Technology has potential to support older adults’ Instrumental Activities of Daily Living (IADLs), activities crucial for independence, such as managing finances, household tasks, and healthcare activities (Charness and Boot, 2009; Horgas and Abowd, 2004; Ramprasad et al., 2019). However, there has been considerably less attention paid to the potential of technologies to support older adults’ participation in enrichment activities, or as Rogers et al. (1998) called them, Enhanced Activities of Daily Living (EADLs). EADLs encompass hobbies, leisure, entertainment, and relaxation activities. A lack of attention to technology to support EADLs has the potential to “lock out” many older adults from enjoyable and rewarding experiences and is further unfortunate because EADLs may play a significant role in the health, well-being, and quality of life of older adults (Hughes et al., 2010; Kuykendall et al., 2015; Menec, 2003).

A host of barriers can possibly preclude meaningful interactions with technology to support IADLs and EADLs when technology is not designed considering the needs, preferences, abilities, and experiences of older user and when older adults are not included in the design process (Czaja et al., 2019). By understanding older adults’ perceptions of technology and barriers to adoption, technology can be better designed to encourage technology adoption and use to benefit all activities of daily living, including EADLs (Charness and Boot, 2009). The current report investigated how older adults interacted with a relatively novel—and increasingly prevalent—form of technology to support EADLs: consumer immersive virtual reality (VR) technology. ¹ Of particular interest was whether older adults psychologically experienced presence differently within virtual environments compared to younger adults, as this is one of the defining characteristics of this technology, and potential barriers to use as assessed by measures of workload and system usability.

Recent literature on VR to support IADL activities of older adults suggests that cybersickness, system workload and usability issues (i.e., frustration, fatigue, discomfort), and lack of experience with technology are potential barriers relevant to older adults’ adoption of VR technology (Appel et al., 2020; Huygelier et al., 2019; Lee et al., 2019; Saredakis et al., 2020; Winter et al., 2021; Yen and Chiu, 2021). These issues are also likely relevant to the use of VR to support EADL activities among older adults. Additional insights into VR to support EADL activities may be gained from investigations of older adults’ interactions with another common technology to support leisure: non-immersive video games. The difference between younger and older adults with respect to video game adoption and gameplay is likely due to differences in perceived needs, attitudes, perceptions, and usability barriers. In the United States, only 24% of older adults (ages 65+) report playing video games, compared to 60% of the youngest adult cohort (18–29 years old, Brown, 2017). While 30% of younger adults report playing video games frequently, only 11% of older adults report the same. This may in part be due to video games being designed by younger people for younger people, not considering the knowledge and abilities of older adults (for review, see Boot et al., 2020). For example, the challenge that video games provide may be appropriate for, and enjoyable by, younger gamers, but may frustrate older adult gamers experiencing age-related perceptual and cognitive changes. Game design can also fail to consider that inexperienced older gamers may lack the mental models required to succeed in video games (Harrington et al., 2017). Finally, attitudinal barriers may also play an important role with some older adults perceiving that video games are unimportant or are afraid that they might be perceived as childish by their peers for engaging in video game play (De Schutter et al., 2015). These challenges may also exist when older adults interact with immersive VR and perhaps prevent them from reaping the various benefits of such technologies.

The goal of the current investigation was to better understand older adults’ perceptions of VR environments and VR technology. Of primary interest was sense of presence as this is “the defining aspect of a successful VR experience” (Weech et al., 2019). We define “presence” as the “subjective sense of being in a virtual environment” (Schubert, 2003). Measures of workload were also collected to assess potential usability barriers when navigating virtual environments as current literature suggests this to be of relevance pertaining to older adults’ adoption of VR technology (Appel et al., 2020; Huygelier et al., 2019; Lee et al., 2019; Saredakis et al., 2020; Winter et al., 2021; Yen and Chiu, 2021).

Finally, cybersickness is still a relevant phenomenon to consider in the use of modern consumer VR technology (Yildirim, 2020). Because cybersickness has been identified as a potential barrier to the use of virtual reality in older adult populations (Seifert and Schlomann, 2021), perhaps more so for older adults compared to younger adults (Diersch and Wolbers 2019), it was also assessed. Age effects were of interest in part due to literature suggesting that older adults are more sensitive to simulator or cybersickness (Brooks et al., 2010; Kawano et al., 2016; Keshavarz et al., 2018). However, the relationship between age and cybersickness may be complex, as indicated by a recent meta-analysis finding that participant samples that were on average under the age of 35 generally report higher sickness compared to samples over the age of 35 (Saredakis et al., 2020). Gender may be another relevant factor to consider in the evaluation of VR experiences as older women may be especially susceptible to simulator/cybersickness (Classen et al., 2011; Park et al., 2006; Matas et al., 2015). However, it should also be noted that gender effects in the literature are inconsistent (for review, see Weech et al., 2019).

Forty adults living in the Tallahassee region participated: twenty younger adults (M = 19.7 years, SD = 1.25; 14 females) ² and twenty older adults (M = 71.3 years, SD = 7.41; 9 females). As this was a thesis project, sample size was largely determined by available time and resources. Younger adults were undergraduate students and were recruited through the psychology participant pool at Florida State University. Older adults were recruited through the Institute for Successful Longevity’s (https://isl.fsu.edu/) participant registry. This registry contains contact information for community members in Tallahassee and surrounding areas who have indicated interest in participating in research. All participants were compensated $20 for their participation in a session that lasted approximately 2 hours. Consent was obtained by having participants read and sign a printed consent form after the nature of the study was explained to them. This study complied with APA ethical standards in the treatment of human participants and was approved by the Florida State University Human Subjects Committee (HSC No. 2018.25327).

All supporting data are located on OSF (https://osf.io/vkxde/?view_only=5e36543a3891498da1eeb9a3495698b2). Consistent with previously reported age differences (Boot et al., 2015; Roque and Boot, 2018), older adults were significantly less technology proficient compared to younger adults. Computer Proficiency Questionnaire (CPQ) scores range from 6 to 30, and Mobile Device Proficiency Questionnaire (MDPQ) scores range from 8 to 40, with higher scores representing higher technology proficiency. An ANOVA was conducted on CPQ scores, corresponding to proficiency across a variety of computer tasks. Age (young vs old) and Gender (men vs women) were entered as between-participant factors. This ANOVA indicated that there was a main effect of age [F(1, 36) = 13.130, MSE = 69.809, p < .001, η_p ² = .267] ³ , no effect of gender [F(1, 36) = .594, MSE = 3.160, p = .446, η_p ² = .016], and no interaction between age and gender [F(1, 36) = 1.048, MSE = 1.048, p = .660, η_p ² = .005]. Younger adults (M = 27.58, SD = 0.85) were more computer proficient compared to older adults (M = 24.90, SD = 3.10). An identical analysis was performed on MDPQ scores, corresponding to technology proficiency across a variety of tasks using smartphones and tablets. There was a main effect of age [F(1, 36) = 18.264, MSE = 488.331, p < .001, η_p ² = .337], no effect of gender [F(1, 36) = 1.088, MSE = 27.121. p = .304, η_p ² = 0.029], and no interaction between age and gender [F(1, 36) = 0.848, MSE = 21.139, p = .363, η_p ² = .023]. Younger adults (M = 31.78, SD = 0.41) were more mobile device proficient compared to older adults (M = 24.90, SD = 7.06).

A Repeated Measures ANOVA was conducted on IPQ General Presence data, corresponding to the overall subjective sense of being present in the virtual environment (Schubert, 2003). Age (young vs old) and Gender (men vs women) were entered as between-participant factors and Experience (Meditation, SteamVR Home, Vesper Peak, Xortex) was entered as a within-participant factor. Here, and elsewhere, violations of sphericity were addressed using Greenhouse–Geisser degrees of freedom and MSE. This analysis indicated a main effect of experience [F(2.145, 77.217) = 4.981, MSE = 3.767, p < .05, η_p ² = .122] and a main effect of age [F(1, 36) = 5.824, MSE = 9.290, p < 0.05, η_p ² = .139]. No other effects or interactions were significant. As can be seen from Figure 1A, older adults reported higher levels of general presence compared to younger adults. No a priori hypotheses were made regarding differences between experiences. However, post-hoc t-tests with Bonferroni correction revealed that presence scores were higher for the Vesper Peak experience compared to Meditation (p < .001) and Xortex (p < .05).

Spatial Presence data, corresponding to the sense of being physically present in a virtual environment (Schubert, 2003), were entered into an identical ANOVA. The analysis revealed a main effect of experience [F(2.396, 86.273) = 10.289, MSE = 3.169, p < 0.001, η_p ² = .222] and a main effect of age [F(1, 36) = 4.883, MSE = 4.883, p < 0.05, η_p ² = .120]. Here again, older adults reported higher presence scores compared to younger adults (Figure 1B). Post-hoc t-tests with Bonferroni correction revealed that Meditation was rated as significantly lower in terms of spatial presence compared to all other experiences (all p values <0.05).

Next, we ran the exact same analysis on IPQ Involvement Data, with involvement being defined as the attention one gives to the virtual environment and how involved one is with the experience (Schubert, 2003). This analysis revealed a main effect of experience [F(4.425, 78.355) = 9.260, MSE = 4.425, p < 0.001, η_p ² = .205] and a main effect of age [F(1, 36) = 6.214, MSE = 11.926, p < 0.05, η_p ² = .147]. In general, older adults reported a greater level of involvement compared to younger adults (Figure 1C). Post-hoc t-tests with Bonferroni correction revealed that scores were higher for the Xortex experience compared to Meditation (p < .01) and SteamVR Home (p < .01). Vesper Peak was also rated higher than Steam Home VR (p < .05).

Finally, the same analysis was conducted on IPQ Realism data, with realism being defined as the subjective experience of realism of the virtual environment (Schubert, 2003). This analysis indicated a main effect of experience [F(2.168, 78.037) = 13.253, MSE = 7.063, p < 0.001, η_p ² = .269] and a main effect of age [F(1, 36) = 5.096, MSE = 7.183, p < 0.05, η_p ² = .124]. As can be seen from Figure 1D, older adults reported higher levels of realism. Vesper Peak was reported as more realistic than Meditation (p < .001) and SteamVR Home (p < .001). Xortex was among the least realistic, rated significantly lower compared to Vesper Peak (p < .001) and SteamVR Home (p < .05). Finally, Steam Home VR was reported as being more realistic than Meditation (p < .05).

In summary, in general, older adults reported higher levels of presence within VR environments across multiple measures of presence compared to younger adults (see Figure 2 for scatterplots of these relationships, largely consistent with ANOVA results). Not surprisingly, given the diversity of the experiences participants were asked to interact with, measures of presence varied across experiences. Intuitively, the most photo-realistic experience (Vesper Peak) was rated high in terms of presence, and the most dynamic and game-like experience (Xortex) was rated high in terms of involvement, but low in terms of realism.

An identical analysis was conducted on NASA TLX Mental Demand data (Figure 3A). This revealed a main effect of age [F(1, 36) = 6.98, MSE = 33.337, p < 0.05, η_p ² = .162], a main effect of experience [F(2.150, 77.400) = 15.463, MSE = 39.032, p < 0.001, η_p ² = .300], and an interaction between age and experience [F(2.150, 77.400) = 4.827, MSE = 12.185, p < 0.01, η_p ² = .118]. To probe the nature of this interaction post-hoc ANOVAs were conducted on each experience separately using age as the between-participant factor. This revealed a main effect of age in the SteamVR Home experience exclusively [F(1, 38) = 32.510, MSE = 55.225, p < .001, η_p ² = .461]. Older adults reported higher mental workload compared to younger adults for this experience. With respect to experience, post-hoc t-tests with Bonferroni correction revealed that Xortex was more demanding than Vesper Peak (p < .01) and Meditation (p < .001). Meditation was rated as less demanding compared to SteamVR Home (p < .01) and Vesper Peak (p < .05).

For the NASA TLX Physical Demand data (Figure 3B), there was a significant main effect of experience only (F(1.836, 66.086) = 16.102, MSE = 28.994, p < 0.001, η_p ² = .309). Post-hoc t-tests with Bonferroni correction revealed that Xortex was more physically demanding than all other experiences (all p values <0.01). Vesper Peak was also rated as more physically demanding than Meditation (p < .01).

The same analysis was conducted on NASA TLX Temporal Demand data (Figure 3C). This analysis revealed a significant main effect of experience only [F(1.387, 49.923) = 32.296, MSE = 87.911, p < 0.001, η_p ² = .480]. Not surprisingly, this appeared to be driven by greater temporal demands associated with the fast-paced action game Xortex. Xortex was rated to have a greater temporal demand compared to all other experiences (all p values <0 .001).

The exact same analysis was conducted on NASA TLX Frustration data (Figure 3D). The analysis indicated no significant effects (p values for all main effects and interactions >.21). Older and younger adults did not differ in their feelings of frustration towards the experiences.

Finally, we turn to NASA TLX Effort data (Figure 4). This analysis indicated that there was a significant main effect of experience [F(2.178, 78.409) = 31.794, MSE = 77.386, p < .001, η_p ² = .469] and a significant interaction between gender and age [F(1, 36) = 4.443, MSE = 38.753, p < 0.05, η_p ² = .110]. To probe the nature of this interaction we conducted post-hoc ANOVAs on each experience separately with gender and age as between-participant factors. This analysis revealed a significant interaction between gender and age in the Vesper Peak experience [F(1, 36) = 6.716, MSE = 24.197, p < 0.05, η_p ² = 157]. That is, among older adults, women reported less effort than men, whereas among younger adults, women reported higher levels of effort than did. Xortex was more effortful compared to all other experiences (all p values <0 .001) and SteamVR Home was more effortful compared to Meditation (p < .05), according to post-hoc t-tests with Bonferroni correction.

While in some instances, age differences were evident, for many experiences older adults and younger adults reported similar workload ratings. This was true in general for measures of Temporal Demand and Frustration, and many experiences with respect to Mental Demand and Physical Demand. This was despite clear evidence from the CPQ and MDPQ that older adults were less proficient overall with technology. When differences occurred, older adults tended to report higher levels of demand compared to younger adults, though sometimes age interacted with gender. Not surprisingly, the more complex environments were associated with higher demands compared to experiences like Meditation and Vesper Peak.

The System Usability Scale (SUS) is a well-validated measure of usability that yields a single usability score from 0 to 100. An ANOVA was conducted on SUS data with Age (young vs old) and Gender (male vs female) as between-participant factors. This analysis revealed no effect of age [F(1, 36) = 1.666, MSE = 216.282 p = 0.205, η_p ² = .044], no effect of gender [F(1, 36) = 1.836, MSE = 238.413. p = 0.184, η_p ² = .049], and no interaction [F(1, 36) = 0.072, MSE = 9.356, p = 0.790, η_p ² = 0.002]. Older adults reported positive user experiences (M = 73.38, SD = 9.78) similar to younger adults (M = 77.13, SD = 12.81).

An ANOVA was conducted on SSQ (measuring simulator/cybersickness) data with Age (young vs old) and Gender (male vs female) as between-participant factors. This analysis revealed a main effect of age [F(1, 36) = 5.554, MSE = 37.411, p < .05, η_p ² = 0.134]. Contrary to expectations, older adults experienced less cybersickness (M = 1.35, SD = 1.60) compared to younger adults (M = 3.50, SD = 3.22), though in general, rates of cybersickness for both groups were low. For the distribution of scores, see Figure 5. To address potential issues of skew, the same ANOVA was conducted on log transformed data with similar results [age effect F(1, 36) = 7.778, MSE = .735, p < .01, η_p ² = .178].

Participants were asked about their likes and dislikes at the end of the experiment (see OSF for full responses; https://osf.io/vkxde/?view_only=5e36543a3891498da1eeb9a3495698b2). All older adults reported aspects of the experiences that they enjoyed, including being able to do things they could not do in reality and the ability to see new places. Older participants generally reported their experiences were relaxing, enjoyable, and fun. For dislikes, 45% of older adults reported comments that conveyed that there was “nothing” that they disliked about their experiences. Two participants reported comments related to difficulty using the controllers, and one participant reported disliking the cord on the headset (though this was not unique to older adults). Other comments related to wanting to explore and move through the environment more or wanting more time within the virtual environments. In general, comments were strongly positive with no strongly negative comments. Three older adult participants mentioned the fast-paced or challenging nature of Xortex as a dislike. Younger adult comments were similar to those of older adults. Younger participants generally reported fun and enjoyment. For dislikes, one participant reported dislike of the controllers and one reported dissatisfaction with the headset cord. However, a greater number of younger participants (five) reported dislikes related to Meditation VR due to its general lack of interactivity (while only one older adult mentioned this).

Overall, VR appears to be a promising technology for use with older adults, but several limitations of the current study must be acknowledged. While gender differences were explored in reported analyses, it is important to acknowledge the unequal gender distribution of each group, which may have impacted our ability to assess and understand differences between genders and age groups. Likewise, the shortcomings of self-report measures, which may be biased, should be noted. Although many studies utilize self-report measures of simulator sickness (Appel et al., 2020; Narciso, 2019; Yildirim, 2020), there may be potential in predicting and gauging cybersickness by analyzing participant-level and application-level predictors (Rebenitsch and Owen, 2019; Kim et al., 2021). Even still, self-report measures of cybersickness, workload, and presence are common and well accepted in the VR literature. The timing of the questionnaires should also be noted, especially the presence questionnaires which were administered after each experience. However, in general, Schwind et al. (2019) reported no impact on mean scores when assessing measures during vs after the VR experience. They note that administering the questionnaire within the VR environment could, however, reduce study duration and disorientation.

Reports of workload and usability were generally comparable between younger and older adults. However, sample size (N = 40) was constrained due to available resources. A post-hoc power analysis, for a sample of 40, indicated that the study was only powered to detect large mean differences between groups (minimum d = .92, α = .05, β= .80). Null effects, including the absence of observed age effects on measures of workload and usability, should be interpreted with caution. Small or medium age effects could exist and be revealed in studies featuring larger sample sizes. It is also important to note that workload and usability were measured with respect to a system that was already setup by the experimenter, and with the assistance of an experimenter who provided guidance. In the absence of such guidance, it is likely that both younger and older adults would have had much less positive VR experiences. Older adults, as a result of their less technology experience (Boot et al., 2015), might have experienced greater challenges setting up and using the system and navigating to the appropriate VR program.

Additionally, results reported here reflect experiences after short-term exposure to VR. It is unclear how attitudes and experiences of younger and older adults might change after prolonged use of VR technology. It should also be noted that participants experienced all VR experiences in a fixed order, and perceptions of one experience may have been influenced by ones experienced prior. Larger studies will be able to explore the impact of potential order effects, and future studies should also parametrically manipulate exposure duration to obtain a better sense of how duration impacts all factors discussed here. However, despite these limitations, results provide initial insight into potential age-related differences in the experience of VR and barriers to the use and adoption of VR technology.