Search

The importance of format, graphic layout, and question routing instructions on the quality of survey data has been acknowledged for self-administered survey instruments, but the significance of these factors for questionnaires administered by interviewers has not been recognized equally. This paper examines two studies in which interviewers used different questionnaire designs to administer the same set of survey questions to randomly selected adult respondents in housing units that made up area probability samples of the Detroit metropolitan area. The paper presents empirical evidence of specific effects (questions skipped in error, unprobed answers) directly attributable to the choice of questionnaire design. The analysis shows that questionnaire design choices can either help or hurt the quality of data collected by interviewers. Furthermore, the behaviors of experienced and inexperienced interviewers are affected in similar ways. In other words, interviewing experience does not compensate for format deficits in the design of survey instruments. [ABSTRACT FROM AUTHOR]

Interviewer observations of household characteristics can be useful variables for constructing post-survey nonresponse adjustments, particularly if the observations are good proxies of key variables collected later in the survey. The U.S. Census Bureau and other survey data-collection organizations now systematically design interviewer observations with this purpose in mind. However, because these observations are typically estimates or judgments made by the interviewers, they may be incorrect. The effectiveness of nonresponse adjustments using these observations is hampered by such errors. Thus, knowledge about both household- and interviewer-level factors that affect the accuracy of interviewer observations is important for survey researchers in their efforts to design interviewer observations. To date, only one published study of a face-to-face household survey has attempted to identify these factors. This paper attempts to help fill this gap in knowledge by presenting multilevel models of the accuracy of a housing-unit observation recorded by interviewers in the National Survey of Family Growth (NSFG): whether or not young children under the age of 15 are present in the household. The presence of children in the household is an important correlate of key variables in the NSFG and many other socio-economic surveys, and is related to survey participation in general. The accuracy of this observation was found to be a function of selected household-level factors and interactions between household- and interviewer-level factors. Implications of these results for practice and directions for future research are discussed. [ABSTRACT FROM PUBLISHER]

Fair evaluation of interviewers based on their cooperation rates is complicated in most surveys. Depending on the mode(s) of data collection and the methods used to assign sampled cases to interviewers, different interviewers often work cases that are more or less difficult to interview. Further, in centralized telephone facilities, interviewers are given cases affected by the results of earlier attempts by other interviewers. This paper proposes and evaluates an interviewer performance indicator that attempts to repair these weaknesses and can be used in all data collection modes involving interviewers. For each contact made by an interviewer, the deviation between the outcome of that contact (1 = successful interview, 0 = other) and the predicted probability of a successful interview for that contact (according to auxiliary information and survey paradata) is recorded. Each interviewer then receives a score that is the mean of these deviations across all contacts. This performance indicator gives larger credit to the interviewer who obtains success on difficult cases versus relatively easy cases. The indicator also gives smaller penalties to failures on very difficult cases. We analyze face-to-face and telephone contacts from three different surveys to illustrate the computation of this indicator and its properties. We find that calling-history paradata are the strongest predictors of obtaining interviews in both modes (especially for face-to-face contacts), and produce propensity-adjusted performance indicators that more effectively distinguish among interviewers. [ABSTRACT FROM PUBLISHER]

Audio computer-assisted self-interviewing (audio-CASI or ACASI) has been widely adopted around the world as a method for eliciting more candid responses to sensitive questions in surveys. While few studies have explored the added advantages audio may bring over text-CASI, those that have (e.g., Tourangeau and Smith, 1996, Public Opinion Quarterly 60(2):275-304) found large effects for self administration over interviewer administration, but only modest additional gains from the audio enhancement. In this paper, we explore the use of audio-CASI versus text-CASI in a national survey of fertility-related issues in the United States (the National Survey of Family Growth). In the pretest, male and female respondents were randomly assigned to audio-CASI (n = 299) or text-CASI (n = 312). We compare the distributions of substantive responses between modes and examine a variety of paradata (e.g., keystroke files, time stamps) to examine the use of the CASI instruments. The main study, which interviewed 7,643 women and 4,928 men aged 15-44 in 2002-03, used audio-CASI only, but again we have a variety of paradata and interviewer debriefing items to examine the extent to which subjects made use of the audio enhancements to CASI. Our results indicate that most respondents make Iimited use of the audio features of audio-CASI and accordingly the gains produced by this technology are modest relative to text-CASI. [ABSTRACT FROM AUTHOR]

Traditional statistical analyses of interviewer effects on survey data do not examine whether these effects change over a field period. However. the nature of the survey interview is dynamic. Interviewers' behaviors and perceptions may evolve as they gain experience, thus potentially affecting data quality. This paper looks at how interview length and interviewer evaluations of respondents change over interviewers' workloads. Multilevel models with random interviewer effects are used to account for the clustering of cases within interviewers and individual interviewer characteristics in the 1984, 1988, and 2000 National Election Studies. The 1984 and 1988 NES released sample in four replicates, minimizing the confound between order in an interviewers' workload and sample composition. We find that over the course of the studies, both measures change significantly. Interviewer prior survey experience also was significantly negatively related to the length of the interview. These findings have implications for interviewer training prior to and during studies, as well as suggesting future research to reveal why these behaviors and perceptions change. [ABSTRACT FROM AUTHOR]

This article explores how interviewer gender influences responses to marriage-related questions in face-to-face surveys utilizing the 2006 China General Social Survey (CGSS). The specific marriage items examined concern respondent’s views of the happiness of married people and normative views of marriage and ask: (a) whether married men are happier than unmarried men; (b) whether married women are happier than unmarried women; (c) whether getting married is better than being single; and (d) whether staying married is better than getting divorced. Drawing from social desirability theory, this article examines two specific research questions. First, does the gender of the interviewer influence respondents’ normative views of marriage? Second, do the effects of interviewer’s gender on these questions differ for male and female respondents? The results provide mixed support for gender-of-interviewer effects. Consistent with social desirability theory, the findings indicate that female interviewers elicit more pro-marriage and neutral responses than negative responses to the first two items than male interviewers. For the most part, social desirability theory was not supported for the last two items. Gender differences in the susceptibility to gender of interviewer effects were inconsistent across models. [ABSTRACT FROM PUBLISHER]