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 5-1 summarizes the different scales of measurement, Table 5-2 tells you which type of research question requires each scale of measurement, and Table 5-3 tells you what measures produce what kind of data. Thus, by just reviewing two pages (pp.135-136), 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 scale of measurement. You also have to weigh practical issues, ethical issues, and validity issues.