The sense of presence in virtual environments:

Abstract. The sense of presence is the feeling of being there in a virtual environment. A three-component self report scale to measure sense of presence is described, the components being sense of spatial presence, involvement, and realness. This three-component structure was developed in a survey study with players of 3D games (N = 246) and replicated in a second survey study (N = 296); studies using the scale for measuring the effects of interaction on presence provide evidence for validity. The findings are explained by the Potential Action Coding Theory of presence, which assumes that presence develops from mental model building and suppression of the real environment.

Keywords: sense of presence, immersion, involvement, virtual environments, virtual reality

Zusammenfassung. Präsenz in einer virtuellen Umgebung zu erleben, heißt, das Gefühl zu haben, sich in dieser virtuellen Umgebung zu befinden. Es wird eine Skala zur Messung von Präsenzerleben vorgestellt, die drei Komponenten erfasst: räumliche Präsenz, Involviertheit und Realitätsurteil. Diese Struktur wurde in einer Studie mit Nutzer/inne/n virtueller Umgebungen (N = 246) entwickelt und in einer zweiten Studie (N = 296) repliziert. Experimente zu den Wirkungen von Interaktivität auf Präsenz liefern erste Evidenz für die Validität der Skala. Die Ergebnisse werden auf der Basis der Potential Action Coding Theory interpretiert, die Präsenz als Folge der Enkodierung potenzieller Handlungen und der Unterdrückung der realen Umgebung erklärt.

Schlüsselwörter: Präsenz, Immersion, Involviertheit, virtuelle Umgebungen, virtuelle Realität

Presence in virtual environments is a subjective experience

It is a rare case that a technology is defined in terms of its psychological effects. However, this is what happened with virtual reality: It is defined as a technology that is capable of eliciting a sense of presence in a virtual environment (VE; Steuer, 1992). Sense of presence itself is defined as the sense of being there in the virtual environment (Slater, Usoh & Steed, 1994). This paper describes a three-component measure of subjective sense of presence using self-report scales.

Today’s VEs typically offer their users visual and auditory depictions of real or fictitious environments. Users can interact with the environment, at least in the sense that they can move their viewpoint in the virtual space. Display technologies include PC screens, head-mounted displays (HMDs), and multiple projection screens, interaction technologies include devices such as mice or joysticks, and head and body tracking. A sense of being in the virtual space can already develop with rather simple technological means, as players of 3D screen-based computer games can attest.

The sense of presence must be distinguished from the technological quality of the VE. The former is a subjective experience similar to a feeling. The latter is commonly called immersion (cf. Slater, 1999), referring to objective descriptions of the technology (e.g., the quality of the rendering or the field of view in an HMD). Of course, there can also be subjective evaluations of a system’s immersion, which, however, still need to be distinguished from presence. For instance, the item ”I had the sense that the VE surrounded me“ indicates presence while the item ”The head-mounted display had a large field of view“ evaluates immersion. The questionnaire described below included both types of items, but the focus of this paper is on presence, not technology evaluation items.

The Igroup Presence Questionnaire

Overview of scale development

The three-component Igroup Presence Questionnaire (IPQ) [1] was developed on the basis of previous scales (Carlin, Hoffman & Weghorst, 1997; Hendrix, 1994; Regenbrecht, Schubert & Friedmann, 1998; Slater et al., 1994; Witmer & Singer, 1998). Problems with these measures were that they sometimes confuse presence with technology evaluation (cf. Slater, 1999) and that no factor analyses exploring internal structures were available, although it was acknowledged early on that the presence experience is multi-dimensional (Biocca & Delaney, 1995).

Schubert, Friedmann and Regenbrecht (2001) combined previously published presence items with new items that tapped especially subjective experiences of being enclosed in, directly interacting with, and concentrating on the VE. Generation of new items took place after interviewing expert VR users. Two survey studies explored the component structure of the item pool. In Study 1, 246 participants answered a total of 75 items. [2] Two thirds of the items were borrowed from previously published scales. (Of the 14 items chosen in the end, only five came from previous publications.) Original scale anchors were kept, resulting in different response formats for some items, but this did not influence the factorial structure. Both items on presence feelings and items on technology evaluation were entered, with the goal to show that they load on different factors. A principal component analysis (PCA) with oblique rotation extracted eight components. Of these, three were identified as presence components since they described feelings and experiences with respect to the VE (and not the technology), with a total of 29 items loading on them. (The others described subjective evaluations of the technology’s interactive and immersive capabilities.) The scales were condensed using confirmatory factor analyses (CFA), eliminating items with double loadings. The obtained factor structure was then again tested on a second sample (N = 296, see [2] ). CFA revealed that the modelled structure did also fit this second independent sample, confirming its stability.

Three presence components

The three presence components were identified as spatial presence (SP), involvement (INV), and realness (REAL). For spatial presence, two sample items were “I felt present in the virtual space” and “I had a sense of acting in the virtual space instead of operating something from the outside.” SP tapped the common definition of presence as a feeling of being surrounded by the VE, directly interacting in it (without any mediation by interfaces), and a sense of transportation to another place (cf. Green & Brock, 2000; Schubert & Crusius, 2002). The items loading on the involvement component described the own attention focus on the virtual world, and residual awareness of the real world. Items were for instance “I was completely captivated by the virtual world,” and “I still paid attention to the real environment.” In general, this factor assesses how much the user has the feeling of focusing on the VE instead of focusing on the real world. But again, it is an assessment of a feeling. Finally, the realness component referred to a comparison between the virtual and the real world, or to how real the VE was judged to be. Sample items were “How real did the virtual world seem to you?”, and “How much did your experience in the virtual environment seem consistent with your real world experience?”.

In order to develop short scales from the results of the exploratory PCA, CFAs were conducted. As a starting point, the structure found in the exploratory PCA was entered, with the addition of a latent second order factor that loaded on all three first order latent factors (SP, INV, and REAL). The goal of the CFA was to identify those items that (1) either correlated only with their subcomponent or (2) with all three subcomponents at once. Items of the latter type were taken out of their respective subcomponent and loaded directly on the second order factor. Only one item confirmed this requirement. Interestingly, it was the general definition of presence: “In the virtual environment I had a sense of being there...”. In addition, the process left five items for SP, four items for INV and three items for REAL. The components correlate considerably. In the second independent sample, the correlations were: r_SP,INV = .53, r_SP,REAL = .53, r_INV,REAL = .40, all ps < .001. Internal consistencies are satisfactory, Cronbach’s Alpha_SP = .78, Alpha_INV = .74, Alpha_REAL = .63 (N = 265, cases with missings excluded). [3]

Validation

So far, the only external validation for the scale are studies showing its covariation with hypothesized causes of presence. Schubert, Friedmann and Regenbrecht (1999) showed that SP and INV correlated especially with the users’ feeling that they could interact with the VE in manifold ways, and that they could predict the result of their actions. Furthermore, the perception of dramatic narration and also a judged high quality of the immersion correlated with SP and INV.

A frequently repeated but only rarely tested assumption is that interactivity increases presence in VEs. Regenbrecht and Schubert (2002) conducted three studies to test how perceived, objective, and illusory opportunities to interact influence the three presence components. They found that all three kinds of interactivity increased spatial presence. Involvement was not affected in any of the studies, and realness was reliably affected only by more objective possibilities to interact.

In addition, the three-component structure of the IPQ is supported by results from Lessiter, Freeman, Keogh and Davidoff (2001). In a similar but independent effort, they also pooled a large collection of presence items in one survey study and reached a similar factor structure in exploratory analyses. The fact that Lessiter et al.’s sample was more balanced with respect to gender than the samples described here lessens worries that the current scale development might have been biased by gender. Furthermore, since they surveyed primarily viewers of IMAX and other stereoscopic movies, it seems that the structure is a rather general one.

The Potential Action Coding Theory of presence

But how can we explain that presence comprises these three factors, and not others? A theory of the cognitive processes leading to presence is needed. We have recently proposed such a model: the Potential Action Coding Theory of presence (PACT; Schubert et al., 2001). The core of the theory is that the mental representation of the VE determines the sense of presence. Mentally representing the VE involves two processes: mental model construction and suppression of irrelevant information, similar to comprehension of text (Glenberg, 1997). First, PACT assumes that mental representations of potential bodily actions that can be performed in the VE are constructed. Second, stimuli from the real environment that do not fit the current representation are suppressed. These two processes are ‘echoed’ in the feelings of spatial presence and involvement: The users feel present when their mental representation includes plenty of bodily actions that are possible in the VE, and the users feel involved when they successfully suppress the real environment. Furthermore, we assume that the reality component comes in as judgement performed after the construction (Gilbert & Gill, 2000). In sum, we see the successful construction of a spatial-functional model that codes potential actions of the body in the virtual space as leading to the experience that defines VR - the sense of presence.

References

Fussnoten

This paper is based on research conducted together with Dr. Holger Regenbrecht and Frank Friedmann; their contribution is gratefully acknowledged. I also thank Jan Crusius and an anonymous reviewer for comments on an earlier version of this article.

Anschrift