Content uploaded by Jane Lessiter
Author content
All content in this area was uploaded by Jane Lessiter on Jan 07, 2014
Content may be subject to copyright.
Jane Lessiter
j.lessiter@gold.ac.uk
Jonathan Freeman
j.freeman@gold.ac.uk
Edmund Keogh
Jules Davidoff
Department of Psychology
Goldsmiths College
University of London
New Cross, London SE14 6NW
United Kingdom
Presence, Vol. 10, No. 3, June 2001, 282–297
©2001 by the Massachusetts Institute of Technology
A Cross-Media Presence
Questionnaire:
The ITC-Sense of Presence Inventory
Abstract
The presence research community would benefit from a reliable and valid cross-
media presence measure that allows results from different laboratories to be com-
pared and a more comprehensive knowledge base to be developed. The ITC-Sense
of Presence Inventory (ITC-SOPI) is a new state questionnaire measure whose de-
velopment has been informed by previous research on the determinants of pres-
ence and current self-report measures. It focuses on users’ experiences of media,
with no reference to objective system parameters. More than 600 people com-
pleted the ITC-SOPI following an experience with one of a range of noninteractive
and interactive media. Exploratory analysis (principal axis factoring) revealed four
factors: Sense of Physical Space, Engagement, Ecological Validity, and Negative Ef-
fects. Relations between the factors and the consistency of the factor structure with
others reported in the literature are discussed. Preliminary analyses described here
demonstrate that the ITC-SOPI is reliable and valid, but more rigorous testing of its
psychometric properties and applicability to interactive virtual environments is re-
quired. Subject to satisfactory confirmatory analyses, the ITC-SOPI will offer re-
searchers using a range of media systems a tool with which to measure four facets
of a media experience that are putatively related to presence.
1 Introduction
1.1 Defining Presence
Presence has been used as a global experiential quality metric to evaluate,
develop, and optimize both advanced broadcast and virtual environment (VE)
media systems (Freeman & Avons, 2000; Freeman, Avons, Pearson, & IJssel-
steijn, 1999; Slater, Usoh, & Steed, 1994). Presence is generally defined as a
user’s subjective sensation of “being there” in a scene depicted by a medium
(Barfield, Zeltzer, Sheridan, & Slater, 1995). It has also been defined as “a
perceptual illusion of non-mediation” (Lombard & Ditton, 1997), a definition
that is consistent with the former one as it implies that a user incorrectly per-
ceives a mediated scene to be unmediated. Further, Slater and Usoh (1994)
described presence as “the (suspension of dis-) belief” of being located in a
world other than the physical one.
In addition to these specific definitions, presence has been variously de-
scribed as a “mental manifestation” (Sheridan, 1992a), a (general) “existential
phenomenon” (Draper, Kaber, & Usher, 1998), and a “cognitive state” con-
sistent with a sense of “being there” in an environment, a state that results
282 PRESENCE: VOLUME 10, NUMBER 3
from attending to and evaluating incoming sensory in-
formation (Barfield et al., 1995).
Others have used the terms telepresence (Barfield et
al., 1995; Sheridan, 1992a; Welch, Blackmon, Liu,
Mellers, & Stark, 1996) and virtual presence (Barfield et
al., 1995) to refer to presence in relation to specific
technologies. Draper et al. (1998) define telepresence as
“the perception of presence within a physically remote or
simulated site,” which suggests, de facto, that presence
is a valid construct in relation to experience of the real
(physical) world. This is a contentious issue, a full dis-
cussion of which is beyond the scope of this paper.
However, it is our view that presence is a more useful
concept when it is limited to the study of users’ experi-
ences of mediated presentations. Real-world experience
can be adequately described in terms of more traditional
psychological constructs: such as attention, involve-
ment, and arousal, to name but a few. There seems little
to gain from describing people’s everyday experience in
terms of presence. In the physical world, “being there”
is an invariant, and any variations in sensations of “being
there” are rarely contemplated. Consistent with the def-
initions given above, the reason that presence is relevant
to understanding users’ experiences of media is that an
illusion is generated whereby a user senses that she/he
is located somewhere other than her/his physical envi-
ronment. This illusion can be entertaining (Lodge,
1999), but in addition it might affect users’ perfor-
mance within mediated environments (Welch, 1999)
and support effective therapeutic applications (North,
North, & Coble, 1997). It remains though, at all times,
an illusion. At the very least, though, real-world experi-
ence is useful to presence research insofar as it serves as
a benchmark, or standard, against which to subjectively
judge levels of presence in mediated environments.
Whether there are qualitative differences in the sensa-
tions supported by different media systems is an open
question. However, in a focus group study designed to
investigate viewers’ experiences of stereoscopic televi-
sion, Freeman and Avons (2000) found that observers
used terms commonly associated with virtual environ-
ment display systems, such as “being there” (cf. Usoh et
al., 1999). This suggests that presence may be evoked
by a range of media to varying degrees. As presence is
multiply determined, it is likely that tradeoffs exist be-
tween the different determinants. For instance, displays
that support a high degree of photorealism (such as cin-
ema or high-definition TV) may compensate for an ab-
sence of control and manipulation input devices (hence-
forth, referred to as interactive displays) typically
supported in VEs.
Through a variety of self-report measures, evidence in
support of a range of determinants of presence has been
generated. To more easily compare the relative impact
on presence of its determinants, a general (cross-media),
valid, and reliable measure of presence is desirable. This
paper describes the development of the ITC-Sense of
Presence Inventory (ITC-SOPI), which we propose as
one such measure.
1.2 Determinants of Presence
It is beyond the scope of this paper to provide a
thorough review of all the research that has been com-
pleted into the determinants of presence. For excellent
reviews, the reader is referred to Barfield et al. (1995)
and Draper et al. (1998). Two explanatory approaches
to presence have been used to date: those based on
properties of the technology, and those that compare
and contrast other psychological experiences with pres-
ence (Draper et al., 1998). Partly as a result of these
divergent approaches, the literature has on occasion
confused determinants and correlates of presence. An
additional explanation for this confusion is the genuine
debate as to whether some factors (for example, interest
and engagement) are determinants or correlates of pres-
ence. These issues will not be resolved here. For the
purposes of this paper, we will identify content areas
relevant to the development of the ITC-Sense of Pres-
ence Inventory.
Two general categories of variables can determine a
user’s presence: media characteristics and user character-
istics. This differentiation shares parallels with that of
Slater and colleagues (Slater, Steed, McCarthy, & Mar-
ingelli, 1998; Slater & Usoh, 1994; Slater & Wilbur,
1997) who described “external” (objective) and “inter-
nal” (subjective) determinants of presence. The media
characteristics category can be subdivided into aspects of
Lessiter et al. 283
media form and media content (for example, IJssel-
steijn, de Ridder, Freeman, & Avons, 2000). It is likely
that these categories will interact in determining pres-
ence.
Media form refers to physical, objective properties of
a display medium. Sheridan (1992a) proposed three
major elements of media form that may determine pres-
ence: the extent of sensory information presented, the
degree of control a participant has over positioning his/
her sensors within the environment (such as a turn of
the head to see or hear more of the environment), and a
user’s ability to modify aspects of the environment. An
ideal teleoperation system could support each of these
elements sufficiently to make the mediated presentation
indistinguishable from the physical world, and Sheri-
dan’s theory would predict that such a system would
elicit high presence. By this definition, media form cov-
ers the means by which an image is represented (for ex-
ample, photorealistic video or animated computer
graphics).
We use the media content category to refer to the
overall theme, narrative, or story depicted via a display
system. In experiments conducted to date, this has
ranged from a virtual cliff scenario (Usoh et al., 1999)
to a fast-paced rally car drive (Freeman, Avons, Meddis,
Pearson, & IJsselsteijn, 2000; Freeman & Avons,
2000). Clearly, some aspects of a narrative might
change as a result of variation in media form variables.
For example, in an interactive VE, a user can select the
direction in which to navigate; hence, the order of
events within the VE may vary. However, the overall
theme—what we refer to as media content—remains
unaffected.
Social elements of a displayed environment, such as
acknowledgement of the user through the reactions of
other actors, virtual or real (Heeter, 1992), also contrib-
ute to determining presence (sometimes referred to as
social presence). The reader is referred to Lombard and
Ditton (1997) for a comprehensive review of research
into social aspects of presence. Different display config-
urations are likely to be required to optimize media sys-
tems for social presence rather than (spatial) presence.
Some media systems, such as the immersive teleconfer-
encing system being developed by the European Com-
mission’s IST VIRTUE project (2000), aim to support
both high presence and high social presence.
There is some evidence that user characteristics, such
as a user’s perceptual, cognitive and motor abilities, and
personality traits (such as a willingness to suspend disbe-
lief (Slater & Usoh, 1994)), can be important in deter-
mining presence (Witmer & Singer, 1998). Relevant
individual characteristics are likely to vary with the age
and possibly with the sex of the user. Further, as pres-
ence is a transient experiential state, like mood (Sheri-
dan, 1992a, 1992b), it is susceptible to variation within
the same person, given the same physical conditions on
two separate occasions.
Slater and colleagues have reported preliminary inves-
tigations into the relationships between individual dif-
ferences and presence (Slater et al., 1998; Slater &
Usoh, 1994; Slater & Wilbur, 1997). Using both sub-
jective and objective methodologies, Slater and col-
leagues have demonstrated that users whose primary
representation systems are visual are more likely to expe-
rience presence in a visual virtual environment than us-
ers whose primary representation systems are auditory
or kinaesthetic.
Witmer and Singer (1998) made an important contri-
bution to the measurement of individual differences in
presence. Their Immersive Tendencies Questionnaire
contains 29 items of which 16 relate to one of three
clusters (involvement, focus, and games) that tap both
state (for example, “How mentally alert do you feel at
the present time?”) and trait (“Have you ever remained
apprehensive or fearful long after watching a scary mov-
ie?”) individual differences that putatively moderate the
tendency to feel present.
Attention and involvement are user responses associ-
ated with presence and are likely to be affected by all
three of the categories of determinants previously out-
lined: media form, media content, and user characteris-
tics. Involvement and attention are particularly impor-
tant in the measurement of presence (Barfield &
Weghorst, 1993; Witmer & Singer, 1998) as partici-
pants exposed to a media system and unfamiliar with the
presence concept are likely to use these terms to de-
scribe their experience. For instance, Freeman and
Avons (2000) found that participants described their
284 PRESENCE: VOLUME 10, NUMBER 3
involvement in a 3-D video presentation as “unavoid-
able”. The term flow has been used to describe a state of
augmented concentration, in which the user is unaware
of external distractors, the placement of self in the real
world, and even real time (Draper et al., 1998; Fon-
taine, 1992).
To summarize, several empirical studies have demon-
strated that presence tends to increase as the fidelity of a
reproduction or simulation of the physical world in-
creases.
High-fidelity representations of environments are per-
ceived as being more natural and real—having depth,
space, and continuity—and, consequently, as less dis-
tracting. The naturalness of a visual representation (IJs-
selsteijn, de Ridder, Hamberg, Bouwhuis, & Freeman,
1998) and of a user’s interaction with a media system
(Hendrix & Barfield, 1996a) have both been shown to
correlate with presence. When more attention is allo-
cated to the mediated environment than to the environ-
ment in which an individual is physically located, the
increased mediated sensory input and decreased nonme-
diated sensory input give rise to sensations of being lo-
cated within the mediated environment (Kim & Biocca,
1997). Presence involves the user feeling as though they
are spatially located within an environment portrayed by
a display system, and it is characterized by a sensation of
a strong perception-action link between the display and
the user. Which of the categories of determinants of
presence previously reviewed is most important is a
question that cannot be answered at present due to the
range of different presence measures employed.
1.3 Subjective Measures of Presence
Sheridan (1992a) reasons that, because presence is
a “mental manifestation,” its fundamental measurement
is self-report. Presence has been measured using simple
rating scales (Hendrix & Barfield, 1996b; Slater &
Usoh, 1994; Slater et al., 1994; Welch et al., 1996) that
typically relate to feeling: physically located in a medi-
ated space, that the mediated environment is as real as
the real world, and that the mediated place had been
“visited.” More structured approaches to the measure-
ment of presence that aim to elucidate the presence
construct (and to provide a more reliable and valid in-
dex of the dimension(s) related to presence) have also
been followed. Table 1 summarizes a range of existing
subjective presence measures.
An ideal presence questionnaire that could be used to
evaluate a range of media systems must satisfy a number
of important criteria. First, an understanding of pres-
ence should not be assumed by directly asking respon-
dents how present they feel. Presence is a relatively un-
familiar construct to most nonexperts (Freeman &
Avons, 2000; Freeman et al., 2000). Not understanding
the construct for which they are providing ratings
would reduce participants’ ability to meaningfully dis-
tinguish points along a presence rating scale (Freeman,
2000).
Second, questions should avoid addressing two issues
in one question; this renders the response options con-
fusing and meaningless. Care should be taken in phras-
ing questions that are answered on a Likert-type scale to
ensure that only one (continuous) construct is being
measured (for example, “How much did you feel as
though the mediated environment was a place that you
visited?”).
Third, the response options used should, ideally, be
consistent across items (within individual question-
naires). This renders the questionnaire more user-
friendly and reduces its completion time.
Fourth, presence is likely to be a multidimensional
construct (Barfield et al., 1995, Witmer & Singer, 1998;
Schubert, Friedmann, & Regenbrecht, 1999). An ideal
presence questionnaire should take account of this pos-
sibility and tap a range of characteristics that are puta-
tively related to presence. Indeed, unidimensional pres-
ence ratings have been shown to be potentially unstable
in that they can be affected by prior experience (Free-
man et al., 1999). When a group of observers was asked
to first rate a training stimulus for how interesting it
was, subsequent presence ratings for a test stimulus
were less sensitive to variations in viewing condition
(the presence or absence of stereoscopic depth cues)
than were those of a group asked to rate the same train-
ing stimulus for 3D-ness (Freeman et al., 1999). This
result was taken to indicate that both depth within the
stimulus and interest in the content contribute to pres-
Lessiter et al. 285
ence, but that the importance of each could be varied
by participants’ experience immediately prior to provid-
ing presence ratings. One potential explanation of the
problems of stability and bias associated with simple
presence rating scales is that they treat presence as uni-
dimensional when it is in fact multidimensional. Thus, a
measure that takes account of the potential multidimen-
sional structure of presence may prove to be more ro-
bust.
Fifth, qualitative reports from experimental partici-
Table 1. Summary of Subjective Measures of Presence
N.
items
N.
participants Sample Analysis Dimensions Labels
Slater et al.
(1994)
3 N/A N/A N/A N/A Presence
Slater & Usoh
(1994)
1 N/A N/A N/A N/A Presence
Hendrix & Barfield
(1996b)
3 N/A N/A N/A N/A Presence
Welch et al.
(1996)
2 N/A N/A N/A N/A Presence
Kim & Biocca
(1997)
8 96 Users exposed to
video stimuli
presented on
either 9, 20, or
32 in. TV
Factor analysis 2 Arrival
Departure
Witmer & Singer
(1998)
32 152 VE users Cluster analysis 3 Involved/Control
Natural
Interface Quality
Barfield, Baird, &
Bjorneseth
(1998)
18 Not stated Not stated Factor analysis and
cluster analysis
Not stated (subset relate to
Presence)
Schubert, Friedmann,
& Regenbrecht
(1999)
75 246 90% male; mostly
3-D computer
games
players
Principal
components
analysis (oblique)
8 Spatial Presence*
Quality of Immersion
Involvement*
Drama
Interface Awareness
Exploration of VEs
Predictability and
Interaction
Realness
(believability)*
* Related to “presence” factor following second-order analysis.
286 PRESENCE: VOLUME 10, NUMBER 3
pants suggest that there are distinct similarities in
people’s experiences of, and sensations associated
with, “being there” elicited by different types of me-
diated environment (Freeman & Avons, 2000; c.f.
Usoh et al., 1999). Hence, a general presence mea-
sure that focuses entirely on users’ experiences with
media presentations should be possible. Thus, to en-
sure generalizability to other media systems and to
avoid confounding separate user and system contribu-
tions to presence (Slater, 1999), the questions should
not make reference to specific media system and con-
tent properties. For example, references to input and
interaction devices mean a questionnaire cannot be
applied to the evaluation of presence through nonin-
teractive media. Furthermore, depending on their
phrasing, questions relating to specific technical de-
vices may not measure presence at all (Slater, 1999).
The sense of interacting with a mediated environment
can be probed without making direct reference to
physical properties of the system. Because a user’s
perception of, and ease of interaction with, an inter-
face are important issues to consider when designing
media systems, there is a good rationale for measur-
ing them with supplementary questions/question-
naires.
Sixth and relatedly, a general presence measure
should be piloted on participants of a range of media
systems displaying a range of contents, that is, systems
with variation in media form and media content.
Finally, standard data-reduction techniques, such as
factor analysis, used in the development of question-
naires assume a high respondent-to-item ratio (Kline,
1993). Violation of this criterion reduces the reliability
of the correlation coefficients that are used to compute
the factors (Tabachnick & Fidell, 1989). Comrey (1973
in Tabachnick & Fidell) suggests that a sample size of
50 is very poor, 100 is poor, 200 is fair, 300 is good,
500 is very good, and 1,000 is excellent. However, as a
general index, Tabachnick and Fidell suggest that it is
“comforting to have at least five cases for each observed
variable” (p. 603).
Our research group has developed a new cross-media
presence questionnaire to measure media users’ state of
presence, which aims to have broad application and ad-
heres to the criteria above. In section 2, we describe the
development of the ITC-Sense of Presence Inventory
(ITC-SOPI). We then present the results of an explor-
atory factor analysis, and compare these results to those
of other studies that have investigated the structure and
measurement of presence.
2 Method
2.1 ITC-SOPI Development
Sixty-three items were initially generated that
tapped possible manifestations of different content areas
deemed relevant to presence on the basis of theoretical
and empirical papers: sense of space, involvement, atten-
tion, distraction, control and manipulation (autonomy),
realness, naturalness, perception of time, awareness of
behavioral responses, a sense of social interaction
(parasocial and copresence), personal relevance, arousal,
and negative effects.
Any questionnaire should be as respondent-friendly as
possible and have face validity; that is, it should look like
it is measuring what it purports to measure (Oppen-
heim, 1992; Rust & Golombok, 1989). A five-point
Likert scale (1 ⫽strongly disagree; 5 ⫽strongly agree)
was chosen as the response option for all items. This
consistency makes it easier for respondents to complete
the questionnaire and for responses to be scored. Items
were phrased carefully, simply, and unambiguously, and
were numbered, as this is a familiar questionnaire style
and recommended in psychometric texts (Rust &
Golombok, 1989). The use of double-negatives was
avoided, and each item dealt with just one issue. Social
desirability was discouraged by careful phrasing, by in-
structions to avoid spending too long on any one ques-
tion, and by emphasizing that the first response is usu-
ally the best. The questionnaire was presented in two
parts, A (7 items) and B (56 items), respectively relating
to respondents’ experiences before and during the me-
diated environment. Comprehensive instructions ex-
plained the terms used, the layout, and how to select a
response, and provided an assurance of confidentiality to
increase compliance and honesty. A number of poten-
tially important background variables were also col-
Lessiter et al. 287
lected. These included basic demographics (such as age,
sex, occupation), education, self-rated experience with
computers and frequency of computer games playing,
and self-rated knowledge of VR, TV/film production,
and stereoscopic image presentation using polarized
spectacles. This section requires some modification for
use in non-UK labs, because educational status is tai-
lored to UK standards. Alternatively, the background
information can be replaced with background variables
that are of interest to a particular lab.
2.2 Samples and Procedure
The ITC-SOPI was administered to a total of 604
people following an experience with a mediated envi-
ronment. The sample was divided into six subsamples,
each of which experienced a different level of physical
immersion (for example, 3-D versus 2-D, ability to con-
trol aspects of the environment versus no control, large
screen versus small screen, surround sound versus ste-
reo) and different display contents.
Respondents were recruited from an IMAX cinema,
presenting 2-D and 3-D (viewed using polarized
glasses) films on a 15m ⫻21m screen, n⫽22 and 196,
respectively; a well-known cinema chain in the United
Kingdom, presenting 2-D films on general release, n⫽
238; Goldsmiths College Students’ Union (GCSU) film
night, where VHS films are projected onto a 2m ⫻
1.5m (approx.) display screen with fairly degraded au-
dio/visual fidelity, n⫽26; and two experimental set-
tings based in the Psychology Department at Gold-
smiths College. One of these consisted of participants
viewing a video short (⬍30 min.) on a 28 in. color TV
(n⫽73) before completing the ITC-SOPI. The other
experimental setting entailed participants playing a com-
puter-generated racing game on a consumer videogame
console (n⫽49). The console was connected to the
Platform for Immersive Television (PiT, a controlled
test environment with a 28 in. color TV and set to
mono audio output: see figure 1). Participants navigated
the car around the track using an unsophisticated game-
pad. An incentive was offered for questionnaire comple-
tion (free prize draw entry, free “two tickets for the
price of one” vouchers, cash payment of £3, or course
credit (psychology undergraduates only)).
The content was varied to avoid confounding the re-
sults with a specific type of content. In this way, the sta-
bility of the relationships between questions, irrespective
of content, could be examined. All respondents rated
just one presentation. There were two presentations for
IMAX 3-D, both fictitious; one for IMAX 2-D (a docu-
mentary film); eleven different films for the cinema con-
dition (including one animated film); two films for
GCSU; six video shorts, comprising two animated stim-
uli, football highlights, a black-and-white French film
(no subtitles and eclectic editing), and two crime recon-
structions from a popular British television program;
and one stimulus for the computer games sample (F1
Racing). Although the subsamples had a broad range of
media content in terms of different themes and stories,
in terms of media form, noninteractive, photorealistic
displays were heavily weighted in the sample. These are
potential shortcomings that we address in the discus-
sion.
Fifty-one percent of the sample were male, and re-
spondents were aged between 9 and 73 years (mean age
of 29 years; s.d. ⫽11.24).
Figure 1. Venue example: The Platform for Immersive Television
(PiT) for computer games console respondents.
288 PRESENCE: VOLUME 10, NUMBER 3
3 Results
3.1 Principal Axis Factoring
Data for the 63 items were entered into a principal
axis factoring (PAF) analysis. (For an introduction to
factor analysis, see Tabachnick and Fidell (1989, chapter
12)). In brief, factor analysis is a data-reduction tech-
nique that summarizes patterns of correlations by re-
vealing groups of correlated items, which are called fac-
tors. Each factor accounts for a proportion of the
variance across all items in the data set. This proportion
is represented by a numerical term, an eigenvalue. The
first factor will always account for the largest proportion
of variance in a data set (that is, it will have the largest
eigenvalue). Eigenvalues decrease with each successive
factor. Factor loadings represent correlations between
each item and a factor. The magnitude of the correla-
tions therefore indicate the weightings of variables on
each factor. In a good factor analysis, a variable will load
highly on just one factor, although variables will some-
times cross-load. Ideally, a minimum of three variables
will mark a factor.
Pairwise deletion for cases with missing values was
used to make full use of the data set. This was repeated
with the more conservative listwise deletion. Both pro-
duced broadly similar structures and factor loadings.
Preliminary checks (such as the Kaiser-Meyer-Olkin
measure of sampling adequacy) indicated that the data
were suitable for PAF. Furthermore, examination of the
intercorrelation matrix for all variables revealed a large
number of significant correlations (pⱕ0.001) indicat-
ing good factorability of the data set.
A number of criteria were employed to determine the
number of factors to extract. First, factors with initial
eigenvalues greater than 1 were considered. According
to Tabachnick and Fidell (1989), this figure is “usually
somewhere between the number of variables divided by
3 and the number of variables divided by 5” (p. 635).
For 63 items, somewhere between 13 and 21 would
therefore be anticipated. In fact, twelve factors satisfied
this criterion (16.47, 4.81, 2.61, 2.55, 2.00, 1.84, 1.48,
1.36, 1.27, 1.18, 1.06, 1.05). With 40 or fewer vari-
ables and a large sample size, this criterion is usually a
good index, but, in other circumstances, the criterion of
eigenvalues greater than 1 can over- or underestimate
the number of factors (Tabachnick & Fidell, 1989). In
this instance, with 63 items, the number of factors was
likely to be overestimated.
A second criterion, of examining “elbows” in the
scree plot (a plot of the eigenvalues), suggested either a
two-, three-, four-, or five-factor solution. A third crite-
rion was based on comparing the number and spread of
“marker” variables for each factor solution. (See table
2.) We use two definitions of marker variables. First,
they are defined as any variable with a factor loading
above 0.3 on any factor in the rotated factor matrix
(Watson et al., 1995). A more stringent definition is
that a variable’s primary loading must be 0.2 greater
than any cross-loading (Bedford, 1997), in addition to
having a primary loading above 0.3. When comparing
different factor solutions, the greater the number of
marker variables per factor coupled with a relatively even
distribution of markers that meet the more stringent
criterion indicates a better solution for the data. Note,
however, that fewer marker variables are anticipated
with each successive factor that is extracted in any solu-
Table 2. Marker Variables on Each Factor for the Five
Different Factor Structures
Number of
factors in
solution 1 2 3 4 5
150
2 47 (30) 26 (10)
3 36 (17) 29 (8) 17 (4)
4 28 (14) 24 (13) 16 (6) 8 (6)
5 27 (12) 24 (12) 13 (4) 7 (6) 6 (0)
Values are based on each of two criteria: (i) Watson et al.’s
(1995) criterion of variables that load higher than 0.3 in
the rotated factor matrix, and values in parentheses are
based on (ii) Bedford’s (1997) criterion of variables that
load higher than 0.3 and where the primary loading is
more than 0.2 greater than any cross-loading.
No value is available for the one-factor solution using
criterion (ii) because there was no other factor with which
to compare loadings.
Lessiter et al. 289
tion. The five-factor solution was rejected because, us-
ing Bedford’s more stringent criterion, there were no
marker variables on the fifth factor, as shown in table 2.
Although the two-, three-, and four-factor solutions
each fared reasonably well on the more stringent crite-
rion, the four-factor solution made the most conceptual
sense. Cumulatively, the four factors accounted for
38.3% of the variance in the data.
The four factors were subjected to varimax rotation,
which aims to achieve “simple structure” (Thurstone,
1947). Simple factor structures are easier to interpret as
they avoid ambiguities associated with correlated factors. A
similar solution to that we report here was obtained using
oblique rotation, but, for our preliminary data reduction
purposes, we report the varimax solution.
3.1.1 Factor Structure. The first factor ac-
counted for 14.2% of the variance and was defined by
high loadings from items such as “I had a sense of being
in the scenes displayed,” “I felt I was visiting the places
in the displayed environment,” and “I felt that the char-
acters and/or objects could almost touch me.” These
items indicate a sense of physical placement in the medi-
ated environment, and interaction with and control over
parts of the mediated environment. This factor seems to
encapsulate the definitions of presence discussed in sec-
tion 1—a sense of being there—and was consequently
labeled “Sense of Physical Space.”
The second factor explained 11.1% of the variance
and had high loadings from items including “I felt in-
volved (in the displayed environment),” “I enjoyed my-
self,” and “My experience was intense.” These items
suggest a tendency to feel psychologically involved and
to enjoy the content. This factor was consequently la-
beled “Engagement.”
The third factor explained 7.6% of the variance and
was characterized by questions including “The content
seemed believable to me,” “The displayed environment
seemed natural,” and “I had a strong sense that the
characters and objects were solid.” These variables indi-
cate a tendency to perceive the mediated environment
as lifelike and real, and the factor was subsequently la-
beled “Ecological Validity.”
The fourth and final factor explained 5.4% of the vari-
ance and was almost exclusively characterized by the
variables describing adverse physiological reactions such
as “I felt dizzy,” “I felt nauseous,” “I felt I had a head-
ache,” and “I had eyestrain.” This factor was labeled
“Negative Effects.”
Although all items/questions were generated or se-
lected from previous presence measures for their puta-
tive ability to tap elements of presence, eight questions
failed to load significantly (ⱖ0.30) on any of the factors.
(See table 3.) These were discarded for the revised ver-
sion of the ITC-SOPI.
Table 3. Items that Failed to Load Significantly (ⱖ0.3) on any of the Four Factors which were Deleted from the Revised ITC-SOPI
ITC-SOPI item
Loading
F1 F2 F3 F4
I was aware of the real world ⫺0.23 ⫺0.15 ⫺0.03 ⫺0.11
I wanted to see more of the space in the displayed
environment than I was able to
0.18 0.23 0.07 0.19
I found it easy to forget that I was watching a display 0.19 0.23 0.25 0.19
I had the best viewpoints 0.20 0.19 0.19 ⫺0.05
The temperature of the real world distracted me 0.05 0.03 ⫺0.12 0.26
I was distracted by the quality of the technology 0.15 0.07 ⫺0.17 0.29
I wanted to make specific sounds louder or softer 0.20 0.11 ⫺0.05 0.25
I felt I knew what was going to happen next 0.02 ⫺0.02 ⫺0.11 0.03
290 PRESENCE: VOLUME 10, NUMBER 3
3.2 Reliability
First, to examine the stability of the items on each
of the four factors, the data set was randomly divided
into two subsamples (group 1: n⫽325; group 2: n⫽
279), while ensuring that each of the viewing condi-
tions (IMAX 2-D, 3-D, Cinema GCSU film night,
video shorts, and computer game) was adequately repre-
sented within each subsample. Most of the items
showed the same loadings as those obtained in the large
sample. In an attempt to further reduce the number of
items in the ITC-SOPI, six items that were conceptually
or statistically inconsistent were excluded from the re-
vised questionnaire.
Second, internal reliability coefficients (Cronbach’s
alpha) were computed for each of the four factors.
Items that failed to load above 0.3, or that were deemed
conceptually or statistically inconsistent following the
structure reliability check described above, were ex-
cluded. Alphas were very good: Sense of Physical Space
(21 items) ⫽0.94; Engagement (13 items) ⫽0.89;
Ecological Validity (6 items) ⫽0.76; Negative Effects
(8 items) ⫽0.77. Items that reduced the alpha coeffi-
cient (two for Sense of Physical Space, one for Ecologi-
cal Validity, and two for Negative Effects) were ex-
cluded from the revised ITC-SOPI.
3.3 Validity: Media Format
Preliminary validation of the ITC-SOPI was per-
formed by comparing mean scale scores computed for
each factor across the different media formats (samples).
The results were very positive particularly for the first
factor given that several important issues, such as con-
tent, were not controlled. (See figure 2a–d.) Clearly, a
more stringent validity test would compare matched
content across the different media formats.
Sense of Physical Space showed sensitivity to media
form: IMAX 3-D had the highest score, followed by
interactive computer game, IMAX 2-D, cinema, GCSU,
and video shorts. (See figure 2a.) A one-way analysis of
variance (ANOVA) revealed a significant difference in
Sense of Physical Space scores across the six media sam-
ples (F(5,603) ⫽75.77; p⬍0.001). Post-hoc tests
(Bonferroni) indicated that ratings for Sense of Physical
Space for all the media formats were significantly differ-
ent from one another (p⬍0.05) with the exceptions of
IMAX 2-D versus computer games, video shorts versus
GCSU, and cinema versus GCSU.
Engagement also demonstrated some sensitivity to
media sample: mean scores were highest for IMAX
2-D, followed by IMAX 3-D, computer game,
GCSU, cinema, and video shorts. A one-way ANOVA
revealed a significant difference between the six media
samples (F(5,603) ⫽20.81; p⬍0.001). Post-hoc
tests (Bonferroni) indicated that ratings for Engage-
ment were significantly different (p⬍0.05) between
video shorts and every other condition, between
IMAX 2-D versus cinema, and IMAX 3-D versus cin-
ema. (See figure 2b.)
For Ecological Validity, mean scores were highest for
IMAX 2-D, followed by IMAX 3-D, video shorts, cin-
ema, GCSU, and computer game. A one-way ANOVA
showed a significant difference between the six media
samples (F(5,603) ⫽23.61; p⬍0.001). Post-hoc tests
(Bonferroni) indicated that ratings for Ecological Valid-
ity for all the media formats were significantly different
from one another (p⬍0.05) with the exceptions of
IMAX 3-D versus video shorts, cinema versus computer
game, cinema versus GCSU, and computer game versus
GCSU. (See figure 2c.)
Finally, for Negative Effects, IMAX 2-D had the
highest mean score, followed by IMAX 3-D, computer
game, video shorts, cinema, and GCSU. A one-way
ANOVA revealed a significant difference between the
six conditions (F(5,603) ⫽12.03; p⬍0.001). Specifi-
cally, post-hoc tests (Bonferroni) revealed significant
differences (p⬍0.05) between IMAX 2-D versus cin-
ema, IMAX 2-D versus GCSU, IMAX 3-D versus video
shorts, IMAX 3-D versus cinema, IMAX 3-D versus
GCSU; all other pairs were not significant. (See figure
2d.)
3.4 Revised ITC-SOPI
A total of 44 items were retained following the
analyses detailed above. Items were dropped that either:
(i) failed to load above 0.3 on any factor (eight items);
Lessiter et al. 291
(ii) showed unstable factor loadings when the sample
was randomly split and/or were deemed of low impor-
tance or relevance to their factor (six items), or (iii) ele-
vated the internal consistency of their respective factors
when deleted in the alpha check (five items).
3.5 Factor Intercorrelations
Scores were produced for each of the four factors
by calculating the mean of their respective items (based
on the revised ITC-SOPI described above, where Sense
of Physical Space ⫽19 items; Engagement ⫽13 items;
Ecological Validity ⫽5 items; and Negative Effects ⫽6
items). Pearson’s correlations were run between all the
resultant scale totals (n⫽604). Each of the first three
“positive presence-related” scales were strongly posi-
tively intercorrelated (Sense of Physical Space and En-
gagement: r⫽0.62; p⬍0.001; Sense of Physical Space
and Ecological Validity: r⫽0.52; p⬍0.001; Engage-
ment and Ecological Validity: r⫽0.45; p⬍0.001).
Negative Effects did not correlate with Engagement
(r⫽⫺0.05; n.s.) or with Ecological Validity (r⫽0.06;
n.s.), but did correlate with Sense of Physical Space (r⫽
0.24; p⬍0.001).
4 Discussion
The factor structure identified here, based on re-
sponses to the ITC-SOPI across a range of predomi-
nantly noninteractive media formats, is similar to that
produced by other researchers who have attempted to
elucidate the structure of presence using more-limited
samples (for example, responses from VE users or 3-D
Figure 2. Mean scores for each of the media samples: (a) Sense of Physical Space; (b) Engagement; (c) Ecological Validity;
and (d) Negative Effects.
292 PRESENCE: VOLUME 10, NUMBER 3
computer games players). Informal reports in both VE
(Usoh et al., 1999) and TV-based studies (Freeman &
Avons, 2000) suggest that there are distinct similarities
in respondents’ experiences of “being there” elicited by
different display media.
Each of the four factors is likely to be determined by
the interaction between media form, media content,
and user characteristic variables. In this regard, it is not
possible to draw firm conclusions based on the present
data set, which intended to include as much variety in
these three broad determinants as possible to enhance
the generalizability of the measure. Further studies are
required that measure the effects of manipulations of
specific determinants of presence on the ITC-SOPI fac-
tors. However, we speculate briefly on possible relations
between determinants of presence and each of the fac-
tors we identified in our current dataset, based on re-
sults from the current data set and the constituent items
of each factor.
The first factor in our data set, Sense of Physical
Space, contains two of the three items used in simple
post-test presence scales that have previously been con-
sidered to tap key elements of presence (such as the
scales of Slater et al., 1994). It is of note that, although
these items had primary loadings on Sense of Physical
Space, they also cross-loaded on Engagement and Eco-
logical Validity. Presence is likely to be related not only
to a user’s sense of being located within a spatially con-
tiguous physical environment but also to his/her per-
sonal evaluation of the appeal and the naturalness/be-
lievability of the content within the displayed
environment.
We propose that the prime determinants of the Sense
of Physical Space factor are, in the terminology we use
in the introduction to this paper, media form variables.
Some support for this assertion is given by the prelimi-
nary validity analysis that found that users of high-fidel-
ity presentations (such as IMAX 3-D) gave higher Sense
of Physical Space ratings than did users of lower-fidelity
presentations (such as GCSU). Interestingly, although
the computer-games console condition consisted of a
relatively small field of view and was low in photoreal-
ism, respondents produced relatively high Sense of
Physical Space ratings. This suggests that the ability to
physically control and manipulate aspects of the dis-
played environment (even using unsophisticated control
devices) enhances the sense of being physically located
in that environment. Media form variables may there-
fore interact in compensatory ways, similar to the iso-
presence equivalence classes suggested by Ellis (1996).
Furthermore, it is of note that these results were ob-
tained when content was varied both within and across
the samples piloted.
The second factor in our data set, Engagement, pro-
vides a measure of a user’s involvement and interest in
the content of the displayed environment, and their
general enjoyment of the media experience. That this
factor emerged indicates that attention and involvement
are important elements of a user’s evaluation of a media
experience (Barfield & Weghorst, 1993; Draper et al.,
1998; Witmer & Singer, 1998). The content of the pre-
sentation inevitably influences ratings of Engagement.
Indeed, one of the ITC-SOPI items directly probes re-
spondents about how appealing they found the content.
Other questions that comprise this scale relate to arousal
and emotionality. These are likely to be influenced by
the media content, but also intensified by the media
form. For instance, a fast-paced rally car scene (with
first-person perspective) is likely to be arousing in itself,
but it is likely to be more arousing when presented on a
large screen with surround sound and motion feedback.
By this rationale, Engagement may be determined both
by media content and media form variables.
Questions that comprise our third factor, Ecological
Validity, relate to the believability and realism of the
content and the naturalness and solidity of the environ-
ment. The number, extent, and consistency of sensory
stimulation (media form variables) are therefore likely to
enhance perceived naturalness and, in turn, increase rat-
ings on this scale. Indeed, this is somewhat supported
by the pattern of scores across the media samples pi-
loted. Again, IMAX presentations received the highest
ratings, in spite of the entirely fictitious contents of
IMAX 3-D. This suggests that the greater the immersive
properties of the media form, the less influence that
content has on the presentation’s believability, realness,
and naturalness. Differences in the media form variable,
photorealism, should also produce differences on this
Lessiter et al. 293
scale. Indeed, the computer game sample received the
lowest ratings compared to the other media samples on
Ecological Validity.
Our fourth factor, Negative Effects, is less related to
our first three factors than they are to each other. Wit-
mer and Singer (1998) report that presence and simula-
tor sickness are inversely related. Whilst in the current
study Negative Effects was not strongly correlated (posi-
tively or negatively) with Engagement or Ecological
Validity, it was significantly but modestly (and posi-
tively) related to Sense of Physical Space. Headache,
eyestrain, tiredness, and other negative effects may be
associated with the media form. For instance, some re-
spondents in the IMAX 3-D sample reported (in addi-
tional comments) that the polarized glasses were un-
comfortable. Indeed, IMAX 3-D respondents gave the
second-highest ratings on this scale. IMAX 2-D respon-
dents, on average, gave the highest Negative Effects
ratings. This sample had no variation in content; all re-
spondents viewed Everest, a documentary following the
ascent of a team of climbers of the world’s highest
mountain, and this may have produced higher than av-
erage negative effects such as dizziness and disorienta-
tion. Content is also likely to affect self-reported nega-
tive effects more generally by interacting with user
characteristics (such as personal content preferences).
For instance, if a given content is perceived as boring,
corresponding ratings of tiredness, or even headaches,
may result.
Schubert et al.’s (1999) three presence-related com-
ponents (“Spatial presence”, “Involvement”, and “Real-
ness [comparability to the reality]”) have striking paral-
lels with the first three factors identified in the present
study (Sense of Physical Space, Engagement, and Eco-
logical Validity). Witmer and Singer’s (1998) first clus-
ter, “Involved/Control,” is comparable to an amalgam-
ation of our Sense of Physical Space and Engagement
factors. Indeed, Sense of Physical Space and Engage-
ment were found to correlate strongly and significantly,
lending support to Witmer and Singer’s (1998) notion
that “Immersion” (Sense of Physical Space) and “In-
volvement” (Engagement) are interdependent. Their
second and third clusters, “Natural” and “Interface
quality” are akin to Ecological Validity as identified
here. Our Negative Effects factor is not represented in
either Witmer and Singer’s or Schubert et al.’s struc-
tures, because they did not include questions relating to
adverse physiological effects of exposure to media.
It is less clear how the factors we have obtained relate
to Kim and Biocca’s (1997) two-factor solution. It is
likely that the limited number of items (eight) that Kim
and Biocca used did not provide a comprehensive cover-
age of all the possible facets of presence that were incor-
porated into the ITC-SOPI.
From a theoretical perspective, our data are consistent
with Slater and Wilbur’s (1997) notions of the influence
of immersion on presence. They argue that high pres-
ence is associated with the extent to which an individual
is provided with an inclusive, extensive, surrounding,
and vivid display. This is not to suggest that our ques-
tionnaire measures objective physical properties of the
technologies used; it explicitly measures subjective expe-
rience. However, our preliminary validation check dem-
onstrates that our first factor, Sense of Physical Space, is
sensitive to variation in objective system parameters. It
is our view that these parameters should be treated as
independent variables where researchers have an interest
in their effects on presence. In addition, Sense of Physi-
cal Space and Engagement correspond closely with two
factors that Witmer and Singer (1998) identify as essen-
tial to experience presence (which they term “Immer-
sion” and “Involvement”).
It is not clear whether the four factors we have identi-
fied all contribute to a sense of presence, and, if they do,
whether this occurs additively or in a more complex
manner. We currently recommend that each ITC-SOPI
scale be analyzed separately, as we expect them to be
differentially sensitive to manipulations of particular de-
terminants of presence.
The preliminary results reported in this paper are en-
couraging but require replication with new samples and
evenly distributed types of content. Clearly, one limita-
tion of the ITC-SOPI is that it has not been piloted on
a broad range of what we term “interactive” displays
(that is, those that include a control/manipulation de-
vice). However, the revised ITC-SOPI has now been
distributed to a number of VE labs across the world.
Confirmatory factor analysis of these new data will en-
294 PRESENCE: VOLUME 10, NUMBER 3
able a test of the stability of both the factor structure
and its constituent items. A full exploration of the inter-
relationships between the factors will then be under-
taken. In addition, guidelines for scoring ITC-SOPI
responses will be developed. A more rigorous validation
of the ITC-SOPI is also in progress, using experimental
designs controlling for factors such as novelty and con-
tent. Finally, corroborative evidence will be sought
through behavioral and physiological measures of pres-
ence collected concurrently with the revised ITC-SOPI.
As we discussed in the introduction to this paper, a
standard cross-media presence measure will constitute a
useful tool for the presence research community. It will
enable the comparison of results both within and across
independent research laboratories. Across labs, varia-
tions in media content will clearly affect presence rat-
ings; there is, after all, likely to be substantial variation
in users’ attitudes toward different contents, in turn,
influencing presence. This is an issue that is likely to
affect any presence measure. Nonetheless, it is reassur-
ing that the first ITC-SOPI factor, Sense of Physical
Space, demonstrated sensitivity to variations in media
form in spite of variation in content. A generic tool will
enable researchers to investigate in more detail the rela-
tive importance of the determinants of presence, includ-
ing aspects of media form, media content, and user
characteristics.
Current questionnaire measures have not been devel-
oped for the evaluation of presence across a range of
media because their piloting procedure has involved
very limited samples. Although the present limitations
of the ITC-SOPI in this regard are acknowledged, it has
been designed for general utility without emphasizing
system features of any particular medium. However, it is
recommended that the ITC-SOPI be used in conjunc-
tion with other questionnaires tailored to lab-specific
requirements. These might range from questions focus-
ing on particular features of media content, such as per-
ceived relationships with characters in a mediated envi-
ronment, to the usability of specific system devices.
Subject to further validation, we expect the ITC-SOPI
to offer great utility to presence researchers using very
different media systems ranging from standard broad-
cast displays to advanced, fully immersive VEs.
Author note: The ITC-Sense of Presence Inventory
(ITC-SOPI) is copyright of the UK Independent Tele-
vision Commission. To maintain the validity of the
copyright, the ITC-SOPI cannot be published in this
journal. If any laboratories are interested in using the
ITC-SOPI for research purposes, please contact Dr.
Jane Lessiter (j.lessiter@gold.ac.uk) or Dr. Jonathan
Freeman (j.freeman@gold.ac.uk). Subject to laboratories
agreeing to a short series of conditions, free use of the
questionnaire will be approved. Under these terms, the
ITC-SOPI is currently in use in ten laboratories world-
wide.
Acknowledgments
The generous support of Dr. Nick Lodge and Dr. David Har-
rison (Technology Group) of the UK Independent Television
Commission to the Immersive TV Project at Goldsmiths Col-
lege, University of London is gratefully acknowledged. Dis-
cussions with Cath Dillon (Goldsmiths College, London) and
Dr. Matthew Lombard (Temple University, USA) were also
useful in the early stages of the questionnaire development.
The authors would also like to acknowledge with thanks their
partners from the European Commission’s ACTS TAPES-
TRIES project (1996–1999), in particular Dr. S.E. Avons
(University of Essex) and Dr. Wijnand IJsselsteijn (IPO Cen-
ter for User-System Interaction, NL), with whom several of
the ideas behind the research reported here were originally
discussed. Finally, we are very grateful to the media companies
whose customers completed the ITC-SOPI.
References
Barfield, W., Baird, K. M., & Bjorneseth, O. J. (1998). Pres-
ence in virtual environments as a function of type of input
device and display update rate. Displays, 19, 91–98.
Barfield, W., & Weghorst, S. (1993). The sense of presence
within virtual environments: A conceptual framework. In G.
Salvendy & M. Smith (Eds.) Software and hardware inter-
faces (pp. 699–704). Amsterdam: Elsevier.
Barfield, W., Zeltzer, D., Sheridan, T. B., & Slater, M.
(1995). Presence and performance within virtual environ-
ments. In W. Barfield & Furness, T. A. (Eds.). Virtual envi-
Lessiter et al. 295
ronments and advanced interface design (pp. 473–541) Ox-
ford: Oxford University Press.
Bedford, A. (1997). On Clark-Watson’s tripartite model of
anxiety and depression. Psychological Reports, 80(1), 125–
126.
Draper, J. V., Kaber, D. B., & Usher, J. M. (1998). Telepres-
ence. Human Factors, 40, 354–375.
Ellis, S. R. (1996). Presence of Mind: A reaction to Thomas
Sheridan’s “Further Musings on the Psychophysics of Pres-
ence.” Presence: Teleoperators and Virtual Environments, 5,
247–259.
Fontaine, G. (1992). The experience of a sense of presence in
intercultural and international encounters. Presence: Teleop-
erators and Virtual Environments, 1, 482–490.
Freeman, J. (2000). Subjective and Objective Approaches to the
Assessment of Presence. Unpublished doctoral dissertation,
University of Essex.
Freeman, J., & Avons, S. E. (2000). Focus group exploration
of presence through advanced broadcast services. Proceed-
ings of the SPIE, Human Vision and Electronic Imaging V,
3959–3976.
Freeman, J., Avons, S. E., Pearson, D., & IJsselsteijn, W.
(1999). Effects of sensory information and prior experience
on direct subjective ratings of presence. Presence: Teleopera-
tors and Virtual Environments, 8, 1–13.
Freeman, J., Avons, S. E., Meddis, R., Pearson, D. E., & IJs-
selsteijn, W. (2000). Using behavioral realism to estimate
presence: A study of the utility of postural responses to mo-
tion stimuli. Presence: Teleoperators and Virtual Environ-
ments, 9(2), 149–164.
Heeter, C. (1992). Being there: The subjective experience of
presence. Presence: Teleoperators and Virtual Environments,
1, 262–271.
Hendrix, C., & Barfield, W. (1996a). Presence within virtual
environments as a function of visual display parameters.
Presence: Teleoperators and Virtual Environments, 5, 274–
289.
———. (1996b). The sense of presence within auditory vir-
tual environments. Presence: Teleoperators and Virtual Envi-
ronments, 5, 290–301.
IJsselsteijn, W., de Ridder, H., Hamberg, R., Bouwhuis, D.,
& Freeman, J. (1998). Perceived depth and the feeling of
presence in 3DTV. Displays, 18, 207–214.
IJsselsteijn, W. A., de Ridder, H., Freeman, J., & Avons, S.
(2000). Presence: Concept, determinants and measurement.
Proceedings of the SPIE, Human Vision and Electronic
Imaging V, 3959, 520–529.
Kim, T., & Biocca, F. (1997). Telepresence via television:
Two dimensions of telepresence may have different con-
nections to memory and persuasion. Journal of Computer
Mediated Communication [On-line serial], 3(2). Avail-
able: http://www.ascusc.org/jcmc/vol3/issue2/kim.
html.
Kline, P. (1993). Personality: the psychometric view, London:
Routledge.
Lodge, N. K. (1999). Being part of the fun—Immersive televi-
sion! Keynote address at the Conference of the Broadcast Engi-
neering Society (India), New Delhi, February 1999.
Lombard, M., & Ditton, T. (1997). At the heart of it all: The
concept of presence. Journal of Computer Mediated Com-
munication, [On-line serial] 3(2), Available:
http//www.ascusc.org/jcmc/vol3/issue2/lombard.html.
North, M. M., North, S. M., & Coble, J. R. (1997). Virtual
environments psychotherapy: A case study of fear of flying
disorder. Presence: Teleoperators and Virtual Environments,
6(1), 127–132.
Oppenheim, A. N. (1992). Questionnaire design, interviewing
and attitude measurement. London: Pinter Publishers.
Rust, J., & Golombok, S. (1989). Modern psychometrics: The
science of psychological assessment. London: Routledge.
Schubert, T. W., Friedmann, F., & Regenbrecht, H. T.
(1999). Decomposing the sense of presence: Factor analytic
insights. Paper presented at the Second International Work-
shop on Presence, University of Essex, Colchester, April
1999 [on-line]. Available: http://www.uni⬃jena.de/⬃sth/
vr/insights.html.
Sheridan, T. B. (1992a). Musings on telepresence and virtual
presence. Presence: Teleoperators and Virtual Environments,
1, 120–125.
———. (1992b). Defining our terms. Presence: Teleoperators
and Virtual Environments, 1, 272–274.
Slater, M. (1999). Measuring presence: A response to the Wit-
mer and Singer Presence Questionnaire. Presence: Teleopera-
tors and Virtual Environments, 8, 560–565.
Slater, M., Steed, A., McCarthy, J., & Maringelli, F. (1998).
The influence of body movement on presence in virtual en-
vironments. Human Factors, 40, 469–477.
Slater, M., & Usoh, M. (1994). Representations systems, per-
ceptual position, and presence in immersive virtual environ-
ments. Presence: Teleoperators and Virtual Environments, 2,
221–233.
Slater, M., Usoh, M., & Steed, A. (1994). Depth of presence
in virtual environments. Presence: Teleoperators and Virtual
Environments, 3, 130–144.
296 PRESENCE: VOLUME 10, NUMBER 3
Slater, M., & Wilbur, S. (1997). A framework for immersive
virtual environments (FIVE): Speculations on the role of
presence in virtual environments. Presence: Teleoperators and
Virtual Environments, 6, 603–616.
Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate
statistics. (2
nd
ed.). New York: Harper Collins Publishers.
Thurstone, L. L. (1947). Multiple Factor Analysis. Chicago:
University of Chicago Press.
Usoh, M., Arthur, K., Whitton, M., Bastos, R., Steed, A.,
Brooks, F., & Slater, M. (1999). The visual cliff revisited: A
virtual presence study on locomotion. Paper presented at the
Second International Workshop on Presence. University of
Essex, Colchester, April 1999 [on-line]. Available: http://
www.cs.ucl.ac.uk/staff/m.usoh/bt.html.
VIRTUE (2000). European Commission, Information Society
Technologies Project [on-line]. Available: http://www.
cordis.lu/ist/projects/99-10044.htm.
Watson, D., Clark, L. A., Weber, K., Assenheimer, J. S.,
Strauss, M. E., & McCormick, R. A. (1995). Testing a tri-
partite model: II. Exploring the symptom structure of anxi-
ety and depression in student, adult, and patient samples.
Journal of Abnormal Psychology, 104(1), 15–25.
Welch, R. B., Blackmon, T. T., Liu, A., Mellers, B. A., &
Stark, L. W. (1996). The effects of pictorial realism, delay of
visual feedback, and observer interactivity on the subjective
sense of presence. Presence: Teleoperators and Virtual Envi-
ronments, 5, 262–273.
Welch, R. B. (1999). How can we determine if the sense of
presence affects task performance? Presence: Teleoperators
and Virtual Environments, 8, 574–577.
Witmer, B. G., & Singer, M. J. (1998). Measuring presence
in virtual environments: A presence questionnaire. Pres-
ence: Teleoperators and Virtual Environments, 7, 225–
240.
Lessiter et al. 297
