Publication Type: Invited Paper
Abstract: It has become common practice for forensics competitors, during the introductory sections of their speeches, to pose a research question which functions in many ways like a thesis for the speech and purports to provide a direction for the analysis and conclusions which follow. But are these research questions meaningful, or are they simply a convention which misrepresents the actual nature of the competitive speech? This paper will examine this question by considering such issues as: (1) the history of using research questions in competitive communication analysis, (2) a comparison and contrast of the nature and roles of research questions in competitive vs. other scholarly writing, (3) the degree to which research questions clarify vs. distort the analytical flow of speeches, (4) the impact of rhetorical questions on the other elements of competitive speeches (choice of methodology, application of method, drawing of conclusions), (5) the tendency of research questions to enhance vs. damage the quality of analysis conducted by forensics competitors, and (6) an assessment of the impact/significance of this issue on students preparing for advanced study in the communication discipline.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: Using data from 548 split-ballot experiments in telephone surveys conducted by the Gallup Organization, we tested whether the linguistic structure of questions moderates response order effects by influencing when and how respondents form their answer to questions. In dichotomous, orally-presented questions that explicitly instruct or implicitly encourage respondents to wait until they have heard all the answer choices before formulating an answer, recency effects predominated. This finding suggests that when hearing these questions, many respondents waited until all the response options had been read before considering any of them and many respondents began by considering the last response option they had heard. But among some questions with wordings that appeared to end before they in fact did, no recency effects appeared. This finding suggests that when hearing these questions, at least some respondents began forming their answers early, before all the response options had been read. Consistent with satisficing theory, the effect of linguistic structure was also strongest among respondents with the fewest cognitive skills, measured using their educational attainment (though not among the oldest respondents, another proxy that has been proposed for cognitive skills). Greater question complexity and later placement of a question in a questionnaire were associated with stronger response order effects. Similarly, questions with response choices that were each complete sentences demonstrated stronger response order effects than those with response choices that were single words or short phrases. All this helps to further specify the conditions under which response order effects are likely to occur and the direction they are likely to take. These results also suggest what cognitive mechanisms may be at work: respondents appear to be sensitive to cues communicated through the linguistic structure of questions about when and how to answer orally-presented survey questions.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: This paper examines response times (RTs) in the context of survey research. Response times are heavily studied in experimental and/or cognitive psychology but are only now gaining attention from survey researchers. Most work on response times in survey research so far has considered response times either as predictors or as proxy measures for some other variable of greater interest. As a result, response times have not been the main focus of the research. Focusing on the nature and causes of response times, this paper takes on the response time as a dependent variable and attempts to tackle the mechanisms of response times by seeking an appropriate regression model. Using the survey response model proposed by Tourangeau, Rips and Rasinski (2000), we include both item-level characteristics and respondent-level characteristics that have been demonstrated to have an impact on response time in a hierarchical regression model. The results from the regression model corroborate evidence from cognitive and experimental psychology that response time is affected by question characteristics such as length, number of answer categories, the nature of answer categories, and the location of the question within the questionnaire as well as respondent characteristics such as age, education and experience with the Internet. Two interaction terms--age by question length and age by the location of question--are shown to have an impact on response times, too.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: In the last twenty years, key researchers in the area of fear of crime have expressed concern about early measures of fear (e.g., Ferraro, LaGrange, Warr and Stafford). One of the questions of concern was: “Are there areas right around where you live--that is, within a few blocks--where you are afraid to walk alone?” Some criticized the question because it did not include the word crime, and they argued that questions like these probably measure perceived risk more than fear. These researchers then called for distinguishing between risk and fear and assessing fear of specific offenses to improve upon our measurement. Most recent studies have followed this advice. Using measures of perceived risk and fear of specific offenses as well as behavioral precautions, this study examines which is most correlated with the general assessment question used in prior studies in an attempt to determine whether those prior studies were more closely measuring risk, fear or behaviors.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: This study uses open ended and multiple choice questionnaires to examine the influence of question format on the cultivation effect. Student respondents (N=517) were requested to estimate the prevalence of criminality, single parent families and sexual activity among teens in the society. The given estimates were compared to real world figures and to TV world values that were obtained from a content analysis of one week of prime-time network programming (66 hours from three different channels). The results confirm to cultivation predictions by pointing at differences between heavy viewers and light viewers, so that heavy viewers give extremer estimates, but the findings also indicate that the impact of the question format is greater than the effect of cultivation, and that the amount of viewing and the question format interact so that cultivation is more visible in open-ended questionnaires.