| Analysis of Variance |
Assigning Measurement Levels
The DRUG data set contains results of an experiment carried out to evaluate the effect of four drugs (DRUG) with three experimentally induced diseases (DISEASE). Each drug by disease combination was applied to six randomly selected dogs. The response variable is the increase in systolic blood pressure (CHANG_BP) due to the drug treatment. DRUG and DISEASE are classification or class variables; that is, variables that identify distinct levels or groups. DRUG contains four levels or classes and DISEASE contains three.
 |
Open the DRUG data set. |
Figure 15.2: Data Window
A variable's measurement level determines the way it is treated in analyses. In the data window, measurement levels appear above the variable names, in the upper right portion of the column header. SAS/INSIGHT software supports two measurement levels: interval (Int) and nominal (Nom).
Interval variables contain values that vary across a continuous range. In this data set, the change in blood pressure (CHANG_BP) is an interval variable.
Nominal variables contain a discrete set of values. In this data set, both DRUG and DISEASE contain a discrete set of values. However, since these are numeric variables, by default they have interval measurement levels (Int).
You need to assign both these variables the nominal measurement level (Nom) in order to treat them as classification variables. To do so, use the data measurement level pop-up menu.
|
Click on the Int measurement level indicator for the variable DRUG. |
This displays a pop-up menu.
Figure 15.3: Measurement Levels Menu
The radio mark beside Interval shows the current measurement level. Because DRUG is a numeric variable, it can use either an interval or a nominal measurement level.
|
Choose Nominal in the pop-up menu to change DRUG's measurement level. |
|
Repeat these steps to change the measurement level for DISEASE. |
Check the measurement levels for DRUG and DISEASE in the data window. Both have Nom measurement levels.
Figure 15.4: Data with Nominal Variables DRUG and DISEASE
Copyright © 2007 by SAS Institute Inc., Cary, NC, USA. All rights reserved.