The Psychometric Properties of a Preliminary Social Presence Measure Using Rasch Analysis

Abstract

Social presence is an important construct in computer mediated communication, such as found in online collaborative learning (OCL) settings. It is hypothesized that social presence influences the degree of perceived learning and learning outcomes of OCL group members. However, the construct social presence is contested as many incompatible definitions exist in the research community and so do the many measures of social presence. Also, none of the existing social presence measures has undergone a rigid construct validation process such as proposed by Rasch Measurement theory. As a result, hypothesis testing using these measures produced unreliable findings. To address this undesirable situation, we returned to the original definition of Short et al. [29] and redefined it as the degree to which the other person is perceived as physical ‘real’ in the communication. We present a social presence measure that assesses this perception of realness. Rasch analysis was used to validate the raw social presence measure. Our findings revealed that measuring the degree of realness was excellent for those who have high perceptions of realness of the other (i.e., they could be well differentiated), whereas this was moderate for those who have low perceptions (i.e., they could be less well differentiated). Our conclusion is that the social presence measure is already an improvement when compared to existing social presence measures that emphasize realness but it surely needs further improvement: those who have low perceptions of realness should equally well be differentiated as those with high perceptions of it.