Volume 23 Numbers 1 & 2, 2013
Like What You See? The Effect of Video-Mediated Gazing on Information Recall and Impression Formation
Chris Fullwood & Neil Morris
When the use of videoconferencing became practicable on a global scale a debate developed concerned with whether adding a visual channel to remote communications enhances technology-mediated interactions. Indeed, the uptake of video-mediated technologies is largely based on the assumption that visual signals add value to communication (Campbell, 1998; Sellen, 1997). If visual signals did not significantly augment remote communication then audio-only transmissions (for example via the telephone) would suffice for all forms of remote interaction and have the advantage of being cheaper and technologically simpler (Fullwood & Finn, 2010). Thus it is of some considerable importance to establish the value of visual signals for remote communication since this augmentation involves a large investment in technology. Added to this debate, there has been some speculation over whether visua l signals communicated via video links have the same communicative impact as those communicated face-to-face. For instance, the manner in which video systems are set up often results in a difficulty in communicating normal variations in gazing behaviour, for example eye contact. This issue can be partially addressed by simply advising users to gaze into the camera at certain points in their communication to give the impression that eye contact is being made; a practice which has been dubbed “video-mediated gazing” (Fullwood, 2007). The current study aims to test whether this particular practice can lead to benefits frequently associated with eye contact in face-to-face communication, namely improved information recall and more positive person perceptions.
Video-Mediated Communication and Impression Formation
Although early research into VMC focused largely on how well the technology could be used to complete various real-world tasks (e.g. see Boyle, Anderson, & Newlands, 1994; Doherty-Sneddon et al., 1997; O’Malley, Langdon, Anderson, & Doherty-Sneddon, 1996), more recent research has shifted focus onto the psychological impact of engaging with the technology. One area that has received attention is the impression formation process. Although videoconferencing technology can be used by the general public for recreational purposes (i.e. with the use of a personal webcam and software applications like Skype), VMC is also often utilised for more specialised operations where it is important to gather a feel for communication partners (for example, business applications such as interviews) (Fullwood & Finn, 2010 a>). In these contexts, assessing the manner in which the technology impacts upon person perceptions is a central concern. When it comes to making judgements about other individuals, non-verbal cues (for example, facial expressions, eye contact and aspects of appearance) are said to be particularly important (e.g. see Argyle, 1994). In a face-to-face setting, first impressions are often formed with great rapidity and in many cases on the basis of very limited information (Ambady, Bernieri, and Richeson, 2000; Bar, Neta, and Linz, 2006). For instance, accurate impressions of personality have been shown to occur when raters have been exposed to targets face-to-face for only brief periods of time (4-10 seconds) and even after only being presented with a simple static photograph (e.g. see Ambady et al., 2000; Be rry, 1990). Moreover, once impressions have been formed they tend be fairly persistent and resistant to change (Asch, 1946). People tend to be very good at judging the dispositions and personalities of others face-to-face. Indeed, studies suggest accurate judgments on a number of elaborate social attributes, for example dominance, warmth and hierarchy (e.g. see Ambady et al., 2000; Berry, 1990; Brothers, 1997; Funder, 1987; Hassin & Trope, 2000; Zebrowitz, 1997). However, the question remains: does the use of video technologies interfere in any way with the impression formation process?
Investigations into aspects of the impression formation process in a VMC context have considered a variety of real-world applications, for example interviews (e.g. Chapman & Rowe, 2001), long distance collaboration (e.g. Bos, Olson, Gergle, Olson, & Wright, 2002) and sales tasks (e.g. Greenspan, Goldberg, Weimer, & Basso, 2000). In each of these contexts, the successful completion of the task may hinge upon gathering a favourable impression of another individual. In this sense, it would seem imperative that accurate impressions are formed via VMC. One task for which this would appear to be particularly relevant is the interview, and a number of investigations have considered impression formation in this context. In a study in 2001, Chapman and Webster asked participants to rate an interviewee on a number of measures includin g interpersonal and communication skills. Participants either watched an interview take place face-to-face or via a videoconference. Their findings suggest that video-mediated interviewees may be at a disadvantage as candidates are better able to express both verbal and non-verbal cues, immediate feedback and natural language in a face-to-face setting. A further study by Chapman and Rowe (2002) assessed the influence of interview structure and interview medium on interviewee ratings of interviewer friendliness and performance and the perceived attractiveness of organisations. The findings suggest that video-mediated interviewees prefer very structured interviews, whereas face-to-face interviewees prefer unstructured interviews. The fact that video-mediated interviewees prefer structure may say something about the limitations of communicating via a video link in this context. For example, it may be more difficult to interpret, or even see, b ody language cues thus making the interviewee less inclined to ad lib and to be more comfortable with a clear structure (Fullwood & Finn, 2010). The study also found that organisations which used video-mediated selection procedures were perceived less favourably and that interviewees had less positive perceptions concerning the interviewer’s performance during the interview. Overall, there seems to be evidence here supporting the idea that individuals are perceived less favourably when they communicate via video technologies and one might therefore argue that these perceptions are also less accurate.
In a further study, Chapman, Uggerslev, and Webster (2003) gathered evidence to suggest that face-to-face interviews are deemed as fairer by interviewees. Moreover, face-to-face interviewees are more likely to accept a job offer than video-mediated interviewees and interviewees rate their interview performance less satisfactorily in video-mediated interviews. Evidence to the contrary however suggests that interviewers do not evaluate interviewees less favourably via video (Straus, Miles, & Levesque, 2001) and even a potential bias in favour of videoconference applicants (Chapman & Rowe, 2001). These mixed findings perhaps suggest that although the cue-restricted context of a video interview may disadvantage some individuals, there is also the possibility that some people will benefit from being interviewed remotely, for example those individua ls who are shy or find face-to-face interviews anxiety-provoking. It is also possible that some interviewers are aware of the limitations of VMC, and therefore overcompensate in their evaluations of video-mediated interviewees to level the playing field.
Long distance collaboration presents another real-world context in which individuals need to make judgements about others. Bos et al. (2002) argue that in our modern global economy, more and more business relationships are being formed in technology-mediated contexts, for instance supplier-purchaser relationships. Unlike the one-off context of the interview, remote collaborations of this kind involve perpetual contact. In other words, individuals build up a relationship over time and are not necessarily forced into making swift decisions about another’s character (as would be the case in the interview for example). Building trust is of paramount importance in these types of business relationships (Bos et al., 2002). If collaborators cannot successfully establish and maintain trust then it is likely that they will be less open about sharing information with one another and communications will be more carefully monitored and constricted, which will inevitably impact upon productivity(Bos et al., 2002; Das and Teng, 1998). Moreover, a lack of trust between co-workers may alter the nature of collaborations with fewer instances of contact or avoidance of contact altogether (Herbsleb, Mockus, Finholt, & Grinter, 2000; Teasley et al., 2000).
In evaluating the use of technology for long distance collaborations, Bos et al. (2002) compared four different communication channels (face-to-face, video, audio and text chat) using a social dilemma game. Their findings support the notion that, comparative to face-to-face interaction, trust takes longer to establish and once established it is more fragile (i.e. individuals are more susceptible to change their mind about whether they trust another person). Bos et al. (2002) argue that VMC “narrows the visual field and masks both visual and verbal cues.” As a consequence, the absence or attenuation of these cues makes it increasingly difficult for individuals to decide whether someone else is trustworthy and therefore they may also feel less inclined to communicate their own trustworthiness.
In conjunction with the ideas discussed above, Handy (1995) asserted that “trust needs touch”, implying that co-present, face-to-face contact is essential for trust formation. Although Bos et al.’s findings (2002) support the notion that trust is harder to come by in a technology-mediated context, Zheng et al. (2002) demonstrated that trust could be facilitated with simple, non-technological interventions. In their study, trust, as measured through engagement in a social dilemma task, was found to be built more rapidly when participants met face-to-face before taking part in the task. However, other forms of pre-task intervention were also deemed to be relatively successful, for instance taking part in a “getting to know you” style text-chat session. It would seem that these types of informal ice-breakers can help to break down the barriers prior to en gagement in more formal activities. Other forms of more technologically-based interventions have also been tested. Nguyen and Canny (2007) for instance suggest that spatial distortions, in particular a loss of gaze reciprocity, can negatively impact upon the trust formation process. In their study, the authors compared performance on a cooperative investment task between a spatially faithful VMC set-up (i.e. one which allowed all participants to have a unique and correct perspective as opposed to a shared one which was dictated by the camera) and a “normal,” spatially distorted video system. Findings reveal that trust formation was negatively affected by the spatially distorted set-up in so far as participants displayed less overall trust and were less resilient to breakdowns in trust. These findings support the notion that the manner in which video systems are “normally” set up may distort the perception of non-ver bal conduct, therefore ultimately impacting upon person perceptions.
Further supporting evidence for the influence of VMC on person perceptions comes from Fullwood (2007). Rather than focus on a real-world application in which judgements are made of others (i.e. interviews), Fullwood designed a task with the specific intention of drawing the attention of the participants away from the impression formation aspect of the research. A simple mind-reading task was employed in which participants were requested to “transmit” a list of colours from their mind to the mind of their communication partner, who in turn had to indicate which colour was being transmitted. Participants completed the task either face-to-face or via a video link. The video link set-up employed the use of a monitor with a camera placed directly above it. Participants were requested to sit one metre from the monitor and the scope of the image included the head and shoulders of the participant. After the task had been completed, parti cipants were asked to rate their partner on a number of dimensions using a five-point Likert scale, including mind-reading ability and measures of likeability, intelligence and communication skills. Findings indicate that video-mediated partners were rated less favourably in terms of how likeable and intelligent they were. The author concluded that the technology may interfere with the impression formation process for a number of reasons. For example, for most videoconferencing set-ups it is not possible to make eye contact due to camera placement. As eye contact has been shown to play an important role in impression formation (e.g. see Exline and Eldridge, 1967; Hemsley and Doob, 1978; Thayer and Schiff, 1974; Wheeler et al., 1979) this may then have a knock-on effect for person perceptions via VMC. Moreover, considering that the judgement s of individual traits are often influenced by the perception of an initial trait (e.g. people who are deemed to be more attractive may also be seen as more likeable, friendly etc), there is the potential for a “reverse halo-effect” to come into operation. In other words, there would be a possibility of a domino effect in which the perception of one negative trait may lead to negative perceptions on other dimensions (i.e. someone who is perceived as less credible could also be perceived as less likeable, intelligent, etc.) (Asch, 1946; Thorndike, 1920).
Explaining Poorer Person Perceptions in VMC
In a recent review of VMC and impression formation, Fullwood and Finn (2010) propose a number of potential explanations for why impressions can sometimes be less favourable and less accurate when people are evaluated in a VMC context. First, the quality of the video system may influence the manner in which visual signals are communicated. In systems that have delays or poor visual or audio quality, non-verbal and verbal signals may be degraded, distorted or attenuated. This issue may be of particular relevance in an interview context where technological barriers may also interfere with the feedback process. Interviewees may find it increasingly difficult to judge their performance during the interview due to delayed or distorted feedback and conversely this may affect their confidence (Chapman et al., 2003). Further supporting evidence for the manner in which the distortion of visual signals can impact upon person perceptions comes from Huang (2005). Taking part in an arctic survival task, participants who were perceived to be “taller” due to the manipulation of the distance and angle of the camera were deemed to be more influential in the task than participants who were perceived as “smaller.” Therefore, even when the technology is optimal (i.e. no transmission delays) the visual image may still be distorted and may therefore still influence the manner in which the individual is perceived by others.
A further issue with VMC is appearance consciousness. Fullwood and Finn (2010) argue that users of video-mediated technologies, particularly inexperienced ones, may feel as if they are being observed or monitored. For this reason the user may become more conscious of the manner in which they present themselves and alter their presentation style to a more cautious and reserved one. Finally, Fullwood and Finn (2010) propose that one of the major issues influencing negative perceptions in VMC relates to a difficulty in preserving natural eye contact. Eye contact plays a significant role in impression formation (e.g. see Hemsley and Doob, 1978). People are judged less favourably on a number of dimensions (e.g. honesty, attractiveness and intelligence) when they avoid eye contact. Conversely, people tend to be judged more favourably when making more eye contact. The proble m with VMC is that eye contact cannot be truly preserved due to camera placement. The camera is most often placed above the monitor, and therefore when communicators gaze at their communication partners (i.e. at their image on the monitor) it gives the impression that they are looking in a downward direction. Less favourable impressions may therefore be linked to the perception that individuals are averting their gaze or avoiding eye contact altogether (Storck and Sproull, 1995). Although technological solutions like the one used by Nguyen and Canny (2007) may go some way to addressing this problem, these types of systems are not currently being utilised on a global scale (Fullwood and Finn, 2010).
Advising Users on Good Practice: Video-mediated Gazing
There would appear to be a number of reasons why impressions of others are often less favourable via VMC, however many of these issues would seem to be easily reconciled. One particular non-technological strategy that has been employed previously with a good level success involves the use of video-mediated gazing, or the act of looking into the camera to imitate natural eye contact (Fullwood & Doherty-Sneddon, 2006). Using an actor to pitch two fictitious soap products, Fullwood and Doherty-Sneddon (2006) investigated the effect of video-mediated gazing on information recall. In one condition the actor looked at the monitor and therefore would have appeared to be looking in a downward direction for the entire pitch. In the second condition, the actor looked into the camera at specific points in the pitch; therefore giving the impression to the participant that eye contact was be ing made. Approximately one third of the monologue was accompanied by video-mediated gazing for each of the soap products. Participants were played the pre-recorded sales recitations but were informed that the salesman was talking to them via a live feed. After observing the sales pitch, participants were then given a recall test on both of the soap products. In the video-mediated gazing condition, participants recalled significantly more information about the soap products. This finding is important because it replicates similar findings in a face-to-face context (e.g. Fry & Smith, 1975; Otteson & Otteson, 1980).
One interpretation for this effect is that gazing results in increased physiological arousal. The effects of gazing on information recall can be explained with reference to the Yerkes-Dodson law (1908) which predicts an inverted “u” shaped relationship between arousal and cognitive performance. According to Beattie (1981) high levels of inappropriate gazing result in high levels of arousal, which may in turn disrupt cognitive processing. A relationship has also been found between low levels of arousal and poor task performance, as low arousal levels do not result in significant motivation to complete a task. Moderate levels of gazing, however, may result in moderate levels of arousal, which may help to provide the motivation to learn (Yerkes and Dodson, 1908). It therefore follows that gazing behaviour has an effect on recall as it increases arou sal and thus improves cognitive processing. Also, it may be the case that gazing serves as an indicator that something important is going to happen, therefore serving to reinforce salient points in conversation, in much the same way that eyebrow movements are used to reinforce important aspects of speech (Ekman, 1979; Whittaker and O’Conaill, 1997). Although these findings do not directly relate to impression formation per se, what they do suggest is that one function of eye gaze (i.e. to aid memory recall) can be replicated in VMC. For this reason, one might expect that other psychological effects associated with eye contact (e.g. improved person perceptions) can also be replicated with the use of video-mediated gazing.
The current study therefore aims to investigate whether video-mediated gazing can be used to improve person perceptions and information recall. Furthermore, the investigation addresses whether this practice could be adopted readily by users of the technology with limited opportunity for rehearsal or practice. In the Fullwood and Doherty-Sneddon (2006) study an actor was employed for the role of the salesman and was given ample opportunity to rehearse the pitch and to practice video-mediated gazing. Therefore, will video-mediated gazing be as effective when users are merely advised to adopt this practice, and are not given the opportunity to practice presentations until perfected? To test this research question, participants were split into two conditions: one group was advised to use video-mediated gazing and another group was given no instructions whatsoever. Participants took part in two experiments. The first experiment m easured information recall and required each participant to recount a story to a second participant who was then given a memory test on the information presented. The second experiment measured impression formation and required participants to deliver a sales pitch which they were subsequently judged on by a separate group of participants. It was expected that more information would be recalled and that participants would be perceived more favourably in the video-mediated gazing condition.
A sample consisting of twenty pairs (40 participants in total) of Open University students took part in the first phase of the study. Seven of these participants were male and 33 were female. Participants volunteered to take part in the study and were sampled via a sign-up sheet. Participants were therefore assigned to pairs on the basis of which slot they had signed up for. No details of age were taken. Written informed consent was obtained so recordings of dialogues could be made for further analysis.
A further 20 undergraduate psychology students from a large UK university were sampled for the second phase of the study. These participants were unfamiliar with the Open University participants used in the previous phase. Three of these participants were male and 17 were female. No details of age were taken.
A between groups design was used. Pairs were randomly assigned to one of two conditions: advice or no advice. Participant pairs in the first phase of the study completed two tasks: task 1 (story-telling task), and task 2 (sales tasks). The sales recitations from task 2 were recorded and used in phase 2 of the study. A between groups design was therefore also employed for the second phase of the study. Participants were randomly assigned to watch the sales recitations in either the advice or no advice condition. These separate group of participants were however unaware of the experimental manipulation.
For phase 1 of the study, participant pairs were separated by a screen. On one side of the screen a colour monitor (JVC TM-14EK(B)) was mounted in a wooden box, with a video camcorder (Sony CCD-TR2200EPAL) placed directly above the monitor. A microphone was placed to the right of the monitor. The monitor and camcorder on one side of the screen were connected to the other side of the screen and with precisely the same set-up. The monitors were 14 inches in size. Each participant was distanced approximately one metre from the monitor and the scope of the view included the participant’s face and upper body. As the video link involved a simple connection between 2 sites within close proximity, it was possible to produce a high quality feed with no transmission delay and full lip synchronisation. The equipment was set up in the manner described in an attempt to mimic one-to-one videoconferencing systems utilised in previous studies investigating impression formation (e .g. Bos et al., 2002). Furthermore, this particular set-up was chosen to broadly replicate both set-top/appliance and desktop videoconferencing systems which may be utilised for a variety of functions, including meetings and interviews.
One participant in the pair was given a story on “Vernon the very old wolf” with an accompanying sheet containing all the important points from the story in order of occurrence. The other participant in the pair was given a list containing information on a fictitious soap product and an accompanying list of key points.
For phase 2 of the study, the same set-up as the first task was used, except that in room 2 the video camcorder was used to play back the pre-recorded sales pitches. For the second phase of the study a questionnaire was circulated to all participants. The questionnaire contained 8 questions which were designed to assess their perceptions of the salesperson, product and technology. The questions were rated on a 5-point Likert scale, with one being the lowest possible score, and five being the highest possible score. Scores ranged from “not at all” to “very.” Refer to Table 1 in the results section for the exact wording of each question.
A story was devised about an old wolf called Vernon. One participant from each pair in both conditions (advice and no advice) was given 5 minutes to familiarise themselves with the story with the instruction that they would then need to communicate this story to the other participant in the pair, ensuring that they covered all of the important points on the bullet-point list in the order that they were presented. Furthermore, they were told that they should build the story around these points and simply use them as a memory aid, rather than read directly from the sheet. In condition 1, the participants were advised to use video-mediated gazing behaviours and in condition 2 they were given no advice at all. In the advice condition participants were given a brief introduction to the potential benefits of gazing into the camera (e.g. it would appear to the other participant that they were making direct eye contact with them). Furthermore, they were told that they should at tempt to utilise this strategy to reinforce important points in their speech (i.e. ones which were included in the bullet-point list). Participants in the advice condition were further informed that they should look directly into the camera when making these points but should avoid staring into the camera for too long (rather they should look into the camera for the duration of the point that they were making). Participants in the no-advice condition were simply given instructions on how to complete the task and not given the advice on video-mediated gazing.
After familiarising themselves with the story the participants were given a sheet of paper containing all the important points of the story in order of occurrence and were asked to relay the story to the other participant. The sheet containing the important points in the story was placed on a clipboard in front of the participant so that they could view it any time whilst conversing over the videoconference. The important points in the story were also the points in which the other participant would be tested for recall.
After the story had been relayed the second participant in the pair was given a task of selling a product to the first participant. These participants were either advised to use video-mediated gazing behaviours (given the same instructions as above) or were given no advice at all. Participants were given information concerning a fictitious soap product and were given 5 minutes to familiarise themselves with this information. Again, key points were placed on a clipboard in front of the participant. The participants were also requested to look up at the camera at these key points. After this time participants were asked to sell the product to the other participant. Sales recitations were recorded for later analysis and for phase 2 of the study. After this task had taken place, the participant who listened to the story was given a recall test for the information contained in the story. A total score of 25 points was achievable in the memory test and therefore scores could range from 0 to 25.
For the second phase of the study, a separate group of participants were shown the pre-recorded sales recitations and were asked to rate the salesperson’s performance on a questionnaire. Participants were informed that this was a live video feed but were requested not to interact with the salesperson. After the participants had seen the video, they were requested to complete a questionnaire. After finishing the questionnaire participants were debriefed as to the nature of the deception and were also asked whether or not they were aware that the sales pitch was not actually in fact a live feed. None of the participants indicated that they were aware of this and therefore the deception was deemed successful. However with this said, the non-interactive nature of this second phase could be deemed to be lacking in ecological validity as the technology would rarely be used in this way, and may therefore be seen as a potential weakness of this study.
Finally, using recorded dialogues from phase 1 of the study, assessments were made concerning the length of the sales pitch in each condition, the total amount of time spent gazing at the camera in each condition, the total number of gazes at camera in each condition, and the proportion of time spent gazing at camera in each condition. These measurements were recorded in order to test whether participants had successfully adopted the advice. The length of the sales pitch was recorded from the point at which the salesperson began speaking to the point at which speech concluded. Gazing was said to take place when the participant looked in a forward direction (directly at the camera). In the video recording, this behaviour was detected when, from the coder’s point of view, it appeared as if eye contact was being made. A gaze was measured from the point at which gaze at the camera occurred and concluded at the point when gaze was averted from the camera. Due to the d ifficulty of coding such gazing behaviour an inter-coder reliability test was conducted using one other experienced coder. A sub-sample of 8 participant recordings was used. Using Cohen’s Kappa, agreement for total gaze count was at 98.5% and therefore deemed acceptable.
A series of independent samples t-tests were used to test for differences in recall and evaluations of the salesperson and sales pitch between the advice and no-advice conditions. Summary statistics are provided in Table 1. For the recall task scores could range from 0 to 25. For each of the questions relating to perceptions of the interviewee, task and technology, scores could range from 0 to 5.
A significant difference was found between the conditions on perceptions of how friendly the salesperson was perceived (t(18) = 2.39, p < 0.05). The salesman was perceived as being friendlier when users were advised to employ video-mediated gazing. Participants also indicated a significant difference between the two conditions for the question “did you feel the salesperson was talking to you?” (t(18) = 2.64, p < 0.01). Participants felt that the salesperson was more likely to be speaking to them in the advice condition compared to the no advice condition. A significant difference was also found between the conditions on how much the participants enjoyed taking part in the task (t(18) = 1.78, p < 0.05). Participants enjoyed the task more in the advice condition compared to the no advice condition.
No significant differences were found between the conditions in the number of correct answers on the recall test. Perceptions of honesty, confidence and communicative ability did not differ significantly between conditions. Perceptions of the product also did not differ between conditions, with no significant differences for the ratings of likeliness to purchase product and product uniqueness.
A series of independent samples t-tests were also used to test for differences between total time of sales pitch, total time spent gazing, number of gazes in the sales pitch and proportion of the pitch spent gazing. Summary statistics are provided in Table 2.
A significant difference was found between conditions for the mean total number of gazes in sales pitch (t(18) = 2.22, p < 0.05). Participants in the advice condition used more gazes in total than participants in the no advice condition. A significant difference between conditions was also found in the proportion of the sales pitch spent gazing (t(18) = 2.89, p < 0.01). Participants who were advised to use video-mediated gazing spent a higher proportion of the sales pitch engaged in gazing behaviour than participants who were not advised. No difference was found between the conditions in the total length of the sales pitch in seconds. No difference was also found in the total number of seconds spent gazing, although the advice condition was numerically higher on average.
Findings from this study provide some evidence to suggest that merely advising users to adopt video-mediated gazing is a beneficial strategy. Moreover, this tactic can be adopted with some success by “normal” users of the technology who get limited opportunity for practice. In terms of impression formation, participants rated the salesperson as more friendly and felt that he/she was more likely to be talking to them personally in the advice condition. Furthermore, participants indicated that they enjoyed taking part in the task more so in the advice condition. Although differences between the two conditions on the impression formation aspect of the study were limited to 3 of the 8 questions only, the findings are still interesting nonetheless and imply that merely advising users to adopt this practice can result in more positive impressions. As eye contact has been shown to play a significant role in person perceptions (i.e. see Exline and Eldridge, 1967; Hemsley and Doob, 1978; Thayer and Schiff, 1974; Wheeler et al., 1979) it would seem that the perception of eye contact via a video link can affect judgements of individuals much in the same way as real eye contact does in a face-to-face context.
The study findings however did not entirely support the predictions. The fact that perceptions of honesty, confidence and communicative ability were unaffected by the conditions was somewhat surprising, particularly as increased levels of eye contact is likely to be positively associated with these traits. Although more positive evaluations of the salesperson were evident for some dimensions in the advice condition, it is possible that the full benefits of this strategy have not been realised due to the limited opportunity for interaction in this task. Bos et al. (2002) argue that technology-mediated interactions result in delayed trust formation and this may be one reason why ratings of honesty were modest for both conditions. Participants may not have been given sufficient time in the task to reach a desired level of trust and this could be why some ratings were not as positive as anticipated.
Furthermore, no difference was found between the conditions on the recall task, however in Fullwood and Doherty-Senddon’s study (2006) participants remembered significantly more information in the video-mediated gazing condition than the no-gazing condition. In order to ascertain whether the lack of a recall difference between conditions was a result of participants failing to successfully adopt the advice, the recorded sales pitches were assessed for video-mediated gazing behaviour. Furthermore, it was necessary to test whether participants in the no-advice condition adopted video-mediated gazing behaviours even though they were not advised to. Participants in the advice condition spent on average 31.2% of the sales pitch gazing at the camera. This is similar to Fullwood and Doherty-Sneddon’s study in which approximately one third of the sales pitch was accompanied by video-mediated gazing, therefore indicating that a sufficient level of gazing had occurred. Findings also indicate that participants in the advice condition gazed at the camera more often and spent a higher proportion of the sales pitch gazing at the camera compared to those in the no-advice condition. Even though participants in the no-advice condition did adopt some sort of video-mediated gazing behaviour, it was significantly less frequent than participants in the advice condition. This would seem to suggest that on the surface users adopted the strategy effectively. However, the results also reveal that there was no difference between the conditions in the average amount of time spent gazing at the camera in seconds. This could mean that although participants in the advice condition gazed more often, individual gazes were more fleeting.
There may be a number of further potential explanations as to why the no recall difference was found. First, perhaps there was a potential ceiling effect as both groups of participants recalled approximately on average 9 items correctly out of 25. Although neither group could recall all of the information, they still have similar memory capacities. Indeed, the 9 items recalled by both groups appears to fit nicely with Miller’s (1956) magic number of 7 plus or minus 2. Although the sales pitch by the second participant was initially considered to be sufficient as a distracter task, it would appear that the recall task was too easy and therefore unable to discriminate between the conditions. Second, participants in the no advice condition did gaze at the camera on some occasions. In the Fullwood and Doherty-Sneddon (2006) study, the control group consisted of a condition in which no gazing occurr ed at all. Therefore, it is likely that reduced amounts of gazing are better than no gazing at all. Finally, although adopting a significant amount of gazing behaviour, participants often gazed at inappropriate times. In the Fullwood and Doherty-Sneddon (2006) study, due to memorisation and practice, it was ensured that the salesperson looked at the camera at the exact points in the sales pitch that were later tested for recall. In this study however, the lack of time allocated to memorise and practice meant that the participants, although using a high level of gazing behaviour, did not always look up at the camera at points in the sales pitch that were later tested for recall. Another explanation for this may be because the clipboard containing key points distracted significantly from mutual gaze communication. Furthermore, looking at important points on a clipboard would present less opportunity for gazing into the camera as the y cannot be done simultaneously. An analysis of a sub-sample of 5 of the recordings supports these ideas: only 52.63% of the items on the recall test were accompanied by video-mediated gazing. This may suggest then that if video-mediated gazing were successful that is likely to serve as an indicator that something important is going to happen by reinforcing salient points in conversation, much in the same way that eyebrow movements have been demonstrated to reinforce important aspects of speech (Ekman, 1979; Whittaker and O’Conaill, 1997).
Overall, this study indicates that merely advising participants to adopt video-mediated gazing behaviours leads to some level of success, however training (or indeed a higher level of preparation) may be required in order to realise the full benefits from Fullwood and Doherty-Sneddon’s (2006) study. Furthermore, the fact that some participants naturally use some gaze to the camera would indicate that this behaviour is not necessarily an unnatural response to using the media. One must be cautious however when advising users to adopt this practice. For example, in advising users to gaze at the camera this would also mean that there would be fewer opportunities for these individuals to pick up important non-verbal information that is being communicated by their communication partners. One must therefore weigh up the relative benefits of adopting this strategy (i.e. the potential for a more positive impression) against the potential drawbacks (i.e. limited opportunity to view one’s conversational partner). Additionally, it may also be the case that requiring the communicator to adopt a strategy that they would not naturally use may place an additional cognitive burden on them (at least in the short term). Perhaps then one might argue that video-mediated gazing is more suited to very specific communication tasks, for example those which hinge on putting across a favourable impression (e.g. interviews) or ones in which communication occurs mainly in one-direction or lack reciprocity (e.g. lectures or public speeches). A further caveat is that in advising users of the technology to gaze at the camera, care must be taken to ensure that this does not occur too regularly. Indeed, Doherty-Sneddon et al. (1997) demonstrated that over-gazing during VMC results in longer dialogues and increased time to complete a cooperative problem-solving task. They argue that over-gazing may be distracting and ultimately interferes with cognitive processing much in the same way proposed by Beattie (1981). Further studies in this area may wish to explore the use of video-mediated gazing in a more applied setting, for example distanced interviews, to see if the strategy results in improved perceptions of job candidates. Furthermore, it may also be worthwhile investigating practice effects for this strategy with a longitudinal study. For example, will more profound benefits be realised when users get the opportunity to practice video-mediated gazing over time?
Ambady, N., Bernieri, F. J., & Richeson, J. A. (2000). Toward a histology of social behavior: Judgmental accuracy from thin slices of the behavioral stream. Advances in Experimental Social Psychology, 32, 201–271.
Bos, N., Olson, J., Gergle, D., Olson, G., & Wright, Z. (2002). Effects of four computer-mediated communications channels on trust development. In Proceedings of the SIGCHI conference on human factors in computing systems: Changing our world, changing ourselves (CHI '02). (pp. 135-140). New York, NY: ACM. doi: 10.1145/503376.503401
Chapman, D. S., & Rowe, P. M. (2001). The impact of videoconference technology, interview structure, and interviewer gender on interviewer evaluations in the employment interview: A field experiment. Journal of Occupational and Organizational Psychology, 74(3), 297-298.
Chapman, D. S., & Rowe, P. M. (2002). The influence of videconference technology and interview structure on the recruiting function of the employment interview: A field experiment. International Journal of Selection and Assessment, 10, 185-197.
Chapman, D. S., & Webster, J. (2001). Rater correction processes in applicant selection using videoconference technology: The role of attributions. Journal of Applied Social Psychology, 31(12), 2518-2537.
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(2), 105-125.
Exline, R. V., & Eldridge, C. (1967). Effects of two patterns of a speaker’s visual behavior upon the perception of the authenticity of his verbal message. Paper presented at the meeting of the Eastern Psychological Association, Boston.
Fullwood, C., & Finn, M. (2010). Video-mediated communication and impression formation: An integrative review. In Adam C. Rayler (Ed.), Videoconferencing: Technology, impact and applications. New York: Nova Science, pp. 35-55.
Greenspan, S., Goldberg, D., Weimer, D., & Basso, A. (2000). Interpersonal trust and common ground in electronically mediated communication. In Proceedings of the 2000 ACM conference on computer supported cooperative work (CSCW '00) (pp. 251-260). New York, NY: ACM. doi: 10.1145/358916.358996
Herbsleb, J. D., Mockus, A., Finholt, T., & Grinter, R. E. (2000). Distance dependencies and delays in global cooperation. In Proceedings of the 2000 ACM conference on computer supported cooperative work (CSCW '00) (pp. 319-328). New York, NY: ACM. doi: 10.1145/358916.359003
Nguyen, D. T. & Canny, J. (2007). Multiview: Improving trust in group video conferencing through spatial faithfulness. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI ’07) (pp. 1465-1474). New York, NY: ACM. doi: 10.1145/1240624.1240846
Sellen, A. (1997). Assessing video mediated conduct: A discussion of different analytic approaches. In K.E. Finn, A.J. Sellen & S.B. Wilbur (Eds.), Video Mediated Communication. Mahwah, NJ: Lawrence Erlbaum.
Straus, S. G., Miles, J. A., & Levesque, L. L. (2001). The effects of videoconference, telephone, and face-to-face media on interviewer and applicant judgments in employment interviews. Journal of Management, 27(3), 363-381.
Teasley, S. D., Covi, L., Krishnan, M. S., & Olson, J.S. (2000). How does radical collocation help a team succeed? In Proceedings of the 2000 ACM conference on computer supported cooperative work (CSCW '00) (pp. 339-346). New York, NY: ACM. doi: 10.1145/358916.359005
Whittaker, S., & O’Conaill, B. (1997). The role of vision in face to face and mediated communication. In K. E. Finn, A. J. Sellen & S. B. Wilbur (Eds.), Video Mediated Communication. Mahwah, NJ: Lawrence Erlbaum.
Zheng, J., Veinott, E., Bos, N., Olson, J. S., & Olson, G. M. (2002). Trust without touch: Jumpstarting long-distance trust with initial social activities. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI ’02), (pp. 141-146). New York, NY: ACM. doi: 10.1145/503376.503402
Copyright 2013 Communication Institute for Online Scholarship, Inc.
This file may not be publicly distributed or reproduced without written permission of