Public Opinion Quarterly - current issue

Editor's Note

Miller, P. V. — 2008-03-14

A Comparison of Address-Based Sampling (ABS) Versus Random-Digit Dialing (RDD) for General Population Surveys

Link, M. W., Battaglia, M. P., Frankel, M. R., Osborn, L., Mokdad, A. H. — 2008-03-14

Valid and reliable public health data are becoming more difficult to obtain through random-digit dial (RDD) telephone surveys. As a result, researchers are evaluating different survey designs (i.e., sampling frame and survey mode combinations) as complements or alternatives to RDD. Traditionally, mail surveys of the general public have been limited due to a lack of a complete sampling frame of households. Recent advances in electronic record keeping, however, have allowed researchers to develop a sample from a frame of addresses (e.g., the U.S. Postal Service Delivery Sequence File, which appears to provide coverage which rivals or possibly exceeds that obtained through RDD sampling methods). To test the use of this frame for surveying adults aged 18 years and older across a wide geographic area, a pilot study was conducted as part of the 2005 Behavioral Risk Factor Surveillance System (BRFSS). The pilot compared use of a traditional, RDD telephone survey methodology to an approach using a mail version of the questionnaire completed by a random sample of households drawn from an address-based frame. The findings indicate that the mail survey approach can achieve higher response rates in low-response-rate states (<40%) than RDD (particularly when two mailings are sent). Additionally, the address frame with mail survey design provides access to cell phone only households and offers cost savings over the telephone approach. The resulting sample, however, significantly overrepresents non-Hispanic whites and people with higher levels of education.

A "Brute Force" Estimation of the Residency Rate for Undetermined Telephone Numbers in an RDD Survey

Kennedy, C., Keeter, S., Dimock, M. — 2008-03-14

In random digit dial (RDD) telephone surveys, some proportion of the sampled telephone numbers cannot definitively be classified as eligible or ineligible. Each call attempt to these numbers generally results in either a busy signal or a ring with no answer. According to the profession's standard definitions, the proportion of these unresolved numbers that are, in fact, eligible is known as "e" and should be accounted for in response rate calculations. We used call records and a directory listed indicator from a survey with an extended calling period (and other features – letters, incentives, etc.) to resolve as many unknown cases as possible and thus derive an empirical estimate (.47) for "e." We found additional support for this estimate by matching the unresolved telephone numbers from a separate survey to residency information from a commercial data vendor. The estimate of "e" applies to national RDD surveys featuring approximately a six-call design.

Effects and Costs of Tracing Strategies on Nonresponse Bias in a Survey of Workers with Low-Back Injury

2008-03-14

Declining rates of participation are an increasing challenge for studies that involve telephone surveys. This study examined the costs of a telephone survey methodology that used increasingly intensive tracing methods to track a pool of claimants who had sustained occupational back injuries. It also compared the respondent sample to people who refused the survey and/or were not located or contacted. 3,181 claimants were drawn from a database maintained by the Missouri Division of Workers’ Compensation (DWC) and 1,475 completed a telephone interview. The DWC database provided data reflecting monetary and disability outcomes for all potential participants; telephone interviews provided additional self-reported data. More intensive tracing strategies improved the representativeness of the sample, yielding more women and minority participants. Relative to less intensive techniques, advanced tracing efforts located people whose DWC records were more similar to claimants not located. While the hands-on tracing efforts reduced the apparent bias of the respondent sample, costs increased as tracing strategies intensified – over $98,000 was spent to trace 1,027 claimants who were never located or contacted. The results suggest some guidelines that may facilitate decision-making for researchers and funders who must balance the trade-off between costs and nonresponse bias when planning surveys.

The Social Roots of the Partisan Gender Gap

Burden, B. C. — 2008-03-14

I suggest that the gender gap in party identification is dependent on question wording and asymmetric stereotypes about men's and women's partisan preferences. A survey experiment reanalyzes the gender gap by comparing the standard partisan battery to an alternative version that emphasizes feelings rather than thoughts. Bringing question wording into closer alignment with theory causes the gender gap to shrink. This happens in part because the "feel" questions find women to be less Democratic than did the "think" questions. Moreover, reduction of the gender gap occurs mostly among highly sophisticated women and not those usually susceptible to question wording effects. Contrary to popular wisdom, men and women appear to be more, not less, alike politically when feelings are primed.

Social Desirability Effects and Support for a Female American President

Streb, M. J., Burrell, B., Frederick, B., Genovese, M. A. — 2008-03-14

Public opinion polls show consistently that a substantial portion of the American public would vote for a qualified female presidential candidate. Because of the controversial nature of such questions, however, the responses may suffer from social desirability effects. In other words, respondents may be purposely giving false answers as not to violate societal norms. Using an unobtrusive measure called the "list experiment," we find that public opinion polls are indeed exaggerating support for a female president. Roughly 26 percent of the public is "angry or upset" about the prospect of a female president. Moreover, this level of dissatisfaction is constant across several demographic groups.

An Experiment on the Measurement of Political Knowledge in Surveys

Sturgis, P., Allum, N., Smith, P. — 2008-03-14

In a series of articles, Mondak and colleagues argue that the conventional way of measuring political knowledge in surveys is flawed. Personality related "propensity to guess" underestimates the level of political knowledge in the population and distorts estimates of between group differences, when a DK alternative is offered. This has led Mondak to recommend the use of closed-ended items on which DKs are not explicitly offered, following best practice in the field of educational testing. In this article, we present the results of an experimental study which calls into question the wisdom of this approach. Our results show little evidence of partial knowledge concealed within DK responses; when people who initially select a DK alternative are subsequently asked to provide a "best guess," they fare statistically no better than chance. We conclude that opinion researchers should be cautious about adopting Mondak's recommendations for the design of political knowledge items in surveys.

Does "Yes or No" on the Telephone Mean the Same as "Check-All-That-Apply" on the Web?

Smyth, J. D., Christian, L. M., Dillman, D. A. — 2008-03-14

Recent experimental research has shown that respondents to forced-choice questions endorse significantly more options than respondents to check-all questions. This research has challenged the common assumption that these two question formats can be used interchangeably but has been limited to comparisons within a single survey mode. In this paper we use data from a 2004 random sample survey of university students to compare the forced-choice and check-all question formats across web self-administered and telephone interviewer-administered surveys as they are commonly used in survey practice. We find that the within-mode question format effects revealed by previous research and reaffirmed in the current study appear to persist across modes as well; the telephone forced-choice format produces higher endorsement than the web check-all format. These results provide further support for the argument that the check-all and forced-choice question formats do not produce comparable results and are not interchangeable formats. Additional comparisons show that the forced-choice format performs similarly across telephone and web modes.

How Too Little Can Give You a Little Too Much: Determining the Number of Household Phone Lines in RDD Surveys

Merkle, D., Langer, G. — 2008-03-14

Accurately assessing the number of household phone lines in random-digit dialed (RDD) surveys for use in weighting has become more complicated over time. This study evaluates phone line weighting by asking an in-depth battery of seven questions designed to measure the number of available land lines in the home more precisely. The results show that this weighting correction is not so simple, and that inadequately parsing the type and use of household phone numbers produces an inaccurate weight. Failing explicitly to remove irrelevant phone numbers, especially cell phones, produces a larger-than-appropriate weighting adjustment. Potentially more important on a practical level is that, when done correctly, and even when "overdone," phone line weighting had little impact across a diverse set of opinion and demographic variables.

The Polls--Review: Reply to Berinsky and Druckman: Success Still Matters

Gelpi, C., Reifler, J. — 2008-03-14

The Polls--Trends: Labor Unions in the United States

Panagopoulos, C., Francia, P. L. — 2008-03-14

In this study, we analyze data on public opinion and attitudes toward labor unions from the iPOLL Databank at the Roper Center for Public Opinion Research (University of Connecticut), the American National Election Study, and the Current Population Survey. Despite recent developments that suggest labor unions are in decline, we find organized labor has maintained reasonably strong public support. Although the data indicate that Americans remain skeptical about how much confidence they can place in unions and their leaders, the results make clear that the public continues to recognize the need for unions to protect the rights of workers. These results hold potentially important implications for the future of organized labor in the United States.

Sarah E. Igo. The Averaged American: Surveys, Citizens, and the Making of a Mass Public

Kohut, A. — 2008-03-14

Manuscript Referees, 2006-2007

2008-03-14