Publication Type: Conference Paper/Unpublished Manuscript
Abstract: This paper tests the assumption that survey items that are equivalently worded are understood in the same way in different settings. The paper uses Item Response Theory (IRT) to identify items about attitudes towards family dissolution that have measurement variant characteristics (i.e. differential item functioning) across groups. The data are from the comparative project “Status of Women and Fertility” conducted in communities in India, Malaysia, Pakistan and the Philippines. The results indicate that most or all of the studied items have the same meaning within similar cultural/ linguistic countries (e.g. Pakistan) but not within heterogeneous settings (e.g. India and Malaysia).The paper also shows that respondents use different criteria in responding to attitude questions depending on the sex of the object item (i.e. whether the husband is leaving the spouse or vice versa). Pakistani, Indian, and ethnic Malay women are significantly less likely to approve of separation if the wife is leaving the spouse. Such pattern of responses is much less prevalent among men respondents.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Overall, this paper provides an analysis of how speakers turn beginning design may exhibit aspects of their turn’s planned shape and type. In other words, the paper illustrates that speakers may utilize specific linguistic components that are produced in the beginning part of their turn in order to signal that they are in the process of initiating a new activity which has been occasioned by some prior talk. Specifically, this paper provides a conversation analytical examination of the sequential placement and the interactional function of the German lexical item genau (exactly) in everyday German conversation. The data corpus for this study consists of 25 hours of German telephone conversation and face-to-face interaction between relatives and close friends. An examination of the sequential placement of the German adverb genau in combination with its semantic content revealed that this particular linguistic component may perform a specific interactional function depending on where it is placed in a turn. The analysis of turn initial components suggests that one particular placement for such a lexical item is in an environment in which an item is brought to the speaker’s mind. Genau seems to mark an epistemological shift in the speaker’s mind, namely remembering an item that he or she wanted to mention in the talk and that has just been occasioned by the prior talk. Furthermore, by inserting the linguistic component speakers signal to their co-participant that a mentionable item that has just been remembered will be mentioned in the unfolding turn. In general, the analysis demonstrates how the grammatical structure of turn-beginnings may provide information about the unfolding turn and the subsequent talk/action which may be developed.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Although political knowledge is measured by items with varying difficulty levels, communication studies have not scrutinized the interaction effects between item difficulty and media use on political learning. Drawing from American National Election Study 2004, this analysis investigates whether item difficulty moderates the impact that media use and interpersonal communication have on political learning. We find a noticeable interplay between item difficulty and media effects. While television impacts knowledge about easy political items, newspapers affect the moderately difficult ones, and the contributions that interpersonal talk makes to political knowledge do not depend on item difficulty. Also, communicative behaviors might increase the easy knowledge, whereas socio-structural variables have more explanatory power for the relatively difficult political items. We discuss theoretical and practical implications, and also suggest a new statistical model that combines measurement model for ordered items with multi-level modeling.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: Political event data---categorical data showing who did what to whom (and when) derived from news reports---are frequently converted to an interval level measurement by assigning a numerical scaled value to each event. All of these scaling systems rely on expert assessments to assign the numerical scores, and those assessment do not take into account the actual data that will be aggregated, instead relying on some abstract generalization (or wild-assed guess) of the appropriate value.
This paper uses item response theory (IRT)---a technique originally developed for the scaling of test scores---to derive scales inductively, using event data Israeli interaction with Lebanon and the Palestinians for 1991-2007. In the IRT model, the probability of an event being reported in an interval of time for a particular dyad (for example Israel and Lebanon) and news source is modeled as a logistic function. The aggregate score can then be computed as a function of these probabilities, with rare events being given a higher weight than common events. Scaled monthly scores determined through Rasch scaling, the simplest of the IRT methods, correlate well with scores computed using Goldstein's widely-used scale. The Rasch scales are less successful in reconciling the somewhat divergent sets of events derived from the \textit{Agence France Presse} and Reuters news services. These results will be compared to those obtained with the Goldstein scale, as well as determining whether IRT-derived scales can be used to normalize events derived from multiple news sources.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: This paper uses an item response model to assess the appropriateness of the votes used in interest group scores, in much the same way that educational testing methods assess the appropriateness of questions used in standardized tests.