Seminar: 3. The Influence of Witness Appearance and Demeanour on Witness Credibility 107 3. The Influence of Witness Appearance and Demeanour on Witness Credibility: a theoretical framework CHRIS FIFE-SCHAW Dept of Psychology, University of Surrey, Guildford GU2 5XH INTRODUCTION This article reviews the literature on witness demeanour and assesses its importance as a factor that may influence jurors' deliberations. This is a topic that has received relatively little research attention in comparison to other fea- tures of criminal trials but is becomingincreas- ingly recognized as a possible source of bias in jury deliberations (LeVan, 1984; Stone, 1991). This article draws on Petty and Caccioppo's (1986) Elaboration Likelihood Model of persua- sion (ELM) as a useful framework for under- standing how such factors may influence jury decision making and as a guide for framing future research questions. To begin with I should be clear about what is intended by the term 'demeanour'. In keep- ing with Stone's (1991) definition, "demeanour" excludes the content of evidence, and includes every visible or audible form of self- expression manifested by a witness whether fixed or variable, voluntary or involuntary, simple or complex.' (p. 822) The important point for the discussion here is that the concern is with cues available to jurors about the witness that are logically or empiri- cally demonstrated to be unrelated to the con- tent, accuracy or quality of the evidence being given. By far the greater part of the research lit- erature on jury responses to witnesses has been concerned with the relationship between the perceived confidence of witnesses and the accuracy of their testimony (Wells and Loftus, 1984). However, since there is a generalized bias in favour of believing that more attractive people are honest (Dion et al., 1972) there seems to have been little concern that attrac- tiveness and other demeanour cues in addition to confidence might influence juries' evalu- ations of witnesses' evidence. Eyewitness testimony is often central to a criminal case and is widely regarded by the public as among the most damning kind of evidence available. All too frequently, though, the apparent value to a case of having an eye witness overrides a consideration of the probability, though it may be small, that an eye witness has made an identification error (see Stephenson, 1992, for a review). A great deal of research has focused on sources of eyewitness error (e.g. Loftus, 1979) and there is a growing body of research that has looked at the relationship between percep- tions of witness confidence/certainty and witness accuracy. Eyewitness certainty is par- ticularly interesting, since it has been shown that even though there may be no real relation- ship between confidence and accuracy, jurors seem to believe witnesses when they appear confident and express certainty about their evidence. For example, Wells et al. (1979) found that jurors' attributions of eyewitness confidence accounted for 50 per cent of the variance in jurors' decisions to believe wit- nesses, even though witness confidence was not found to be related to witness accuracy (see also Lindsay et al., 1981; Wells and Leippe 1981; Wells and Lindsay 1985; Bennett and Fife-Schaw, 1993). Indeed, Bothwell et al. (1987) comment that attorneys have known for a long time that jurors rely heavily on a wit- ness's degree of confidence as an index oftheir accuracy and take active steps to either rein- force or undermine perceptions ofconfidence in the service of their case. In a meta-analyticinvestigation Bothwell at al. (1987) concluded that we should be very cautious in assessing the utility of confidence for predicting accuracy in actual cases, as they found a correlation of only r = +0.25 taken across a corpus of 35 crime simulation studies. Further, in a review of studies between 1974 and 1982, Wells and Murray (1984) analysed research on the confidence-accuracy relation-