Major Sources of Measurement Error

MAJOR SOURCES OF MEASUREMENT ERROR.

In personal interviewing, the researcher must deal with bias and cost.  While each is discussed separately, they are interrelated. These are as follow:

a.         Sampling Errors.    A second criterion of a good sample design is precision. No  sample will fully represent its population in al respects. The numerical descriptors that describe samples may be expected to differ from those that describe populations because of random fluctuations inherent in the sampling process. This   is called Sampling Error and reflects the influences of chance in drawing the sample members. Sampling error is what is left after all known sources of systematic variance have been accounted for. In theory, sampling error consists of random fluctuations only, although some unknown systematic variance may be included when too many or too few sample elements possess a particular characteristic.

b.         Nonresponse Error.          Availability is an important source of error for personal interviews. This no response error occurs when you cannot locate whom you are supposed to study. It is an especially difficult problem when using a probability sample of subjects. In such a case, there are predestinated persons to be interviewed; the task is to find these respondents. If you are forced to interview substitutes, an unknown but possibly substantial bias is introduced. One study of no response found that only 31 percent of all first calls (and 20 percent of all first  calls in major metropolitan areas) were completed. The best  first-call of 52 percent was for rural male respondents contacted after 6 PM  on a weekday.

c.         Response Error.     When the data reported differ from the actual data, response error occurs. There are many ways such errors can happen. Errors can be made in the processing and tabulating of data. Errors occurs when the respondent fails to report fully and accurately. One study found that liquid asset holdings were typically underestimated by as much as 25 to 50 percent. Other data, such as income and purchases of consumer durables, are more accurately reported. Respondents have difficulty in reporting fully and accurately on topics that are sensitive or involve ego matters. Consistent control or elimination of such respondent bias is a problem that yet to be solved. The best advice to use trained interviewers who are knowledgeable about such problems.

d.         Interviewer Errors.             It is also a major source of response bias. From the introduction to the conclusion of the interview, there are many points where the interviewer’s  control of the process can affect the quality of the data. There are three different kinds of error concerning the interview techniques just discussed.

(1)       The sample loses credibility and is likely to be biased if interviewers do not do a good job of enlisting respondent cooperation.

(2)       The precision of survey estimates will be reduced and there will be more error around estimates to the extent that interviewers are inconsistent in ways that influence the data.

(3)       Answers may be systematically inaccurate or biased when interviewers fails to appropriately train and motivate respondents or fail to establish an appropriate interpersonal setting for reporting what is called for.