When we worked through the examples during Week 7 we only looked at 2 factors. However, it is very rare that we are only interested in looking at 2 factors. In the blog I talk about adding a third factor and using the Mantel-Haenszel statistic Take a look at that blog entry and re view the material we discussed during Week 7.
The following data is taken from a SAS example. Macknin, Mathew, and Medendorp (1990) conducted a trial trying to determine whether steam inhalation had an effect on patients with the common cold. For more information on the study please refer to the paper whose reference will be listed at the end of this page.
The patients were exposed to two 20 minute steam treatments. Each patient then self -assessed on the following 4 days the level of their nasal drainage on a 4-point scale (1=mild symptoms, 2=moderate symptoms, 3=moderate symptoms, 4=severe symptoms). There were a total of 30 participants in the study. The null hypothesis is that the distribution of the symptom severity scores is the same on each of the four days for each patient.
| Patient ID | Study | day | Patient ID | Study | day | ||||
| 1 | 2 | 3 | 4 | 1 | 2 | 3 | 4 | ||
| 1 | 1 | 1 | 2 | 2 | 16 | 2 | 1 | 1 | 1 |
| 2 | 0 | 0 | 0 | 0 | 17 | 1 | 1 | 1 | 1 |
| 3 | 1 | 1 | 1 | 1 | 18 | 2 | 2 | 2 | 2 |
| 4 | 1 | 1 | 1 | 1 | 19 | 3 | 1 | 1 | 1 |
| 5 | 0 | 2 | 2 | 0 | 20 | 1 | 1 | 2 | 1 |
| …. | |||||||||
| 15 | 2 | 3 | 3 | 3 | 30 | 3 | 3 | 3 | 3 |
Using the blog entry highlighted above and the material we reviewed in Week 7 – do you accept or reject the Null hypothesis stated above? How did you draw your conclusion?
Here is the data already for you in SAS. Review the coding used, does it make sense? If not, please let me know.
SASsy Fridays 