-----+
|A1   |A2   |
+-----+-----+
|30.02|29.89|
+-----+-----+
|29.99|29.93|
+-----+-----+
|30.11|29.72|
+-----+-----+
|29.97|29.98|
+-----+-----+
|30.01|30.02|
+-----+-----+
|29.99|29.98|
+-----+-----+

while for scheme 2 we would arrange data as follows:

+-----+-----+
|30.02|A1   |
+-----+-----+
|29.99|A1   |
+-----+-----+
|30.11|A1   |
+-----+-----+
|29.97|A1   |
+-----+-----+
|30.01|A1   |
+-----+-----+
|29.99|A1   |
+-----+-----+
|29.89|A2   |
+-----+-----+
|29.93|A2   |
+-----+-----+
|29.72|A2   |
+-----+-----+
|29.98|A2   |
+-----+-----+
|30.02|A2   |
+-----+-----+
|29.98|A2   |
+-----+-----+

As explained on this short tutorial on Minitab 2-Sample-t-test, any scheme can be used by filling out the correct option in the 2-Sample-t dialog box. In your case, that means selecting Height as Samples and Class as Subscripts, with Samples arranged in one column. After running your t-test, you may want to check usual assumptions (normality and homogeneity of variance).

$\endgroup$
2
$\begingroup$

Two series of measurements can be given in two separate columns as well in Minitab.

Considering the example in Wiki

  1. Enter two samples in two columns
  2. Go to Stat Menu -> Basic statistics -> 2-sample-t-test

And select "samples in different columns".

You can choose "sample is one column" if you have arranged measurements in one column.

enter image description here

Click "OK".

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.