How We See Changes How We Feel: Investigating the Effect of Visual Point-of-View on Decision-Making in VR Environments
The immersive and embodied nature of virtual reality (VR) has seen this medium applied to numerous research projects that consider how humans might react in simulated scenarios. For example, a common academic application is in using VR experiences as a means to promote prosocial outcomes, inspiring empathy through placing people in realistic environments where their decisions have seemingly authentic consequences on others (i.e. the virtual agents). Even though numerous pieces of work have explored this application of VR, more needs to be done to investigate how specifically we can effectively design VR experiences that feel meaningful and can induce empathy.
In our exploratory work, we considered one dimension that we believed might affect people’s feelings in this regard — their point-of-view (POV) in the VR world. Although VR experiences are almost always shown from the first-person POV (1PP), there have been experiences that use a third-person POV (3PP) as well. We wanted to study how being placed in different POVs in the context of a moral dilemma might affect 1) what ethical decisions people make and 2) how people feel about those decisions.
What Did We Do?
We built a VR experience that takes users through repeated scenarios of a commonly-studied moral dilemma: a car driving up to an intersection loses control and users can choose to swerve or stay in the same lane; for either choice, some group of (virtual) people will perish. We varied these scenarios in terms of utilitarianism (saving more versus fewer people), passenger/pedestrian bias (saving passengers + driver versus pedestrians), and interventionism (continuing in the same lane versus swerving).
To compare whether people would make different decisions based on different POVs, we performed a between-subject study in which each participant was shown these scenarios from one of three perspectives — from the driver’s eyes (first-person/1PP), from a view above the driver but following them (third-person/3PP), or from a static, unmoving perspective (static/SP). To compare how people felt in terms of empathy towards the virtual agents and the meaningfulness of the scenarios based on different POVs, we performed a within-subject study in which participants were shown extra scenarios from the two remaining perspectives.
Our findings revealed something potentially intriguing — that even though people’s dimensions of moral judgment across the three POVs barely showed any significant differences (as reflected in their choices), their self-reported, subjective feelings of empathy and meaningfulness clearly did.
confidence intervals. The dashed line indicates a one-vs-rest comparison between the 1PP metrics versus the
combined 3PP and SP metrics.
Extending This Finding
We wanted to dig deeper into this specific finding, so we performed a follow-up study focussed on this hypothesis. This time, each individual user was shown common scenarios from all three POVs and answered a longer survey based on more specific questions that tied to empathy and meaningfulness. This more thorough analysis allowed us to affirm that people’s subjective feelings did change across POVs. We found that people’s feelings regarding the scenarios were statistically different across four key questions, relating to emotional resonance, focus, and subjective meaningfulness. The 1PP perspective generally elicited the highest values for these metrics, followed by 3PP and then SP.
significant differences across the mean metrics across the 3 groups. Error bars depict 95% confidence intervals.
Similarly to the findings before, despite clear disparities in internal feelings, we found that people’s actual decisions did not show much change across POVs — on average, only about one change (out of a potential six) was made regarding their decisions across each of the POVs.
So Why Is This Important in the End?
Our finding that different perspectives can induce different levels of meaningfulness and empathy towards characters is useful to VR developers. Knowing how different perspective affects levels of perceived meaningfulness and emotional resonance can open the door for designers and developers to better orchestrate how they want players to engage with VR experiences.
Perhaps more important, however, was our unexpected finding — that focussing squarely on peoples’ decisions fails to disentangle the full scope of human complexity and moral judgment when presented with an ethical dilemma. Thus, researchers designing reflective and meaningful experiences should perhaps not only consider the final decisions that humans make but also what introspective factors go into making them as well.
Please check out our full paper here.
