 |
StatXact Example 7
Clinical investigations often involve data in the form of ordered categories. Moses, Emerson, and Hosseini in a survey of research articles published in the New England Journal of Medicine found that 20% of these articles had ordinal categorical data. The data below were part of this survey. Appropriate analyses for such data can be carried out easily and efficiently with StatXact. Indeed StatXact was the package of choice for Moses, Emerson and Hosseini in their expository article on "Analysing Data from Ordered Categories", published in Medical Uses of Statistics, 2nd Edition, edited by Bailar and Mosteller, NEJM Books, Boston, 1992.
Adverse Reactions in Children Infected with Varicella
A randomized clinical trial of Interferon and placebo was conducted on 44 children infected with varicella (chicken pox) (Arvin, et al., 1982). One of the end points of the study was to determine whether Interferon is more effective than placebo in preventing adverse effects. There were four ordinal categories of adverse effects. The number of children falling in each category, by treatment, is tabulated below:
|
Categories of Adverse Effect |
| Treatment |
None |
Life Threatening |
Death in 2-3 Weeks |
Death in 1 Week |
| Interferon |
21 |
0 |
2 |
0 |
| Placebo |
15 |
3 |
1 |
2 |
The first step is to input these data into StatXact. The StatXact screen below shows how easy this is. You simply fill in a 2 x 4 contingency table using a friendly spreadsheet-like editor as shown below:
Once the data are in, you run the exact Wilcoxon rank sum test by typing WI /EX in the StatXact Command Bar, or by choosing from the Statistics menu "Two Independent Samples (with or without Strata)", and on that menu "Wilcoxon-Mann-Whitney...". The following output screen is displayed.
Top
The exact p-value is 0.0743, while the approximate p-value is 0.0502. Although in this case, neither p-value indicates that interferon is any more effective than a placebo, notice that the exact p-value is about 50% larger than the approximate one computed by asymptotic theory. This example illustrates both how easy it is to analyze categorical data with StatXact, and how different the exact and approximate answers can be. A few minutes familiarity with the package and you can easily generate your own exact p-values or confidence intervals. No other commercial software makes it this easy, and certainly none gives you exact p-values for anything more than a simple 2 x 2 table.
|
|
|
|
| |
Example Submissions
Over the years, most of the examples have come from in-house statisticians, but we would like our examples to better reflect the actual uses customers have for our products.
Therefore, we would like to ask for data from our customers for use as examples. Any data we use would be properly cited, and with your permission, could be used for examples in our manuals, in marketing material, or even for one of our live webcasts.
This is a chance to have your hard work published and made known to many other statisticians in your field and in other disciplines.
 Get Acrobat Reader
|
|
