Environment & Stress & Performance

Master Thesis

This thesis explores how situational stressor design in virtual environments influences user engagement and cognitive efficiency. The project compares two task-integrated VR experiences—an interview setting designed to evoke stress and a more relaxed coffee shop environment—while participants complete a series of cognitive tasks. Through environmental design, NPC interaction, user surveys, and task-performance data, the study examines how stress affects sensory, emotional, cognitive, and behavioral engagement. The findings suggest that thoughtfully designed VR environments can successfully evoke situational stress and increase cognitive engagement, while potentially reducing users’ cognitive efficiency.

Master’s Thesis, Northeastern University, 2024

Player Experience in Task-Integrated VR Applications: Environmental Stressor Design in Relation to User Engagement and Cognitive Efficiency

https://doi.org/10.17760/D20662898

1. Problem

cognitive tasks within environments that may naturally evoke stress. However, stress is often treated as a consequence of the task itself, while the influence of the virtual environment—including its spatial atmosphere, lighting, materials, sound, social context, and character interactions—is less frequently examined.

This project was motivated by the question of whether environmental design can actively shape users’ psychological and cognitive experiences in VR. By investigating how situational and environmental stressors influence engagement and task performance, the study aims to support more intentional design decisions: reducing unnecessary stress in learning and rehabilitation applications, while realistically recreating pressure in simulations designed for training and stress management.

2. Related Work: Environmental Stress and Outcomes

2.1. Environmental Stress

Environmental stress theory suggests that people’s emotional and physiological responses are shaped by both situational demands and environmental stimuli. In task-oriented settings, stress may arise from performance expectations, uncertainty, limited control, and the perceived consequences of failure. At the same time, characteristics of the surrounding environment—such as noise, crowding, lighting, materiality, color, spatial openness, and access to nature—may intensify or reduce this experience.

This research informed the project’s distinction between two forms of stressor design: situational stressors, created through the context and demands of the experience, and environmental stressors, embedded in the sensory and architectural qualities of the virtual space.

2.2. VR as an Emotional Environment

The immersive and spatial nature of VR allows users to experience virtual situations as psychologically meaningful environments rather than as content viewed from a distance. Through presence and sensory involvement, users may respond to virtual settings in ways that resemble their reactions to real-world environments.

This makes VR valuable not only for simulating tasks, but also for intentionally constructing emotional conditions. Environmental cues, social interactions, and narrative context can therefore be used to evoke different levels of comfort, tension, urgency, or stress within a controlled setting.

2.3. User Engagement

User engagement is multidimensional rather than a single measure of whether someone “likes” an experience. In this study, engagement was examined through four dimensions:

Sensory involvement, including immersion and presence
Emotional engagement, including emotional valence and arousal
Cognitive engagement, including mental effort and perceived task demand
Behavioral engagement, represented through focus and flow

This framework made it possible to examine whether stress affected different aspects of the user experience in distinct ways, rather than assuming that higher stress would uniformly increase or decrease engagement.

2.4 Stress and Cognitive Efficiency

Stress can alter attention, working memory, inhibition, problem-solving, and decision-making. While moderate challenge may encourage greater mental effort, increased cognitive engagement does not necessarily lead to better task performance. Stress may require users to invest more effort while simultaneously reducing the speed or accuracy of their responses.

This tension between engagement and efficiency became a central focus of the project: whether a more stressful VR environment would make users feel more cognitively involved, while also making it harder for them to perform cognitive tasks efficiently.

3. Research Questions

H1: Stressor designs in VR can deliver stress to players.

H2: Stressor designs in VR will enhance users’ emotional engagement.

H3: Stressor designs in VR will enhance users’ cognitive engagement.

H4: Stressor designs in VR will hinder cognitive efficiency.

H5: Engagement level positively correlates with cognitive efficiency.

H6: Neuroticism positively correlates with reported stress level.

4. Experiment Design

Participants & Experimental Procedure

Thirty participants between the ages of 18 and 44 were recruited from Northeastern University. All participants had prior experience with, or anticipated future experience in, academic or job interviews so that the interview scenario would feel personally relevant.

Participants were screened for eligibility before the study and were randomly assigned to one of two VR conditions: the stressful interview environment or the low-stress coffee shop environment. Each participant received $10 for approximately 30 minutes of participation.

Each study session followed the same sequence:

- Participants completed a demographic survey and cognitive screening.
- The researcher introduced the VR controls and response procedure.
- Participants were fitted with a Meta Quest 2 headset.
- After entering the virtual environment, participants closed their eyes for 30 seconds and then spent two minutes adapting to the scene.
- Participants completed the same five cognitive tasks while interacting with the NPC.
- After the VR experience, they completed a post-experience survey measuring perceived stress and engagement.

Reaction time, response accuracy, survey responses, and personality measures were collected for later analysis.

4.1 The System: Translating Stress into Environmental Stress

To investigate how environmental stressors shape user experience and cognitive performance, I designed two task-integrated VR environments with contrasting emotional conditions: a formal interview room and a relaxed coffee shop.

The two environments shared the same approximate floor area, ceiling height, spatial complexity, lighting intensity, and task structure. This allowed the study to isolate the effects of selected environmental and situational cues while maintaining comparable spatial conditions.

The interview environment was designed to evoke mild situational stress. It combined a high-pressure social context with darker colors, synthetic materials, cool lighting, limited biophilic elements, mechanical background sounds, and a formally dressed NPC with a serious tone of voice.

In contrast, the coffee shop environment was designed as a low-stress condition. It used warmer lighting, natural materials, lighter colors, outdoor views, biophilic elements, ambient sounds of birds and wind, and a casually dressed NPC with a friendly conversational style.

Both environments included the same sequence of cognitive tasks, allowing differences in stress, engagement, and cognitive efficiency to be examined in relation to the surrounding virtual environment rather than the task content itself.

Cafe

Detailed comparison & design table see:

Master’s Thesis, Northeastern University, 2024
Player Experience in Task-Integrated VR Applications: Environmental Stressor Design in Relation to User Engagement and Cognitive Efficiency
https://doi.org/10.17760/D20662898

Conference Room

4.2 The Tasks

Rather than treating the VR environment as a passive background, the cognitive tasks were embedded directly into each scenario as part of the participant’s interaction with the NPC. To evaluate cognitive performance across the two VR conditions, I integrated five short cognitive tasks into each environment. The tasks were selected to engage multiple executive functions, including working memory, attention, inhibition control, and problem-solving.

Participants completed the same tasks in the same sequence, regardless of the environment they experienced. Keeping the task content consistent allowed differences in performance to be examined in relation to the surrounding VR condition rather than differences in task difficulty.

The task set included:

Backward Digit Span: recalling a sequence of numbers in reverse order to assess working memory and attention.
Running Working Memory: identifying whether a set of letters appeared at the end of a previously presented sequence.
Compound Arithmetic: solving a multi-step equation to assess problem-solving and attention.
Continuous Subtraction: repeatedly subtracting a given number while maintaining intermediate results in working memory.
Stroop Task: identifying font colors while suppressing interference from the written color words.

Together, these tasks provided a controlled way to compare how participants processed information and performed under different levels of environmental and situational stress.

Detailed tasks description & rationale see:

4.3 The Measurements

4.3.1 Surveys

To capture participants’ subjective experiences, a set of survey measuring perceived stress and four dimensions of engagement was developed. The engagement items were adapted from established instruments, including the User Engagement Scale, the Extended Reality Presence Scale, and the NASA Task Load Index. Each dimension addressed a different aspect of the participant experience:

Sensory involvement: immersion, presence, and awareness of the surrounding virtual environment
Perceived stress: stress associated with the scenario, cognitive tasks, environmental design, and NPC interaction
Emotional engagement: positive/negative valence and arousal, eg. frustration, discouragement, satisfaction, and accomplishment
Cognitive engagement: mental demand / cognitive efforts
Behavioral engagement: engagement flow

4.3.2 Reaction Time

To measure the cognitive efficiency in the VR experience, a custom reaction-time logging system in Unity was designed.

For each task, the NPC presented the question and then gave a verbal signal—“Please answer”—to establish a consistent starting point. Participants responded verbally and used an “OK” hand gesture to indicate that they had finished. The researcher then pressed the controller trigger, prompting Unity to record the elapsed response time.

Audio recordings were used to verify the timing data and participant responses. Reaction time was combined with response accuracy to calculate an Inverse Efficiency Score (IES), allowing performance speed and accuracy to be evaluated together. A lower IES represented faster and more accurate task performance.

4.4 Data Analysis Framework

The analysis combined self-reported survey data with behavioral performance data to examine how the two VR environments influenced stress, engagement, and cognitive efficiency.
First, participants’ perceived stress scores were averaged and compared between the interview and coffee shop conditions to evaluate whether the environmental and situational stressors successfully created different levels of stress.
Engagement was analyzed across four dimensions—sensory involvement, emotional engagement, cognitive engagement, and flow—as well as an overall engagement score. Average scores were calculated for each dimension and compared between the two groups.
Cognitive efficiency was evaluated using participants’ reaction time and response accuracy. These measures were combined into an Inverse Efficiency Score (IES), calculated by dividing average reaction time by accuracy. Lower IES values indicated faster and more accurate performance.
Finally, correlation analyses were conducted to examine whether engagement was associated with cognitive efficiency, and whether participants’ neuroticism scores were related to their reported stress levels.

5. Results

H1. Perceived Stress Level Between Groups
To test out Hypothesis 1, which suggests that Stressor designs in VR can deliver stress to
players. A between-group T-test was done. Results showed that participants in the experimental
group (Conference room, Group E) were observed to have significantly higher stress level feedback than the Control
Group (Cafe, Group C), with t = - 3.82, df = 28, p < 0.01.

H2/3. Engagement Level Between Groups

To examine Hypothesis 2 and 3, namely stressor designs in VR will enhance users’ emotional and cognitive engagement, process 3 and 4 were conducted.
After calculating the average of each item, the participants' feedback scales on engagement were all normally distributed. A between group T-test was done to look for differences in all four aspects and the overall level of engagement feedback between the two groups. Four aspects including Sensory Involvement (SI), Emotional Engagement (EE), Cognitive Engagement (CE)
and Flow (FL) are each tested as an independent variable. The overall engagement level (Engagement) was derived from the average of all four aspects above.

In the T-Test, no significant differences were found in various metrics, including SI, EE, FL and overall engagement level. However, according to the results, the experimental group was similar to or slightly lower than the control group in all aspects of engagement feedback. In the CE feedback, with p value < 0.001, the cognitive engagement feedback of the experimental group was significantly higher than that of the control group.

H4. Cognitive Efficiency Between Groups

An independent T-test was done to evaluate IES score, which represents the cognitive efficiency differences between two groups. The T-test presented significantly lower scores of IES among the control group, with t = -2.214, df = 28, p = 0.017, which suggested higher cognitive efficiency.

H5. Engagement and Cognitive Efficiency

A Pearson’s Correlation analysis was conducted to examine the relationship between engagement level and cognitive efficiency. With r = 0.237 and p = 0.897, no significant signs showed that there’s a correlation between engagement level and IES (Left).
In this context, a negative correlation was expected to be found, which means that higher engagement level results in lower IES score. As a result, another correlation analysis was done between cognitive engagement level and cognitive efficiency, based on the closer relation between the two. With r = 0.247 and p = 0.187, the correlation between overall engagement and IES was not significant (Right).

H6. Neuroticism and Stress

A correlation test was conducted to evaluate the relationship between participants’ neuroticism scores and reported stress levels. With r = -0.149 and p = 0.432, a weak negative correlation was found.

6. Discussion

The findings demonstrate that VR environments can be intentionally designed to evoke distinct psychological conditions. Participants in the interview environment reported significantly higher stress than those in the coffee shop condition, suggesting that situational context, environmental cues, and NPC behavior worked together to communicate a recognizable level of real-world pressure.

However, increased stress did not improve engagement uniformly. Participants in the stressful condition reported significantly higher cognitive engagement, indicating that they invested more mental effort in completing the tasks. At the same time, no significant differences were found in sensory involvement, emotional engagement, flow, or overall engagement. This distinction suggests that requiring more cognitive effort does not necessarily make an experience more immersive, enjoyable, or engaging as a whole.

The stressful environment also reduced cognitive efficiency. Participants performed the tasks more slowly and less efficiently, even though they reported greater cognitive engagement. This reveals an important tension between effort and performance: users may feel more mentally involved while simultaneously becoming less effective at completing the task.

No significant relationship was found between overall engagement and cognitive efficiency, suggesting that subjective engagement should not be treated as a direct indicator of successful performance. The study also found no meaningful relationship between neuroticism and reported stress, indicating that the designed environment and scenario may have played a stronger role than this individual personality measure within the study.

Overall, the findings emphasize that stress should be treated as an intentional design variable in task-oriented VR. Learning and rehabilitation applications may benefit from reducing unnecessary environmental pressure, while simulations for interviews, emergency response, or professional training may require carefully controlled stress to recreate realistic performance conditions.