 |
|
 | |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|||||||
|
In epidemiologic studies involving large numbers of subjects, attention must be paid to standardizing data collection methods so that all subjects are interviewed and examined in the same way. These considerations become even more important when data collection takes place over a long period of time, either in large cross-sectional studies or in longitudinal designs, because study personnel may change over the course of the study, and individual data collectors may develop specific idiosyncrasies or "drift" in their measurements. Careful development and pre-testing of self-report questionnaires and interview instruments will increase the probability that all subjects interpret the questions in a similar way. Interview specifications, providing standardized methods of probing for more information and for answering subjects' questions about the meaning of terms, are also essential. Interviewers must be trained in the general principles of interviewing, as well as in the use of the specific interview instrument. In studies
involving clinical examinations, it is likely that multiple examiners
will need to be employed and that each examiner will need to conduct
examinations on a different subset of the study subjects. Under
these circumstances, it is essential to insure that, to the extent
possible, different examiners would find similar clinical signs
and symptoms or arrive at a similar diagnosis for the same subject.
There are several steps that can be taken in order to maximize
inter-examiner reliability, which are explored in the screens
that follow.
|
|||||||
