Computer-Mediated Communication Versus Vocal Communication and the Attenuation of Pre-Interaction Impressions

Abstract

Conflicting theoretical approaches yield divergent predictions about the effects of telephones versus computer-mediated communication (CMC) in the persistence or dissipation of pre-interaction expectancies. Moreover, different theoretical orientations and their underlying assumptions often invoke different methodologies, which can bias the results of research. The current studies articulate and assess rival hypotheses from alternative theoretical paradigms to uncover how CMC and vocal communication affect interpersonal impressions. Methodological issues in past CMC research are evaluated that limit the generalizability of previous findings in the area. Experiments employing alternative assumptions and methods indicate that CMC is functionally equivalent to vocal communication in its ability to ameliorate expectancies and that in some cases it can be superior in transmitting positive impressions.

Notes

The authors express their gratitude to Caleb Carr, Morgan Green, Jinsuk Kim, Erin Spottswood, Shushanna Uhe, and Brandon Van Der Heide for their assistance with data collection and coding, and to Elly Konijn and two anonymous reviewers for their helpful suggestions.

1. Epley and Kruger (2005) describe their research and the interface they used as “e-mail” (see p. 416), although the software that they employed, Chatter 2.03, “is a program that allows more or less real time text message communication with other computers” (Info-Mac, 1996). They argue that it represents e-mail because “participants were not actually engaged in a fully interactive chat session, but were instead trading fully formed messages back and forth, just as they would in e-mail” (Personal correspondence, N. Epley, June 12, 2009). Nevertheless, messages were exchanged immediately rather than asynchronously over some extended period of time. E-mail is widely considered asynchronous and in this way distinct from real-time chat (e.g., Herring, 2001; Honeycutt, 2001; Kalman, Ravid, Raban, & Rafaeli, 2006; Kiesler et al., 1984; Latane & L'Herrou, 1996; Riva & Galimberti, 1998; Zhou, Burgoon, Nunamaker, & Twitchell, 2004). By having a real-time discussion, cues that e-mail users employ to help form impressions, such as the lag time between responses (Kalman et al., 2006; Walther & Tidwell, 1995)were unavailable. The current study also employed a synchronous chat system in order to control this factor, since differences between e-mail and chat are not inherently interactivity-related and are beyond the scope of the current investigation.

2. The demographic diversity of participants in Studies 1 and 2, albeit small, represents another departure from Epley and Kruger (2005). That research limited participation to Caucasians in order to control for the potential influence of subjects' actual race in the process of impression change, because the bogus stereotype inductions relied in part on racial cues. In the present study, interviewees'sex determined whether they participated in Study 1 or 2, but they were randomly assigned to conditions demographic characteristics notwithstanding. All analyses reported in the remainder of this research were repeated but without the data from cases in which interviewees were not Caucasians, and no changes occurred among the patterns of findings which obtained when all participants' data were included.

3. Although the same characteristics were used by Epley and Kruger (2005), the utility of the picture and descriptions' effects in arousing stereotyped expectancies was pretested in a 2 × 2 × 2 × 2 between-subjects factorial experiment using different subjects (N = 173), reflecting 16 different combinations of race, GPA, major, and greatest high school achievement. While the attributes did not each produce a main effect on intelligence perceptions, comparisons indicated that the combination of attributes described above generated both the largest significant differences on intelligence ratings, t(20) = 5.22, p < .001, and the greatest absolute differences between respective means (intelligent M = 5.66, SD = .38, n = 10, compared to unintelligent M = 4.14, SD = .85, n = 12).

4. Epley and Kruger (2005) reported one additional experiment in which actual persons acted as interviewees in the CMC condition, rather than using transcriptions of voice-condition interviewees. They report no differences from their other experiments. However, in this experiment (and others), interviewers and interviewees were directed not to say anything during their conversations aside from the questions and answers, constraining interactivity in the manner discussed throughout this article.