Log in / Register
Home arrow Environment arrow Research Methods in Anthropology: Qualitative and Quantitative Approaches

Disadvantages of Telephone Interviewing

  • 1. If you are doing research in Haiti or Bolivia or elsewhere in the developing world, telephone surveys are out of the question, except for some urban centers, and then only if your research is about relatively well-off people. About 97% of all households in the United States have telephones (Belinfante 2009). This makes national surveys a cinch to do and highly reliable. But the distribution of telephones is uneven, which makes some local surveys impossible to do by phone. On the Rosebud Sioux Reservation—an area of about 5,000 square kilometers and home to about ten thousand tribal members—telephone penetration is about 75% (FCC 2007).
  • 2. Telephone interviewing using RDD is convenient, but it’s no lazy way out: It can take 1,000 calls to get 200-300 interviews—and that’s starting with a list of working home phone numbers.

People are getting tired of phone surveys and are opting out (Morin 2004). In a careful study in 2003, the contact rate (the number of people you can actually talk to in a sample of working home phone numbers) was 79% (down from 90% 6 years earlier) and the refusal rate among people contacted was 66%. All in all, the final response rate was about 25% (Keeter et al. 2006). This biases the outcome for some questions, but the good news is that it doesn’t produce consistent bias in the results of surveys (Groves 2006). You can always get nearly 100% study sample completion by replacing refusers with people who will cooperate. (If you do that, make an extra effort to get at least some of the refusers to respond so you can test whether cooperators are a biased sample [box 9.4].)

BOX 9.4


There are companies that sell telephone numbers for surveys. The numbers are chosen to represent businesses or residences and to represent the varying saturation of phone service in different calling areas. Even the best sample of phone numbers, though, may not be enough to keep you out of trouble. During the 1984 U.S. presidential election, Ronald Reagan's tracking poll used a list of registered voters, Republicans and Democrats alike. The poll showed Reagan comfortably ahead of his rival, Walter Mondale, except on Friday nights. Registered Republicans, it turned out, being wealthier than their counterparts among Democrats, were out Friday nights more than Democrats were, and simply weren't available to answer the phone (Begley et al. 1992:38). [1]

can keep them on the line for a remarkably long time (up to an hour) by developing special ‘‘phone personality” traits. Generally, however, you should not plan a telephone interview that lasts for more than 20 minutes.

4. And finally, this: It has long been known that, in an unknown percentage of occasions, hired interviewers willfully falsify data (Boyd and Westfall 1955). When an interviewer who is paid by the completed interview finds a respondent not at home, the temptation is to fill in the interview and get on to the next respondent. It’s particularly easy for interviewers to cheat in telephone surveys—from failing to probe, to interviewing unqualified respondents, to fabricating an item response, and even to fabricating whole interviews. Kiecker and Nelson (1996) hired 33 survey research companies to do eight interviews each, ostensibly as ‘‘mop-up’’ for a larger national market survey. The eight respondents were plants—graduate students of drama, for whom this must have been quite a gig—and were the same eight for each of the surveys. Of the 33 interviewers studied, 10 fabricated an entire interview, 32 fabricated at least one item response, and all 33 failed to record responses verbatim.

You can eliminate most cheating by training and monitoring phone interviewers. Presser and Zhao (1992) monitored 40 trained telephone interviewers at the Maryland Survey Research Center. For the 5,619 questions monitored, interviewers read the questions exactly as worded on the survey 91% of the time. Training works.

Still, no matter how much you train interviewers .... Johnstone et al. (1992) studied 48 telephone interviews done entirely by women and found that female respondents elicited more sympathy, while male respondents elicited more joking. Men, say Johnstone et al., may be less comfortable than women are with being interviewed by women and wind up trying to subvert the interview by turning it into teasing or banter.

  • [1] Telephone interviews must be relatively short, or people will hang up. There is someevidence that once people agree to give you their time in a telephone interview, you
Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
Business & Finance
Computer Science
Language & Literature
Political science