T Tests

Quick reference

• 00:04 Hi, I'm Ray Sheen.
• 00:05 Now this next set of tests are probably the most commonly use hypothesis tests
• 00:10 in the analyze phase.
• 00:12 The T tests are a simple way for
• 00:14 you to determine if there is a difference between to data sets.
• 00:18 I'll start again with the hypothesis test decision tree.
• 00:23 We are working with normal discrete and continuous data, and two data sets or
• 00:27 samples.
• 00:29 There are two T tests to discuss, the two sample T test, and the paired T test.
• 00:34 I'll start with the most widely used T test, and that is the two sample T test.
• 00:39 This test is used to determine if the means from two samples
• 00:42 are statistically different.
• 00:44 If not, the samples can be combined and treated as one population.
• 00:49 If different, the differences may lead to an understanding of the root cause or
• 00:53 cause of the problems that are being investigated.
• 00:56 It can also be used to demonstrate whether a change has been effective.
• 01:01 So the null hypothesis for
• 01:02 this test is that the mean of one sample equals the mean of the other sample.
• 01:07 And the alternative hypothesis is that they are different.
• 01:10 And as with some of the other tests, you could choose to write the alternative
• 01:13 hypothesis, so that one was greater than or less than the other.
• 01:18 The actual mathematics of this test will change depending upon whether the two
• 01:22 samples have equal variances.
• 01:25 So the F Test will need to be done first in order to make the correct selection.
• 01:30 If we didn't do the F Test, assume unequal variances.
• 01:33 The p value may not be quite as precise but at least we'll err on the safe side.
• 01:39 In Excel go to the data analysis menu on your data ribbon.
• 01:43 Then select either t-Test Two Sample Assuming Equal Variances or
• 01:47 t-Test Two Sample Assuming Unequal Variances.
• 01:51 The data entry is done the same way we've done it on other tests.
• 01:54 In Minitab select the pull down menu Stat, Basic Statistics and then 2-Sample t.
• 02:01 Enter the data in the same manner in which you have with other tests.
• 02:04 You then select option button and can select the item for equal variances.
• 02:09 Otherwise, Minitab will assume unequal variances.
• 02:13 With both Excel and Minitab, you will get results with a p value that you can use to
• 02:18 either reject the null hypothesis or fail to reject the null hypothesis.
• 02:23 Next I'd like to discuss the paired T test.
• 02:26 Paired T test is a special case test.
• 02:29 It determines if the mean from two samples that are linked or
• 02:32 paired together, have a similar mean.
• 02:35 Pairing requires that every data item in one data set be paired
• 02:40 with a unique data item in the other data set, a one to one pairing.
• 02:45 In Lean Six Sigma projects, this is often done for before after tests.
• 02:50 So one obvious characteristic of the paired data samples
• 02:53 is that they must have exactly the same number of items in each sample.
• 02:57 Also, when recorded in the sample,
• 02:59 the paired items must be in the same order to ensure that the pairing is correct.
• 03:04 The hypothesis is similar to the Two-Sample T-Tests.
• 03:08 The null hypothesis is that the mean from the two samples are the same and
• 03:12 the alternative hypothesis is that they are not the same.
• 03:16 In Excel the test is done by selecting the Data Analysis menu
• 03:20 from the Data Ribbon then select t-Test Paired Two Sample for Means.
• 03:25 Then enter the data ranges as you have with other Excel tests.
• 03:29 In Minitab select the stat pull down menu, then select basic statistics, and
• 03:34 select Paired T.
• 03:36 Again enter the data columns the same way we have with other tests, and
• 03:39 you can select the greater than or less than from the options panel.
• 03:44 Let's spend just a few moments and
• 03:45 consider why we would ever need a Paired T test.
• 03:49 One condition that I've already mentioned is the Before After test.
• 03:53 In this case, items in the data set are tested to establish a baseline value.
• 03:58 Then the improvement or change is made to the process.
• 04:02 Now all the items are tested again to determine
• 04:05 if the mean value has changed in a statistically significant manner.
• 04:09 An example could be an average time for each operator to complete a unit of
• 04:12 production before and after some new aids or tools were installed.
• 04:17 Another condition is the control test.
• 04:20 In this case, some items in the data set are controlled and
• 04:23 some are exposed to the improvement.
• 04:26 All items go through the same process at the same time.
• 04:29 The performance of the control item is compared to the improved items.
• 04:34 An example of this would be to use two different cleaning processes
• 04:38 on different portions of the same carpet.
• 04:41 A third condition is matching subjects.
• 04:43 In this case,
• 04:44 the item under investigation is already a pair, such as package and label.
• 04:49 Although the items may be different in some important characteristics, they both
• 04:53 go through the same conditions or process, and the results are then compared.
• 04:58 This is often used in behavior testing with husbands and wives.
• 05:04 Whether it's the two sample T Test or
• 05:06 the paired T Test, you will likely find that these tests are the most commonly
• 05:10 used ones that you'll work with during the analyze phase of your project.

Lesson notes are only available for subscribers.