Survey Research Modes

In-Person Interviews

The in-person or face-to-face interview occupies a preeminent position in scientific survey research. Such interviews typically are carried out in respondents' homes, although, historically, survey firms (“houses”) have used other venues as well. In-person interviews have been a major data [Page 2572]collection device since the inception of survey research and—when used in conjunction with probability samples of populations of interest—still are considered by many scholars to be the gold standard for valid and reliable measurement. The claim is that experienced, well-trained interviewers doing in-person surveys can gather higher quality data than would be possible using any other data-gathering mode. Working in the face-to-face mode, interviewers can build rapport and trust with respondents, thereby minimizing social desirability biases that threaten data quality. Also, although in-person surveys have a long history, they are not technically primitive. Using computer-assisted personal interview techniques, sophisticated experiments can be conducted without recourse to a researcher's lab. Data can be e-mailed back immediately from the field to project headquarters, ensuring the close and continuous monitoring of a survey's progress.

Despite its virtues, the face-to-face mode has significant drawbacks that have prompted some researchers to consider alternative data collection modes. One very serious problem is the high cost. National surveys with large probability samples have become extremely expensive, and their cost is escalating rapidly. Another problem is declining response rates. Although some major in-person surveys continue to have very high response rates, others have experienced a sizable erosion in the percentage of people willing to participate.

Telephone Interviews

Starting in the 1980s, social scientists and commercial pollsters reacted to these emerging problems by employing computer-assisted telephone interview methods, with probability samples chosen by random digit dialing procedures. In addition to their relatively low cost, these surveys can be conducted quickly, with fieldwork completed in a fraction of the time needed for a face-to-face study. This is a very important consideration for researchers interested in monitoring the dynamics of public opinion. Using telephone interviewing techniques, researchers can conduct daily interviews with random subsets (replicates) of larger samples to produce finegrained portraits of the evolving public mood during national election campaigns. Yet another desirable feature is that telephone interviewing facilitates survey experiments.

These pluses notwithstanding, the attractiveness of the telephone mode has faded in recent years. Response rates have fallen substantially, with implications that are not yet well understood. Although some studies suggest that the impact of low response rates on data quality is not particularly pernicious, deleterious effects may vary by subject matter. Coverage problems are also increasing as people abandon landline telephones for cell phones that are not reached using traditional techniques. And there is emerging evidence of an enhanced social desirability bias relative to other modes.

...