Up,Down,On,Off? Effects of Camera Decisions on Interpersonal Perception During Videoconferencing

Study Motivation

The use of videoconferencing grew rapidly during the COVID-19 pandemic (Iqbal, 2022) and is expected to remain popular in the coming years (Ozimek, 2020). With videoconferencing becoming commonplace, users seek to discover how to best appear on camera or determine whether one’s camera even needs to be on (see, e.g., Baterna, 2022; Goin & Elias 2020; Mathis, 2020; Thompson, 2021). During videoconferencing, the user has many decisions regarding how they will appear. For example, they can either turn their camera on or off, , mute or unmute themselves, change the angle of their camera’s angle, or sit close to or far from their camera.

Most of the research regarding camera positioning was conducted before the COVID-19 pandemic (e.g., Bekkering, 2004; Bohannon, 2013, Gray 2013; Thomas & Pemstein, 2015) (see, e.g., Fauville et al., 2022 for a more recent study on this topic). Furthermore, scant research has examined how turning one’s camera off during a videoconference impacts interpersonal processes (for one recent study, see Park & Whiting, 2020). Most recent work on camera usage focuses on the personal benefits of turning one’s camera off (see, e.g., Bailenson, 2021; Fauville et al., 2021). Thus, there is a gap in the research regarding how self-presentation during videoconferencing, such as camera usage and angle, influences interpersonal perception.

Interpersonal Perception in Videoconferencing

In human interaction, how someone looks, sounds, and behaves shape how other will perceive them. These perceptions give clues about someone’s emotional state, attitudes, and abilities (Kenny, 2020). Interpersonal perception matters because how someone is perceived can influence how others interact with them interaction (Ferguson & Bargh, 2004). The three types of interpersonal perception discussed in the present study are perceived competence, effort, and experience.

Perceived competence reflects an individual’s perceived ability, intelligence, skill, creativity, and efficacy (Fiske et al., 2006). In the workplace, perceived competence predicts of trust (Mayer et al., 1995) and respect (Oleszkiewicz & Lachowicz-Tabaczek, 2016). Effort is a moralized trait in Western cultures. Individuals exerting more effort are praised (Ogletree et al., 2013) and may be perceived as more honest and virtuous (Amos et al., 2019). In the present study, we defined perceived experience how much task-specific experience or knowledge one has.

Present Study

The primary aim of the present study was to assess the effects of camera usage and angle on perceived power during a videoconference. However, we included several exploratory dependent variables (perceived competence, effort, and experience (see Gruber et al., 2023). These exploratory variables are the focus of the present paper. Because these variables were exploratory, no hypotheses were advanced. However, the study sought to gain initial data regarding the following three research questions:

Participants

We recruited two hundred and twenty-eight participants using an online research participation system at a large, southeastern university. Participants had an average age of 20.39 years (SD = 2.91). Most participants were female (57.0%) and White/Caucasian (68.4%). The majority (76.7%) spent between zero and two hours per week using videoconferencing for school or work. The study was reviewed and approved of by the University’s Institutional Review Board.

Design

The present study employed a 2x2x2 between-subjects experimental design. The first independent variable was camera usage, referring to whether both characters in the vignette had their cameras on or if only one of the character’s had their camera on. The second independent variable was camera angle. This refers to whether a character’s camera was placed above their monitor (high angle) or below (low angle). Both characters had high angles, or one character had a low angle while the other one had a high angle. The third independent variable was actor, as it was necessary to account for the fact that the actors themselves could influence the results. Thus, the two actors (referred to here as A.B. and M.G.) took turns playing each role. Participants were randomly assigned to one of the eight conditions.

Materials

Vignettes

In this study, participants watched scripted video vignettes of two characters, “Mary” and “Sarah,” completing the NASA Moon Survival task (Hall & Watson, 1970). Two members of the research team served as actors in the study. In the exercise, Mary and Sarah were ostensibly given the hypothetical scenario of being stranded on the Moon with a list of intact supplies. They had to rank the items in order of importance for survival, first as individuals and then as a pair. In the present study, participants were told that Mary and Sarah had done their individual rankings off camera, and that they would be watching a video of them completing the group negotiation on a videoconference. In the video, each character contributed the same number of items to the final ranking. This meant that that social judgments could be based on the manipulated variables (camera usage and angle) rather than the outcome of the exercise.

The videos were approximately four minutes long in each condition. They were recorded in an empty office using external webcams and an external microphone. See Figure 1 for the recording setup. After recording, the raw footage was edited so that the characters appeared to be on a Zoom call. See Figure 2 for screenshots of what the final videos looked like in each condition.

OPEN IN VIEWER

Figure 1.

Recording Setup.

OPEN IN VIEWER

Figure 2.

Video Screenshots.

Procedure

Participants accessed the study on Qualtrics at a time and location of their choosing. After consenting to participate, they watched their assigned video. Next, they completed the subjective assessments. Finally, they completed demographic information and were thanked for their time.

Perceived Experience

There was a significant two-way interaction between camera placement and actor on perceived experience, F(1, 219) = 6.98, p = .009, η_p² = .03. A.B. played Mary, follow-up t-tests did not indicate significant differences between when Mary had the high angle (M = 4.42, SD = 1.40) and the low angle (M = 3.98, SD = 1.43), t(110) = 1.66, p = .100, d = .31. When M.G. played Mary, there was also no significant difference between when Mary had the low angle (M = 4.38, SD = 1.36) compared to the high angle (M = 3.91, SD = 1.27), t(113) = -1.90, p = .060, d = .35. However, the effect sizes suggest the presence of effects potentially if the sample was larger, the manipulations were stronger, or if the measure was more sensitive. See Figure 3 for a visualization of these results.

OPEN IN VIEWER

Figure 3.

Perceived Experience.

Perceived Competence

There was a significant three-way interaction between camera usage, angle, and actor on perceived competence, F(1, 219) = 6.66, p = .01, η_p² = .03. This interaction was followed up using two-way ANOVAs to examine a potential camera usage x angle interaction across the two actor conditions. There were no significant interactions of main effects when A.B. played Mary. But, when M.G. played Mary, there was a significant two-way interaction between camera usage and angle, F(1,111) = 4.18, p = .043, η_p² = .04. Follow-up t-tests looking at the data when both cameras were on indicated no significant differences between when Mary was captured by the low angle (M = 4.57, SD = 0.95) and when Mary was captured by the high angle (M = 4.10, SD = 0.84), t(51) = -1.89, p = .065, d = .52. However, note the medium effect size. There was no significant difference in perceived competence when only Mary had her camera on and she had the low angle (M = 4.03, SD = 0.90) vs. the high angle (M = 4.89, SD = 0.52), t(60) = 0.83, p = .205, d = .21. There was also a significant main effect of camera usage, F(1,111) = 6.06, p = .015, η_p² = .05. Overall, Sarah appeared more competent when both cameras were on (M = 4.30, SD = 0.91) than when her camera was off (M = 3.95, SD = 0.71). See Figure 4 for a visualization of these results.

OPEN IN VIEWER

Figure 4.

Perceived Competence, M. G. as Mary.

Perceived Effort

There was a two-way interaction between camera usage and actor on perceived effort, F(1, 219) = 9.21, p = .003, η_p² = .04. Follow-up t-tests showed that when M.G. played Mary, there was a difference in perceived effort depending on camera usage. When both characters had their camera’s on, Sarah was perceived as putting in more effort (M = 4.24, SD = 0.73) than when her camera was off (M = 3.71, SD = 0.84), t(113) = 3.62, p < .001, d = .68. When A.B. played Mary, there was no difference in perceived effort when both cameras were on (M = 3.85, SD = 1.00) and when only Mary had her camera on (M = 4.02, SD = 0.85), t(11) = -0.95, p = .346, d = 0.18. See Figure 5 for a visualization of these results.

OPEN IN VIEWER

Figure 5.

Perceived Effort.

Perceived Experience

There was a two-way interaction between actor and camera angle. Follow-up analyses did not find significant differences. However, the effect sizes indicate that real effects between these variables may be present, especially if one-side alternative hypotheses and more sensitive measures are used. Trends in the data suggest that when A.B. played Mary, Sarah was perceived to be more powerful when Mary had the high angle. But, when M.G. played Mary, the relationship between camera angle and experience was reversed. Now, Sarah was perceived as having more experience when Mary had the low angle. Thus, having a low camera angle either help or hindered perceived experience, depending on who was playing each character. The low angle hurt A.B. but helped M.G to appear more experienced.

Perceived Competence

There was a three-way interaction between actor, camera angle, and camera usage on perceived competence. Specifically, there were only effects of camera usage and angle when M.G. played Mary. When both cameras were on, Sarah was perceived as being more competent when Mary had the low camera angle. The follow-up test for this effect was not significant, but the medium effect size indicates that a real difference exists. Thus, having the low angle seemed to negatively impact Mary’s perceived competence. When only one camera was on, there was no difference in perceived competence between the high and low angles. These results suggest that having a low camera angle negatively impacts perceptions of competence, but only when the visuals of both actors can be compared. When Sarah’s camera was off, it did not matter that Mary had the low camera angle. Because the average perceived competence score stayed about the same in both high angle conditions, we can assume that Sarah’s camera being off is what drove the decrease in her perceived competence in the low angle condition. This assumption is further supported by the observed main effect of camera usage. Overall, Sarah was perceived as being more competent when both cameras were on than her camera was off. Because competence is a known antecedent of trust (Mayer et al., 1995), and trust has a fundamental role in human interaction, this finding is particularly interesting and relevant.

Perceived Effort

There was a two-way interaction between camera usage and actor on perceived effort. When A.B. played Mary, there was no difference in perceived effort between the camera usage conditions. However, when M.G. played Mary, Sarah was perceived as putting in more effort when both cameras were on compared to when her camera was off. Thus, Sarah turning her camera off seemed to convey that she was putting in less effort into the task. This result extends the research on camera usage. Park and Whiting (2020) found that there is a negative bias toward individuals who turn their cameras off, though this effect depends on the perceived reasons for turning the camera off. In the present study, it seems that turning one’s camera off was negatively perceived, though the effect was moderated by who the individual was that turned their camera off.

Limitations

There were several limitations of the present study. Because the present study gathered initial, exploratory data on the outcome variables, only single-item measures were used. Single-item measures may not as readily capture complex constructs compared to multi-item measures, which may explain the non-significant follow-up tests (Loo, 2002). Another limitation was that only female actors were used. Perceptions of nonverbal behavior sometimes depend on the gender of the individual (Cook et al., 2023; Henley & Harmon, 1985). For instance, Cook and colleagues (2023) found that, overall, female actors in a videoconference were perceived as more trustworthy than male actors. Thus, the results of the present study may not extend to male characters or mix-sex groups. Additionally, participants were outside observers of the interaction. Burgoon and Newton (1991) suggest that outside observers in interpersonal perception research tend to view interactions more favorably than active participants. Thus, the results of the present study may not generalize to research that has participants complete the social negotiation task themselves. Furthermore, this limitation means that the results of the present study may not generalize to real-world instances of videoconferencing.

Future Research

The limitations of the present study pave the way for future research. First, future studies ought to include multi-item measures of interpersonal perception measures. Additionally, future work could also include objective measures of interpersonal perception, such as use intent (as in Bekkering, 2004). For instance, participants could indicate which character they would want to work with in the future on a similar task. Future research ought to also replicate the study with male actors or mix-sexed dyads to examine the presence of gender effects. Finally, the present study ought to be replicated in the laboratory with participants completing the task with a partner. This would allow for more external validity, as video may not readily capture the dynamic nature of interpersonal interaction (Burgoon & Newton, 1991).