Description

Observation of facial expression can be used to study the emotional components of presence (Huang et al., 1999). Facial expression can either be scored by human observers, or automatically recognized. There are several methods to aid manual scoring; the most generally used are the Facial Action Coding System (FACS) and the Maximally Discriminative Affect Coding System (MAX) (Ekman, 1985).

Research

Facial expression has been suggested for presence research by Huang et al., but not yet been used.

Sources

Description

Prothero and Parker (2003) hypothesize that presence is “an illusion of position and orientation”. According to the rest frame hypothesis, a particular reference frame, the “rest frame”, is selected as the comparator for spatial judgments. The sense of presence in an environment is thought to reflect the degree to which that environment influences the selected rest frame. Presence can accordingly be measured as the degree to which virtual cues overwhelm real cues.In “nulling”, this approach is used by asking subjects to determine the point at which 2 stimuli counterbalance each other.

Research

Participants (n=12, within-subjects design) wore a HMD providing visual self-motion while sitting in a rotating chair providing conflicting inertial self-motion cues. In each of 2 sessions, participants saw 2 visual conditions consisting of meaningful vs. non-meaningful (random pixels) material. Presence was measured by a visual-inertial crossover measure and a subjective presence rating (1 item, 7-point scale). The crossover measure was based on participants signalling their perception of the left/right extremes of chair motion by switching a toggle.
As predicted, inertial crossover velocity was higher for the meaningful visual condition. For subjective presence, the difference was in the same direction but not significant. A very small correlation (r=0.06) was found between the measures. A larger correlation, approaching significance (r=.38, p<.07) was found between the differences between visual conditions for both measures.

Sensitivity: measure distinguished between different conditions
Reliability: not reported
Validity: moderate correlation with subjective presence measure

Sources

Description

The behavioural realism approach, introduced by Freeman et al (2000), is based on the principle that the more similar a display becomes to the environment it represents, the more observers will respond in the same way that they would respond to the environment itself.
Postural responses are proposed as a presence measure. One advatage is that they are not mediated by high level cognitive processes, so unlikely to affect concurrent subjective evaluation. A second advantage is the capacity to produce differential levels of response.
Postural responses occur under the illusion of observer motion, or vection. It has been argued that measures of vection and presence should be related (Freeman et al, 2000; Ohmi, 1998); if a user experiences vection in an environment, s/he is more likely to feel present in that environment.

Research

Hoshino, Takahashi, Oyamada, Ohmi, & Yoshizawa (1997) measured body sway and subjective physiological discomfort while viewing 3D video images of a boat with a rolling background (n=8, within-subjects design). Independent variables were rolling frequency (5 levels) and the system (HMD, 70 inch 3D display, TV). Body sway was measured using the “Quick Mag” motion analysing system, which takes videos of a marker put on the head of the subject. Results showed that body sway was highest for HMD, lower for 3D display and lowest for TV. The same pattern was found for the physiological discomfort scores. In this same experiment, but reported in a later publication (Ohmi, 1998), body sway was measured while participants watched real-world non-stereo video clips taken from a moving car or train. Body sway was found to be proportional and in the same direction as centrifugal acceleration induced by the stimulus material.

Freeman et al. (2000) investigated the effects of stimuli displaying left and right turns in rapid forward motion. In their study (n=24, within-subjects design), participants viewed videos either taken from inside a racing car traversing a bendy track (moving stimulus) or taken from the side (still stimulus). Stimuli were presented both monoscopically and stereoscopically. A Flock of Birds magnetic position tracker placed at the back of the neck was used to measure participants’ x, y, and z positions. After each stimulus, participants rated their presence, involvement, vection, and simulator sickness on continuous scales. Results showed a strong correlation between postural movements in both viewing conditions; observers moved in the same direction as the car. There was significantly more movement for moving than for still stimuli, in both viewing conditions. The increase in motion was larger for stereoscopic than for monoscopic viewing conditions, an effect that just failed to reach significance. Subjective presence, involvement, and simulator sickness ratings were higher for stereoscopic presentation and moving stimuli. Subjective vection ratings were higher for moving stimuli, but there was no effect of stereoscopic presentation. No correlation was found between postural movement and subjective presence ratings.

A replication of this study was conducted by IJsselsteijn, De Ridder, Freeman, Avons, and Bouwhuis (2001) using a larger display. By comparing the results to those of Freeman et al. (2000), the effect of screen size was studied. Effects of stereoscopic presentation and motion were comparable to the previous study. Results showed a significant effect of screen size on subjective sense of presence, but only for the moving stimulus. No significant difference in postural responses was found between Freeman et al. (2000) and IJsselsteijn et al. (2001).

Sources

Description

Slater, Usoh, and Chrysanthou (1995) have used a presence measure based on conflicting real and virtual cues. Participants were asked to point towards a radio, which was present both in the real and the virtual world. The position of the real radio was changed after participants entered the VE, so the sound came from a different location. The idea is that a high degree of presence will lead participants to point towards the virtual rather than the real radio.

Research

In an experiment investigating the influence of shadows on sense of presence in a VE, participants (n=8, within-subjects design) were asked to select the spear closest to the wall out of five. Each participant took part in 5 trials, in 1-4 (varying per participant) of which shadows were displayed. Measures included subjective presence, measured by 6 items, and the objective pointing measure. Results showed a positive relation between the number of trials with shadows and both subjective and objective presence measures (but only for visually dominant participants). Objective and subjective presence were significantly correlated.

Sensitivity: correlation between condition (presence of shadows) and the pointing measure
Reliability: not reported
Validity: significant correlation between the pointing measure and a subjective presence measure (SUS questionnaire)

Primary Source

Description

Held & Durlach (1991) first proposed to use reflex responses as a measure for presence. Loomis (1992) suggests reflex responses as a method for discriminating between presence (the experience of being in a VE) and distal attribution (the experience of being in touch with a VE). Loomis argues that the observer will only show reflex responses to stimuli of s/he experiences them as “real”.

Research

Description

Sheridan (1992) first suggested using socially conditioned responses to virtual social encounters, such as grasping for an object that is handed over, shaking hands, or utterances, as indicators of social presence. IJsselsteijn, De Ridder, Freeman, and Avons, (2000) also propose a broad range of social behaviours such as facial expressions, gestures, body and head movements, eye contact, vocal cues, turn-taking behaviour, use of space, and verbal expressions.

Research

Bailenson (2001, 2003) and his colleagues carried out several experiments about interpersonal distance in virtual environments. In one experiment, (study was previously described in section Bailenson et al. Questionnaire) participants were immersed in a virtual room in which a virtual male agent stood. In each trial they were asked to walk up to the agent and remember certain features and labels on the front and back of the agent’s shirt. Position and orientation of participants were tracked. Photographic realism of the agent’s face and the degree of mutual gaze between agent and participant were varied. Results show that participants maintained more space around agents than around nonhuman-like objects. Female participants maintained more interpersonal distance between themselves and agents who engaged them in eye contact. There was no effect of realism.
Two later experiments (n=80, between-subjects design) replicated and extended these results. Some participants were led to believe that the agents were avatars, controlled by real people. It was found that participants maintained personal space bubbles around virtual humans that were similar in size and shape to bubbles maintained about real humans. Participants gave an avatar more personal space than an agent, unless the agent displayed realistic gaze behaviour. Participants showed more avoidance of approaching virtual humans when they thought they were controlled by a computer.

Sensitivity: measure discriminated between different conditions (avatar vs. agent, gaze behaviour)
Reliability: not reported
Validity: not reported

Primary Source