Most of the steps to getting a sensitive measure are pretty straightforward. A test with 100 items will probably be able to better distinguish between people than a test with 1 item, a 1-7 rating scale is more sensitive than a "yes/no" questions, a valid test is more sensitive than a less valid test, a reliable measure is more sensitive than a less reliable measure.
Lack of sensitivity is only one way in which a measure may prevent you from finding the answer to your research question. Another way that a measure can let you down is if it doesn't provide you with the scale of measurement you need to answer your research question. Table 6-1 summarizes the different scales of measurement, Table 6-2 tells you which type of research question requires each scale of measurement, and Table 6-3 tells you what measures produce what kind of data. Thus, by just reviewing two pages (pp. 155 and 157), you can review the key issues involved in scales of measurement.
The chapter concludes by explaining that choosing a measure involves more than thinking about sensitivity and scales of measurement. You also have to weigh practical issues, ethical issues, and validity issues.
Back to Chapter Menu