Issue
Article
Vol.30 No.3, July 1998
Article
Issue
Issue |
Article |
Vol.30 No.3, July 1998 |
Article |
Issue |
One of the simplest examples of non-verbal communication is that of proximity, i.e. the distance at which conversants stand (or sit) from each other when interacting. This is known to be guided by certain social rules which take into account personal relationships, culture, personality and the purpose of the discussion, violation of which leads to various psychological and behavioural effects.
One such effect is persuasion. Sales people will tend to approach customers closer than other business people. Video has in the past been shown to lead to more effective negotiation and bargaining than audio only conditions in experimental tasks. So far though, it remains unknown whether or not similar effects of proximity exist within video-mediated communication -- or indeed even whether or not proximity can be conveyed through video. This study investigates the role of the proxemic information within the video image in a formal persuasion environment using both subjective and objective measures of attitude and behaviour.
A simulation of a customer seeking financial advice from an advisor across a live video link had two proxemic conditions -- the image of the advisor appearing either very close or far away. Subjective impressions of both the task and technology were analysed by questionnaire and showed no subjective differences between conditions. Structural analysis of the dialogues however, indicated customers to be more interactive in the close condition. When the advisor appeared close the customer spoke more often, said more, took longer and made more instances of overlapping speech. The similarity between these results and those of familiarity studies are discussed and a hypothesis of perceived close proximity resulting in perceived familiarity is offered.
The study shows that proxemic information is preserved in video conferencing and produces effects similar to those of face-to-face interactions but less pronounced. This is explained by the video conveying only visual proxemic information compared to the multimodal information available in face-to-face interaction.
Keywords: Proxemics, video-mediated communication, banking, dialogue, turn-taking, familiarity.
One of the reasons video is seen as an improved technology for communication than audio only systems is the ability it has to transmit non-verbal information, present in face-to-face interaction but absent with the loss of the visual channel. Numerous studies have been carried out investigating eye gaze and gesture, however little work has investigated proximity i.e. the distance away an individual appears to be -- arguably the most fundamental element of non-verbal communication.
In face-to-face interaction, proximity is an important source of non-verbal communication information. Social rules appear to exist regarding physical closeness when people interact, depending upon their relationship (Allegier & Byrne, 1973), culture (Vaksman & Ellysman, 1979), personality (Leary, 1983) and situation (Alexander & Rudd, 1981). Hall (1966) categorised interaction distances for seating as intimate space (6-18 inches), personal space (18 inches -- 4 ft), and social space (4-12 feet). For each particular type of interaction there is an appropriate distance which corresponds to one of these distances. For example, talking to a close friend may occur within personal space, whereas talking to a stranger will usually occur within the social space.
Of these categories, the most heavily researched is personal space. The term was actually first proposed by Sommer (1959), and has been regarded as a type of territory common to both humans and animals (Scheflen, 1970; Vine, 1975). Personal space is roughly circular but with more space in front (Cook, 1978; Argyle, 1988). This shape of space has generally been accepted as the reason that people tend to turn their orientation away from people who approach too close -- thus decreasing the invasion into their personal space.
Since it was found that humans are very good at judging their proximity to others, researchers soon began to investigate the effects of breaking the normal rules. Should one person stand closer than usual then compensatory behaviours are likely and psychological consequences may also occur.
Compensatory behaviours (reviewed by Capella, 1981, 1983, 1994) include:
Interestingly, the time taken to leave the situation actually tends to increase (Poyatos, 1975) suggesting perhaps that this is not simply a situation where someone feels uncomfortable and thus attempts to leave as fast as possible.
As well as these behavioural effects of proximity, a number of psychological consequences may arise. Paterson (1968) found that confederates who stood closer were perceived as warmer and liking the subjects more. Rosenfeld (1966) also showed that people who wish to seek approval will also stand closer than normal. The reason for this can only be speculated but could possibly be explained by applying the theory of cognitive dissonance (Festinger, 1957). The dissonant situation is that although you do not know the person who stood near you, you let them stand close, where normally only your good friends stand. This dissonance can be resolved by thinking "I like this person".
There is also reason to believe that close proximity increases powers of persuasion. Mehrabian & Williams (1969) showed that close proximity increased intended and perceived persuasiveness. Other factors which are known to increase persuasion include liking or involvement, louder and faster voices, looking more and touching (Argyle 1988, p.260) all of which also bear similarity to closer proximity. Being too close in certain situations however, leads to a decrease in influence (Albert and Dabbs, 1970). This persuasion effect may well be related to the effect of liking, if you let someone approach your personal space, then you must trust them and like them and therefore you will be more persuaded by them.
All studies cited so far have been carried out in copresent situations. With the advent of multimedia technology, it is important to find out whether or not proxemic information can be preserved across video conferencing, and if so then do similar psychological and behavioural effects occur?
The finding that adding a visual communication channel improves performance in tasks with a social element such as negotiating and bargaining (Short et al., 1976; Reid, 1977; Williams, 1977) may be related to the effect of proxemics. When the other person can be seen, it is less likely that a negotiation will end in deadlock. It is unknown though, whether changing the perceived distance on the video would cause differences in persuasion or if the inclusion of video leads to a level of presence which remains unchanged by differences in the nature of the image. The present research set out to investigate this.
One study which investigated spatial presence in video telephony is by Muhlbach, Bocker & Prussog (1995) who found a stereoscopic videoconferencing system to increase telepresence compared with a monoscopic one. This however involved wearing special glasses and was thus impractical for real world applications.
This question is interesting not only scientifically -- concerning whether or not such information can be transferred using video -- but also from an applied point of view. If there are effects of perceived distance then there will be implications for any interaction which takes place using a video link. One such type of interaction involves video call centres which are beginning to be set up in preference to audio call centres, particularly within the financial services. Video call centres provide experts which a customer may contact for help or advice regarding a transaction such as a loan through a self service kiosk. From the perspective of the bank this is a sales situation. If perceived distance has effects on persuasion, likeability, length of interaction and so on then these factors all relate to questions of how to get the best out of such a technology in order to promote sales.
Studies into the use of video conferencing have generally shown it to be unsuccessful at creating face to face style interactions from the point of view of the behaviours and dialogues of the individuals involved (e.g. Anderson et al., 1997). The aim of the present study is to investigate the extent to which proxemic information is transmitted across a video link and what effects this information has on both an individual's subjective impression of another and the nature of the communication used in a video-linked interaction.
In order to investigate this a banking scenario was used. Participants were playing the role of a customer (C) seeking advice on investment from a financial advisor (FA). The appearance of the financial advisor was varied by either appearing very close (close condition) or far away (far condition), while the FA's view of the customer was kept constant. The FA was attempting to give good investment advice to the C but also to try to persuade the C to take a particular option. This task was chosen due to it having a social element (persuasion) and as a realistic application of video technology.
If video mediated communication does preserve the same proxemic information as face to face interaction and causes the same pattern of effects as in face-to-face interaction then the following hypotheses can be made.
HYPOTHESIS 1. When the FA appears close then the interactions should take longer.
HYPOTHESIS 2. When the FA appears close then the persuasion will be more effective.
HYPOTHESIS 3. When the FA appears close then the C would form a more positive impression of the FA.
HYPOTHESIS 4. Turns should be shorter in the close condition.
Participants role-played the part of a customer seeking advice on how best to invest £5000 inheritance. The customer had gone to their local automated branch of their bank to seek advice, sat down at one of the kiosks and used the application to find the section offering financial advice.
At this point participants began a discussion with a financial advisor, permanently available across a live video link. There were four options available to the customer, each of which presented a page of information. The information was taken from real leaflets advertising real accounts in banks and could be viewed one page at a time. Both the FA and the C could alter the page in such ways as using pointers to highlight areas of text.
The options offered were as follows:
BONUS ACCOUNT: A high interest but taxable account offering saving with 120 days notice required for penalty free withdrawals.
TESSA: A 5 year tax free savings plan offering access only to 80% of the interest but guaranteed high interest with no risk.
SHARES: A high risk option capable of making the greatest capital gains but guaranteeing nothing.
PEPs: A lower risk option than the shares but still based on the stock exchange. Opportunity for higher gains than the no risk accounts but lower than the shares.
The role of the FA was played by a postgraduate psychology student. They were instructed to attempt to subtly persuade customers towards the bonus account as this would make the most money for the bank. The advice should, however, appear to be independent as it was most important for the customer to be happy with the advice given. The FA was also asked to look into the camera as much as possible in order to give the impression of looking at the customer. The FA was paid £5 per trial.
The customer received instructions immediately prior to the experiment outlining the scenario to them. They were a young industrial designer, reasonably comfortably off, with a fairly stable job. The rest of the character's situation was left up to them.
The independent variable in this experiment was perceived distance (proximity) of the FA to the customer. There were two levels to this, close and far. The close condition involved the FA being seen as virtually just a face. The far condition allowed plenty of background and a head and shoulders view of the FA.
It is worth noting that only the customer's view of the FA changed and not vice versa. The advisor always saw the customer at a distance in between that of the two conditions. The reason for this is that this study was only interested in the effects of perceived distance on the customer, and so only visual information was altered. Had the visual information been altered for both parties then any effects found could be due either to the visual information directly or as a result of the reaction to changes in each others behaviour. Issues of generalisability to real life interactions are answered by the fact that in this experiment and most likely in reality the FA is effectively working to a script and it is therefore unlikely that their behaviour will alter much depending on perceived distance. Were this work to be generalised to compare with spontaneous social interaction as opposed to this more formalised scenario then this issue would require to be addressed. It should also be pointed out that experiments using confederates also work to scripts, rather than acting spontaneously so the effects caused by reacting to behaviour changes are not being investigated.
The customer was faced with a screen consisting of a whiteboard containing four pages of coloured, textual information (one for each option). It was possible for either the C or the FA to highlight areas on the board by way of arrows etc. but not to remove any of the information.
Situated towards the upper left hand side of the whiteboard there was a video window which used the rtvc (adapted) SunVideo system. The videowindow was 384 x 288 pixels (115mm width x 85mm height) in size. Video and audio were transmitted directly rather than across a network, ensuring broadcast quality frame rate with no appreciable audio lag. This high quality video avoids the deleterious effects caused by poor frame rate and audio lag shown in the past (e.g. Cohen, 1982; Johannsen, 1984; Tang & Isaacs, 1993; Isaacs & Tang 1994; Whittaker, 1995).
The cameras used were JVC videomovie GR-AX 55E. They were placed to the side of the screen as close as possible to the video window in order to get as close to eye contact as possible.
Perceived distance was altered by way of the zoom on the camera. Very close was taken as cropping the video window around the face and very far away allowed a view including the whole of the upper torso. It is worth acknowledging that there is more to perceived distance than simply visual cues, but in the experiment no differences in sound were made, and thus only visual proxemic information was altered.
30 participants (15 males and 15 females; age range 17-32, median = 19) were all undergraduates and graduates from the University of Glasgow. No participant had ever done any banking by video conferencing equipment before, and no-one rated themselves as very inexperienced with computers. Participants were told they were going to take part in a psychology experiment investigating banking by video conferencing. It was explained that they would be audio and video taped for research purposes but that these would be treated as strictly confidential. Participants were informed that there was no time limit and that at the end there would be a questionnaire. Participants were paid £5 for their participation. All participants signed their consent.
At the end of the experiment participants completed a short questionnaire which was designed to investigate their experience of three areas.
PRIOR EXPERIENCE: Prior experience of both video conferencing and computers in general was monitored to ensure any effects found were not simply due to level of experience with the technology.
THE TASK: This section included questions about how the individuals felt about the realism of the task, qualities of the financial advisor (e.g. likeability, sincerity, helpfulness), and quality of the advice given.
THE TECHNOLOGY: Questions in this section investigated how participants felt about the technology itself, including whether they considered they would be willing to use it in real life, and how it compares to other methods of banking.
To measure objectively the actual participant behaviour in terms of communication processes, analysis of the dialogues was carried out. Of specific interest within the dialogues were:
BACKCHANNELS: Backchannels are brief responses indicating agreement, understanding or attention which do not disrupt the flow of the speaker. In this study they were taken to be "mhm", "uhhuh", and "mmm" when said on their own without disrupting the speaker. It should be noted though that sometimes these same words could be used as answers to questions and then would not be considered backchannels.
TURNS: Turns are the most widely studied feature of dialogue and have in the past -- like proxemic information -- been shown to be related to familiarity (Boyle, Anderson & Newlands, 1994). A turn is defined as beginning when one person speaks and lasting until another speaks. A pause, no matter how long, does not automatically end a turn. Backchannels are considered turns but do not constitute ending the speakers turn when they are uttered. Overlaps on the other hand do end the speakers turn and are counted as turns themselves. Thus a period of speech by one person during which there are 2 backchannels uttered by the other person has a total of 3 turns in it. Had the backchannels been interruptions however, there would have been 5 turns.
OVERLAPS: Overlaps were considered to have taken place if one or more words of the second speaker's contribution were perceived to have overlapped the first speaker's contribution. This is the same definition as Boyle et al (1994) used, although they point out that the definition of overlaps and interruptions has been far from consistent across the literature.
WORDS: Words counted included all utterances, including "uhhuh", "mhm", and "mmm".
Analysis of the questionnaires yielded no significant differences on any of the subjective scales measured (including impressions of the advisor, persuasibility, impressions of the technology or likelihood to use the technology in "real life").
PERCEIVED DISTANCE: The questionnaires showed that in 6 out of 15 close trials the agent was considered to be too close but only 1 of the far trials the agent was considered to be too far away. This suggests that participants were comfortable with considering the image in terms of close and far away.
THE TASK: All of the options were chosen at least twice (see Table 1). The bonus account (which the FA was attempting to persuade people to take) was taken the most often (46.7% compared with 27.7% for PEP) but this did not differ significantly between conditions.
| Close | Far | Overall | |
| PEP | 28.6 | 26.7 | 27.7 |
| Shares | 21.4 | 6.7 | 13.3 |
| TESSA | 7.1 | 13.3 | 10.0 |
| Bonus Account | 42.9 | 53.3 | 46.7 |
One interesting trend which failed to reach significance due to the small numbers and short rating scales, involved customers perceptions of the amount of persuasion being used by the FA. Table 2 seems to show that more participants in the close condition (67%) perceived persuasion being used than in the far condition (40%).
| Close | Far |
| A lot | 6.7 | 0.0 |
| A little | 60.0 | 40.0 |
| None | 33.3 | 60.0 |
| Don't know | 0.0 | 0.0 |
Interestingly, this did not correspond to lower levels of trust of the advice given, Table 3 indicates that if anything, customers in the close condition trusted the advice more than in the far condition.
| Close | Far |
| Trust a lot | 40.0 | 20.0 |
| Trust a little | 46.7 | 66.7 |
| Neither trust nor distrust | 6.7 | 6.7 |
| Distrust | 6.7 | 6.7 |
| Distrust a lot | 0.0 | 0.0 |
Other overall results of interest included the following:
THE TECHNOLOGY: Questions about the technology, despite again yielding no significant differences between proxemic conditions, did offer interesting overall results about peoples readiness to use such technology to do their banking.
Face to face remained the overwhelming first choice for seeking advice (76%) with leaflets (no human contact) the first choice for 10% of respondents. Video did tend to be rated second however and outscored telephone banking for all but two participants.
When asked how far people would travel to get face to face advice as opposed to video advice being available in their local branch, 24% responded zero, 21% up to 15 min, 34.5% up to half an hour with 20.5 % willing to travel for more than half an hour to get advice.
Participants were given a selection of financial transactions and asked which they would be willing to carry out by video. The results are shown in table 4.
| Close | Far | Overall |
| Yes | 53.3 | 73.3 | 63.3 |
| No | 6.7 | 13.3 | 10.0 |
| Unsure | 40.0 | 13.3 | 26.7 |
It appears from Table 4 that customers see applying for mortgages as fairly inappropriate for the video whereas opening and closing accounts were seen as the most appropriate. An interesting trend to note is that for all transactions with the exception of customer complaints, those in the close condition were more willing to use the technology. This perhaps suggests that those in the close condition were happier with the technology. The small number of participants however, means that caution should be taken in interpreting these results. More caution is thrown by the results outlined in Table 5. Which were gained from an earlier question.
Finally, respondents were given an open ended opportunity to give any other comments on the simulation. Despite it never having been mentioned anywhere in the questionnaire, 27.6% of respondents commented that there was a problem with eye contact. This is an extremely high level of response considering it was unprompted and shows that the lack of eye contact is still a major issue with this type of system.
Transcripts of the dialogues of all trials were analysed and the following results were obtained. 28 transcripts were included in the analyses, 2 transcripts were rejected due to technical difficulties. This left 15 in the close condition and 13 in the far condition.
Data was analysed using one-way ANOVAS (equivalent to independent samples t-test) and 2-way ANOVAS for independent groups.
With proximity as the independent variable the following results were found.
The overall mean time to complete the task was 548 seconds (9.08 minutes). Between conditions, there was a trend [F(1,26) = 3.23, p<0.1] towards interactions in the close condition (mean = 609 sec) taking longer than those in the far condition (mean = 477 sec).
There were significantly more turns in the close condition (mean = 138.6) than in the far condition (mean = 100.6) [F(1,26) = 4.33, p<0.05]. Analysis by role showed that this difference applied to both the C and the FA, as shown in Table 6.
| Far | Close | F | Sig. |
| FA | 47.7 | 66.3 | 4.73 | p<0.05 |
| C | 52.9 | 73.8 | 4.24 | p<0.05 |
No significant differences were found between the number of backchannels uttered in each condition, neither was there a difference between when broken down by role.
The total number of overlaps per condition fell just short of significance [F(1,26) = 3.98, p<0.1], although there was a trend towards there being more overlaps in the close trials (mean = 24.33) than the far trials (mean = 13.54).
When broken down by role however (Table 7), results show a significant difference between the number of overlaps caused by the customer (instances when the customer begins a turn before the FA has finished her turn) across conditions, with more overlaps in the close condition. The number of overlaps caused by the financial advisor did not differ significantly across conditions.
| Far | Close | F(1,26) | Sig. | |
| FA | 7.15 | 10.80 | 1.508 | ns |
| C | 6.23 | 13.53 | 6.096 | p<0.05 |
In order to get a better idea of just how many overlaps were happening, the number of overlaps as a function of the number of turns was analysed. This examines how often overlaps occur rather than simply how many there are.
There was found to be no significant difference between the overall number of overlaps when considered as a function of the number of turns taken [F(1,26) = 2.42, p>0.05].
As with the total number of overlaps however, when broken down by role, significant differences emerge. Table 8 indicates that there is a significant difference between the number of turns per overlap for the customer with there being more overlaps in the close condition.
| Far | Close | F | Sig. |
| FA | 11.60 | 10.58 | .068 | ns |
| C | 11.93 | 6.87 | 5.201 | p<0.05 |
In order to back up the last analysis, overlaps were also considered as a percentage of the number of turns. Once again, overall there is no significant difference between conditions of the number of overlaps when considered as a percentage of the number of turns [F(1,26) = 3.081, p>0.05].
When broken down by role (Table 9) it can again be seen that the customer makes significantly more overlaps in the close condition than in the far condition..
| Far | Close | F | Sig. |
| FA | 14.60 | 14.83 | 0.005 | ns |
| C | 10.63 | 17.15 | 7.613 | p<0.01 |
The total number of words spoken per trial across conditions was not significant [F(1,26) = 2.59, p>0.05]. However, when broken down by role (Table 10) then it can be seen that in the close condition the customer says significantly more words than in the far condition.
| Far | Close | F(1,26) | Sig. |
| FA | 974 | 1177 | 1.401 | ns |
| C | 252 | 453 | 5.243 | p<0.05 |
When considering the number of words spoken as a percentage of the total words spoken then we have a measure of the share of the dialogue. When considered in this way, results show that the customer increases their share of the dialogue from just under 21% in the far condition to just over 26% of the words spoken in the close condition..
| Far | Close | F | Sig. |
| Customer | 20.97 | 26.11 | 5.678 | p<0.05 |
The number of words per turn was also analysed and is summarised in Table 12. Results show that on top of a very large difference in the number of words per turn depending upon role (due to the financial advisor doing all the explaining) there was a trend towards customer turns being longer in terms of the number of words in a turn in the close condition. There was no such trend for the FA.
| Far | Close | F(1,26) | Sig. |
| FA | 19.83 | 18.35 | .660 | ns |
| C | 4.79 | 5.82 | 2.945 | ns (p<0.1) |
Since this type of scenario differs from normal conversation, on which most linguistic experiments are based, it is important to know exactly what communication differences are occurring as a result of the task itself. For this reason data was also analysed with role being the independent variable. This involves considering the two roles as independent trials and the proximity conditions being the same. Table 13 summarises the results.
Table 13 shows that there are no differences inherent in the task which cause the financial advisor and customer to differ in the number of turns or overlaps they make. The task does on the other hand lead to more backchannels being made by the customer, while the financial advisor says more words, more words per turn, and has a greater share of the dialogue.
Early analysis shows that despite no subjective differences being indicated by the questionnaires, subtle effects of proximity are in evidence in the structure of the dialogues.
When looking at these results it is important to remember that the perceived proximity of the other person only differed for the customer. The view seen by the FA remained the same in all conditions and was in between that of the far and close conditions of the customer. Bearing this in mind, the fact that only the customer was showing significant differences in the dialogues is further evidence that they are caused by effects of perceived proximity.
Results initially can be summed up by saying that when the financial advisor appeared very close, the customer made more interruptions, overlapping speech and indeed said more than when the financial advisor appeared far away. This in turn tended to increase the length of the interactions without altering subjective feelings about the individual or the interaction (as measured by the questionnaire).
The task itself leads to the financial advisor saying much more than the customer and holding a greater share of the dialogue. This in turn leads to there being more backchannels being spoken by the customer (presumably to provide more feedback during long periods of the financial advisors speech, the like of which are not often present in customer dialogue).
Let us consider the results in terms of the initial hypotheses that were made based on face-to-face research into proximity.
HYPOTHESIS 1. When an individual appears closer then the interactions should take longer.
There was a trend towards the task taking longer in the close condition, although this did fall just short of significance. This is weak evidence in favour of the hypothesis, if more trials were run it is likely that indeed, time taken would become significant. The trend was in the predicted direction but we must consider this inconclusive at present.
HYPOTHESIS 2. When the FA appears close the persuasion will be more effective.
Results showed no difference between conditions for which account was chosen. Consequently this hypothesis must be rejected. There was a trend however, suggesting that close perceived proximity leads to an increase in the perception of persuasion although not related to success. This would fit in with the findings of Mehrabian & Williams (1969) who found that close proximity was used as a measure of persuasion, although they did not measure the success of that persuasion.
HYPOTHESIS 3. The participants should form a more positive impression of the advisor when they see a closer image.
Results from the questionnaires have shown there to be no significant differences on any of the questionnaire questions. The conclusion must be drawn that across a video link, proxemic information does not cause differences in conscious impression forming, at least in this scenario. It seems likely, therefore that for impressions to be altered requires more proxemic information than that needed to alter behaviour. Hall (1966) suggested that one of the reasons that personal space differed from social space was the fact that at the closer distance, other senses come into play such as the ability to touch. In this experiment there was no physical, olfactory or audio cues to proximity. It is therefore perhaps not surprising that any proxemic effects appear to be smaller across the video link.
HYPOTHESIS 4. Turns should be shorter in the close condition.
The only guide that we have to the length of the turns comes from the data on the number of words per turn. This showed that there was no overall difference in the number of words per turn across conditions but that there was a trend towards the customer saying more words in their turns in the close condition. This would appear to be evidence to reject the above hypothesis.
There are two possible reasons for this result -- the first is that in the more passive far condition customer turns are generally very short -- filled with lots of one word answers -- as a function of the task. For these turns to get shorter is not really possible. Thus a desire to speak more often is not going to decrease the length of the turns -- were the two conversants to have similar speaking roles perhaps the result would be different.
A second reason is that similar results have been shown when eye contact is preserved versus no eye contact (Doherty-Sneddon et al., 1996) and visibility is possible compared to a non-visible condition (Boyle et al., 1994). Suggested to be sign of inefficiency by these researchers it is at least possible to speculate that the close condition is producing some sort of increased visibility or social presence.
In the past, familiarity has been shown to be related to more extended dialogues. Boyle et al. (1994) found that pairs who knew each other produced more extended dialogues with more turns and more words than pairs of strangers. In the present experiment those in the close condition had an extended dialogue in comparison with those in the far condition both in terms of turns (138.6 vs. 100.6) and words said by the customer (453 vs. 252). It is possible that this shows an effect of close perceived distance leading to a feeling of familiarity. This would support the face to face data on social distance described earlier where individuals who stood closer were treated more like friends than those that stood further away.
Boyle et al. also found that familiar pairs had less overlapping speech, whereas in our experiment there was actually more overlapping speech in the close condition, which would go against this theory of familiarity. On the other hand Boyle et al. do not hypothesise over the reasons for familiarity to cause either more turns or fewer overlaps. It may be possible to subdivide familiarity into two factors, one of which governs turns, another of which controls overlapping speech. A reason for this could be that actual familiarity causes better synchronisation of speech because people who know each other and are used to interacting with each other are better and more efficient at predicting ends of turns by their partner, and thus overlap less than strangers. If this was indeed the case then it would not be expected to be the case in the proximity experiment as it is still a conversation between two strangers.
At the same time, a reasonable hypothesis for familiar pairs to take more turns to complete a task could quite conceivably hold no relation to actually knowing someone. It could be based on feeling more at ease with them and so feeling that they could say more and feel less inhibited about speaking whenever they wanted to. The consequences of this hypothesis for the present experiment, however, would be that if customers in the close condition were treating the FA as if they were familiar then they would still feel able to say more, be less inhibited and thus actually make the higher level of turns despite not actually being more familiar. This would also explain why the customer says more words in the close condition than the far but the FA shows no significant difference -- the FA always saw the same view so did not have the same feeling of perceived familiarity. This last point does not apply to turns as turns will always be fairly evenly distributed between the two participants in a conversation due to the definition of a turn ending when the other person speaks.
It is therefore hypothesised that the findings of Boyle et al. are in accordance with the findings of the present study but also a new way of looking at familiarity is suggested. There are differences between actual familiarity and perceived familiarity which may or may not be conscious.
In the close condition, customers said more and the dialogues contained more turns -- they were more interactive. Research into interactivity of dialogues suggests that a consequence of this may be that memory and understanding of the concepts in the dialogue will probably be enhanced in the close condition.
This theory has been proposed by Clark and colleagues and states that conversation involves participants trying to establish the mutual belief that the listener has understood what the speaker meant. This is a collaborative process called grounding, and requires actions by both the speaker and listener (Clark & Schaefer, 1987). This is known as the collaborative view of understanding (Schober & Clark, 1989). In the past there has been a traditional view known as the autonomous view of understanding which asserts that merely hearing and seeing all that happens and having the same background knowledge is sufficient to understand fully. Schober & Clark (1989) showed that people who overheard instructions being given did not perform as well on a task as those to whom the instructions were given. Those receiving the instructions could ask questions, clarify etc., those who overheard could only listen to all that was said.
This suggests quite strongly that participating in an interaction aids understanding and memory in comparison to listening. The implications for the present experiment should be clear. Those who have a more interactive conversation may well understand more than those who sit and listen more. Consequently those in the close condition may well have a better understanding than those in the far condition.
A difference between this study and the Schober and Clark study is that they only went so far as to compare having an interactive conversation with listening to an interactive conversation but not taking part. It could be suggested that there is actually no evidence that once in an interactive conversation (as all participants in the present study were) the level of interactivity matters. This is true, and in future research a test of understanding and recall will be brought in precisely to test this hypothesis, but there does seem to be logical grounds for rejecting it.
The grounds for such a hypothesis lie in the assumption that individuals will ask questions and clarify and query and check understandings only as often as they need to reach their recognition point. In other words, not everyone understands the same concepts at the same rate. Fair enough. However, in this experiment the factor which has led to differences in levels of interactivity has been the perceived distance of the FA. Therefore individual differences in capacity to learn and understand should have been negated by counterbalancing. There are therefore only two possible conclusions that can be drawn, either the FA appearing close led to her being understood less and thus customers required more checking and aligning, or the FA appearing a far distance away inhibits the customers ability or intention to check their understanding. A recall and understanding test would be able to test these hypotheses. If there was no difference in the understanding between conditions then appearing close is decreasing initial understanding, if on the other hand those in the close condition understood better then it would suggest that level of interactivity and understanding are indeed correlated.
Although at present an open mind must be kept on this before reaching a definite conclusion, there does seem a certain common sense logic to the hypothesis that the more you interact the more you understand. Hopefully future research will clear this up as it does leave an ambiguous message for business. If one theory is right then agents should appear far away as they will get through calls quicker and customers will understand perfectly well. If the other theory is correct, and if businesses using this type of technology genuinely value their customers needs and wish them to understand better then they should make their agents appear close so as to ensure customer understanding.
This experiment has shown that proxemic information is preserved in video conferencing but only to a certain extent. The effects that it produces are generally in line with those of face-to-face interactions but on a lower level -- most likely due to the unimodal proxemic information available compared to the multimodal information available in copresent interaction.
A financial advisor appearing to be very close (within personal space) tends to decrease the formality of an interaction, with the customer taking a greater role. There are more turns and more interruptions spoken when the advisor appears closer. Previous work has suggested that more turns are a sign of familiarity but are accompanied by fewer interruptions. This is accounted for by a hypothesis of an unconscious perceived familiarity.
This experiment has provided further evidence of the nature of the worth of analysing dialogue as a tool for examining subtle psychological effects. It also stands as an excellent example of the role of cognitive psychology within the design process of modern multimedia applications.
Albert, S. & Dabbs, J. M. (1970). "Physical distance and persuasion". Journal of Personality and Social Psychology, 15, 265-270.
Alexander, C. N. Jr. & Rudd, J. (1981). Situated identities and response variables. In J. T. Tedeschi (Ed.), Impression Management Theory and Social Psychological Research (pp. 83-103). New York: Academic Press.
Allegier, A. R., & Byrne, D. (1973). "Attraction towards the opposite sex as a determinant of physical proximity". Journal of Social Psychology, 90, 213-219.
Anderson, A. H., O'Malley, C., Doherty-Sneddon, G., Langton, S., Newlands, A., Mullin J., Fleming, A. M. & Van der Velden, J. (1997). "The impact of VMC on collaborative problem solving: An analysis of task performance, communicative process and user satisfaction". In Finn, K. E., Sellen, A., Wilbur, S. (Eds.) Video-Mediated Communication. Lawrence Erlbaum Assoc. Hillsdale. New Jersey.
Argyle, M. (1988). Bodily Communication. Routledge. London and New York.
Argyle, M. & Dean, J. (1965). "Eye-contact, distance and affiliation". Sociometry, 28, 289-304.
Becker, F. D. & Mayo, C. (1971). Delineating personal distance and territoriality". Environment and Behaviour, 3, 375-381.
Boyle, E. A., Anderson, A. H., & Newlands, A. (1994). "The effects of visibility on dialogue and performance in a co-operative problem solving task". Language and Speech, 37, 1-20.
Capella, J. N. (1981). "Mutual influence in expressive behaviour: Adult-adult and infant-adult dyadic interaction". Psychological Bulletin, 89, 101-132.
Capella, J. N. (1983). "Conversational involvement: Approaching and avoiding others". In J. M. Wieman & R. P. Harrison (Eds.), Nonverbal interaction (pp. 113-152). Beverly Hills, CA: Sage.
Capella, J. N. (1986). "Reciprocal and compensatory reactions to violations of distance norms for high and low self-monitors". In M. L. McLaughlin (Ed.), Communication yearbook 9 (pp. 359-376).
Capella, J.N. (1994). "The management of conversational interaction in adults and infants". In M. L. Knapp & G. R. Miller (Eds.) Handbook of Interpersonal Communication 2nd ed. (pp. 380-418).
Clark, H. H. & Schaefer, E. F. (1987). "Collaborating on contributions to conversations." Language and Cognitive Processes, 2, 19-41.
Cohen, K. (1982). "Speaker interaction: video teleconferences versus face-to-face meetings." In Proceedings of Teleconferencing and Electronic Communications, pp 189-199.
Cook, M. (1970). "Experiments on orientation and proxemics". Human Relations, 23, 61-76.
Doherty-Sneddon, G., Anderson A., O'Malley, C., Langton, S., Garrod, S. & Bruce, V. (1997). "Face-to-face and video mediated communication: A comparison of dialogue structure and task performance". Journal of Experimental Psychology: Applied, 3, 105-125.
Efran, M. G. & Cheyne, J. A. (1974). "Affective concomitants of the invasion of shared space: Behavioural, physiological and verbal indicators". Journal of Personality and Social Psychology, 29, 219-226.
Festinger, L. (1957). A Theory of Cognitive Dissonance. Stanford: Stanford University Press.
Grayson, D. (1996). "Designing a multimedia system: effects on consumer performance and attitudes". Unpublished report, available within NCR.
Hall, E. T. (1966). The Hidden Dimension. Garden City, NY: Doubleday.
Isaacs, E. A., & Tang, J. C. (1994). "What video can and cannot do for collaboration: A case study." Multimedia Systems, 2, 63-73.
Johansen, R. (1984). Teleconferencing and Beyond. New York, NY: McGraw-Hill.
Johnson, C. F. & Dabbs, J.M. (1976). "Self-disclosure in dyads as a function of distance and subject-experimenter relationship. Sociometry, 39, 257-263.
Leary, M. R. (1983). Understanding social anxiety: Social, personality and clinical perspectives. Beverly Hills, CA: Sage.
Mehrabian, A. & Williams, M. (1969). "Nonverbal concomitants of perceived and intended persuasiveness". Journal of Personality and Social Psychology, 13, 37-58.
Muhlbach, L., Bocker, M. & Prussog, A. (1995). "Telepresence in videocommunications: A study on stereoscopy and individual eye contact." Human Factors, 37, 290-305.
Patterson, M. L. (1968). "Spacial factors in human interaction". Human Relations, 21, 351-361.
Patterson, M. L. (1977). "Interpersonal distance, affect, and equilibrium theory. Journal of Social Psychology, 101, 205-214.
Patterson, M. L. (1991). "A functional approach to nonverbal exchange". In Feldman, R. S. & Rime, B. (Eds.) Fundamentals of nonverbal behaviour (pp 458-495). Cambridge University Press.
Poyatos, F. (1975). "Cross cultural study of paralinguistic "alternants" in face-to-face interaction". In Kendon, A., Harris, R.M. & Key, M.R. (Eds.) Organization of Behaviour in Face-to-Face Interaction Mouton & Co. The Hague. Paris. Pp 285-314.
Reid, A. (1977). Comparing the telephone with face-to-face interaction. In (I. Pool, Ed.) The Social Impact of the Telephone. Cambridge, MA: MIT Press.
Rosenfeld, H. M. (1966). "Approval-seeking and approval-inducing functions of verbal and non-verbal responses". Journal of Personality and Social Psychology, 4, 597-605.
Sawitsky, J. C. & Watson, M. J. (1975). "Patterns of proxemic behaviour among preschool children. Representative Research in Social Psychology, 6, 109-113.
Scheflen, A. E. (1970). Territoriality and communication. Final Report NATO Symposium on Non-Verbal Communication, 35-38. Cited in Vine (1975).
Schober, M.F. & Clark, H.H. (1989). "Understanding by Addressees and Overhearers". Cognitive Psychology, 21, 211-232.
Short, J., Williams, E. & Christie, B. (1976). The Social Psychology of Telecommunications. London: John Wiley
Sommer, R. (1959). Studies in personal space. Sociometry, 22, 337-348.
Tang, J. C. & Isaacs, E. (1993). "Why do users like video? Studies of multimedia-supported collaboration. Computer Supported Collaborative Work (CSCW), 1, 163-196.
Vaksman, E. & Ellyson, S. L. (1979). "Visual and spatial behavior among US and foreign males: A cross-cultural test of equilibrium theory." Paper presented at the Eastern Psychological Association Convention, Philadelphia. Cited in Patterson (1991).
Vine, I. (1975). Territoriality and the Spatial Regulation of Interaction. In Kendon, A., Harris, R. M. & Key, M. R. (Eds.) Organization of Behavior in Face-to-Face Interaction (pp. 357-388). Mouton Publishers. The Hague. Paris.
Whittaker, S. (1995). "Rethinking video as a technology for interpersonal communications: theory and design implications." International Journal of Human-Computer Studies, 42, 501-529.
Williams, E. (1977). Experimental comparisons of face-to-face and mediated communication. Psychological Bulletin, 16, 963-976.
David Grayson is a Ph.D. student collaborating with the Department of Psychology at the University of Glasgow, Scotland and NCR (Scotland) Ltd. His research interests are the evaluation of the impact of modern multimedia technologies on users in the financial services.
Dr. Lynne Coventry is a consultant in usability aspects of and consumer acceptance of technology at NCR (Scotland) Ltd. Her research interests are the impact of multimedia communication technologies on consumer behaviour and how technology can facilitate personal financial management.
David Grayson
Multimedia Communications Group
Department of Psychology, University of Glasgow
52 Hillhead Street, Glasgow, G12 8QB, UK
david@mcg.gla.ac.uk
Dr. Lynne Coventry
Senior Cognitive Engineer
NCR Financial Solutions
St. Davids Drive, Dalgety Bay
Fife, KY11 5NB, UK
lynne.coventry@stdavids.ncr.com
Issue |
Article |
Vol.30 No.3, July 1998 |
Article |
Issue |