Gender,Immersion,and Change: Evaluating the Impact of VR Training on Self-Efficacy and Gender Stereotypes

Abstract

The construction industry faces workforce challenges including an aging workforce, declining interest among younger generations, and gender disparities, with women comprising less than 11% of workers. While automation and teleoperation emerge as strategies to address workforce shortages, psychological barriers—low self-efficacy (SE) and gender stereotypes (GS)—continue to limit prospective groups, particularly women, from choosing construction as a career path. This study explores the impact of training at various immersion levels on boosting SE and mitigating GS across gender groups. Using a between-subjects experimental design, participants were randomly assigned to two immersion levels of VR technology with balanced gender distribution and completed two sessions of tower crane operation simulations. SE and GS were assessed before and after the intervention. Results demonstrate that training with both VR immersion levels significantly increased SE across all participants, with women showing greater improvements than men—effectively eliminating pre-existing gender gaps. For GS, while main changes were not statistically significant, analyses revealed marginally significant interaction effects. Post-hoc comparisons showed that women and those in the immersive VR condition demonstrated significant improvements post-training. This research provides valuable insights for developing inclusive training approaches in increasingly automated construction environments, potentially helping address industry-wide labor shortages.

Keywords

virtual reality training teleoperation training gender differences self-efficacy gender stereotypes tower crane operation

Introduction

The construction industry faces workforce challenges including an aging demographic, immigration shortages, and declining interest among younger generations (Deloitte, 2024; Home Builders Institute, 2024; Welfare et al., 2021). As the industry increasingly adopts automation and teleoperation to reduce manual labor dependency (Kaluthantirige et al., 2023), tower crane operation exemplifies this transition with its evolution toward remote control systems (Muddassir et al., 2025). While these technological advancements reduce physical barriers to entry, psychological obstacles persist—low self-efficacy (SE) in operating heavy machinery (Song et al., 2021) and gender stereotypes (GS) associating construction with masculinity (Akinlolu et al., 2023). This highlights the need to design training that both develops technical competencies for automation and teleoperation and builds psychological readiness for underrepresented groups entering construction fields. Virtual reality (VR) training offers safe, realistic practice environments (Mikropoulos & Natsis, 2011) while addressing psychological barriers. Immersion level—from desktop VR allowing partial real-world awareness to immersive VR (IVR) fully engaging users’ perceptual attention—may critically influence outcomes. This study examines how VR training at different immersion levels affects SE and GS across gender groups in tower crane operation, with implications for construction industry training.

Background

SE refers to one’s belief in their ability to succeed in specific tasks (Bandura, 1977). Among four sources shaping SE, Bandura identified mastery experiences—where individuals practice progressively challenging tasks and process successful completion experiences—as most powerful. VR contexts offer particularly effective opportunities for mastery experiences due to controlled task designs that facilitate successful performance accomplishments. While training can improve SE (E. R. Francis et al., 2020; Schoenherr, 2024) with potentially greater gains for those with lower initial confidence (Baker et al., 2007; Mackay & Parkinson, 2010), findings on VR immersion’s impact on SE remain mixed (Cyrus Rezvanifar & Amini, 2020; Jimenez et al., 2021). Some studies suggest higher immersion enhances SE development (Makransky et al., 2020), while others find no significant differences between immersion levels (Buttussi & Chittaro, 2018; Shu et al., 2019). Emerging research suggests gender-specific responses: men show greater SE gains in fully immersive environments, while women demonstrate stronger development with non-immersive simulations (Pande & Jepsen, 2025). This highlights the need for gender-stratified analyses when examining immersion effects across genders.

GS refers to generalized assumptions about abilities based on gender (Heilman, 2012), which limit women’s exploration of roles in male-dominated industries such as construction (B. Francis et al., 2017; Powell et al., 2010). Though persistent, these stereotypes can be challenged through interventions like perspective-taking, which VR uniquely enables through avatar embodiment or gender-swapping experiences (Cha et al., 2025; Chang et al., 2019; Peck et al., 2018). Higher immersion intensifies stereotype reduction effects (Pande & Jepsen, 2025). However, explicitly gender-focused interventions may trigger stereotype threat (ST)—a phenomenon in which awareness of negative stereotypes induces anxiety and undermines performance (Ausburn et al., 2009; Spencer et al., 1999). Researchers have identified ST neutralization through identity-safe environments (Spencer et al., 2016). Gender-neutral VR training applies this approach by removing masculine cues typically found on construction sites, allowing task focus without triggering gender-based anxieties. This study examines whether such training reduces GS and if immersion level affects outcomes.

Guided by these empirical and theoretical gaps, we pose the following research questions (RQs): RQ1: How does participant SE in tower crane operation change after VR-based training? Does this vary by gender or immersion level? RQ2: How does participant GS toward the construction industry change after VR-based training? Does this vary by gender or immersion level?

Approach

The objective of this study is to evaluate the effectiveness of VR-based training in improving SE of tower crane operation and reducing GS towards the construction industry. Time (Pre or Post) served as a within-subject variable, while gender (Man or Woman) and immersion level (desktop VR or IVR) were between-subject variables.

Procedure

Following Institutional Review Board (IRB) approval, twenty-four college students (12 women, 12 men) were randomly assigned to desktop VR or IVR groups with gender balance (see Figure 1). Participants completed baseline surveys and pre-test measures of SE and GS. After viewing instructional videos and receiving platform-specific orientation, participants completed five guided tower crane load movement tasks in various scenarios (e.g., high winds, limited visibility, blind lifts, etc.) with real-time audio guidance to develop operational proficiency. Following a 5-min break, they completed another five unassisted tasks simulating independent operation. This two-phase approach (guided practice followed by independent application) built SE through progressive achievement. Post-test measures were collected after the full experience.

Figure 1.

VR tower crane simulator setup. Left: IVR condition; Right: desktop VR condition.

Measures

Baseline Information

Besides demographic information, we collected prior experience with video games (Exp_VG), VR (Exp_VR), and heavy construction equipment (Exp_HCE) as potential covariates. Each experience variable was measured using a single-item 5-point scale (1 = No experience at all, 5 = Extensive experience).

Tower Crane Operation Self-Efficacy

SE was measured using a 13-item scale (α = .978), adapted from Bandura’s (2006) guidelines for measuring SE. Sample items included “Identify the correct joystick movements for various crane operations,” and “Execute a sequence of precise maneuvers to accurately lower the load to its designated position.” Items were rated on a 10-point Likert scale (1 = strongly disagree, 10 = strongly agree).

Gender Stereotype Towards Construction Industry

GS was assessed using a 4-item scale (α = .827), adapted from frameworks by Devine (1989) and Jost & Banaji (1994). Items like “When you think of people who are very good at construction who do you think of?” and “Who do you think other people think of when they think of people who are very good at construction?” were rated on a 7-point Likert scale (1 = almost all men, 4 = gender balanced, 7 = almost all women).

Analytical Strategies

We analyzed baseline differences using chi-square tests for categorical variables, t-tests for continuous variables, and Mann-Whitney U tests for ordinal variables. Covariates were identified via Spearman correlations and group comparisons using chi-square or Mann–Whitney U tests. SE data were analyzed with fixed-effects models, while GS data, with an intraclass correlation coefficient of .946, warranted linear mixed-effects models (LMMs). Model selection was guided by likelihood ratio tests for nested models (Pinheiro & Bates, 2000) and by comparisons of Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and R²-related metrics for non-nested models. Final effects were interpreted via estimated marginal means with Bonferroni adjustments.

Outcome

Baseline comparisons revealed that IVR participants (M = 22.83) were older than desktop VR participants (M = 19.75, p = .003). No significant differences were found between groups in ethnicity (p = .346), Exp_VG (p = .588), Exp_VR (p = .137), or Exp_HCE (p = .568). Exp_HCE was correlated with both SE (r = .363, p = .079) and GS (r = .466, p < .05) and thus was included as a covariate in subsequent analyses.

Results

For RQ1, five fixed-effects models were compared (Table 1); Model 2, which included a significant Gender × Time interaction, was selected for SE modeling. Significant effects were found for Gender (p = .002), Time (p < .001), and Gender × Time (p = .048). Post hoc analysis showed significant SE gains for both women (M_pre = 2.02, M_post = 7.88, p < .001) and men (M_pre = 4.39, M_post = 8.43, p < .001). The initial gender disparity (Est. = 2.37, SE = 0.73, p = .002) in SE was no longer statistically significant at post-test (Est. = 0.55, SE = 0.73, p = .451). Immersion and Exp_HCE effects were non-significant.

Table 1.

Model Comparison Results for SE Analysis.

Predictors	Model 1			Model 2			Model 3			Model 4			Model 5
Predictors	B	CI	p	B	CI	p	B	CI	p	B	CI	p	B	CI	p
Gender	−1.46	[−2.65, −0.26]	.018	−2.36	[−3.83, −0.90]	.002	−1.46	[−2.67, −0.25]	.019	−2.36	[−3.84, −0.89]	.002	−2.26	[−4.31, −0.21]	.031
Time	4.95	[4.02, 5.88]	<.001	4.04	[2.77, 5.31]	<.001	4.77	[3.43, 6.10]	<.001	3.86	[2.28, 5.43]	<.001	3.86	[2.00, 5.73]	<.001
Immersion Level	−0.01	[−0.96, 0.94]	.986	−0.01	[−0.93, 0.91]	.986	−0.19	[−1.53, 1.15]	.776	−0.19	[−1.49, 1.11]	.768	−0.1	[−1.96, 1.77]	.918
Exp_VG	0.26	[−0.30, 0.83]	.347	0.26	[−0.28, 0.81]	.331	0.26	[−0.30, 0.83]	.352	0.26	[−0.28, 0.81]	.335	0.27	[−0.30, 0.83]	.34
Gender × Time				1.81	[0.02, 3.61]	.048				1.8	[−0.00, 3.63]	.05	1.81	[−0.83, 4.45]	.173
Immersion × Time							0.37	[−1.52, 2.25]	.697	0.37	[−1.45, 2.18]	.687	−0.19	[−2.84, 2.45]	.884
Gender × Immersion													0.36	[−2.28, 3.00]	.785
Gender × Time × Immersion													0.01	[−3.72, 3.74]	.994
Model evaluation
R ²	.75			.78			.75			.78			.78
Adjusted R²	.73			.75			.72			.74			.73
AIC	187.96			185.45			189.79			187.26			191.21
BIC	199.12			198.54			202.89			202.23			209.92

Note. Reference category for Gender (1 = Man); reference category for Time (1 = Pre); reference category for Immersion Level (1 = desktop VR).

For RQ2, five LMMs were compared (Table 2); Model 4, including Gender × Time and Immersion × Time interactions, showed the best fit for GS modeling. Both interactions were marginally significant (p = .092). Post hoc analysis showed GS scores increased significantly for women (M_pre = 2.57, M_post = 2.74, p = .021), with no change for men (M = 3.26 at both time points, p = 1.00). GS scores also increased for IVR participants (M_pre = 2.81, M_post = 2.98, p = .021), but remained unchanged for desktop users (M = 3.02 at both time points, p = 1.00).

Table 2.

Model Comparison Results for GS Analysis.

Predictors	Model 1			Model 2			Model 3			Model 4			Model 5
Predictors	B	CI	p	B	CI	p	B	CI	p	B	CI	p	B	CI	p
Gender	−0.61	[–1.35, 0.14]	.143	−0.6915	[–1.44, 0.06]	.101	−0.61	[–1.35, 0.14]	.143	−0.69	[–1.44, 0.05]	.1	−0.42	[–1.40, 0.55]	.439
Time	0.08	[–0.02, 0.19]	.119	~0	[–0.14, 0.14]	1	~0	[–0.14, 0.14]	1	−0.08	[–0.24, 0.07]	.32	~0	[–0.17, 0.17]	1
Immersion Level	−0.13	[–0.72, 0.46]	.688	−0.1292	[–0.72, 0.46]	.688	−0.21	[–0.81, 0.39]	.515	−0.21	[–0.81, 0.38]	.515	0.04	[–0.79, 0.86]	.934
Exp_VG	0.01	[–0.34, 0.36]	.947	0.1263	[–0.34, 0.36]	.947	0.01	[–0.34, 0.36]	.947	0.01	[–0.34, 0.36]	.947	0.02	[–0.33, 0.37]	.906
Gender × Time				0.1667	[–0.03, 0.36]	.106				0.17	[–0.01, 0.35]	.092	~0	[–0.24, 0.24]	1
Immersion × Time							0.17	[–0.03, 0.36]	.106	0.17	[–0.01, 0.35]	.092	~0	[–0.24, 0.24]	1
Gender × Immersion													−0.51	[–1.67, 0.66]	.439
Gender × Time × Immersion													0.33	[0.00, 0.67]	.077
Model evaluation
R² (Marginal)	.15			.15			.15			.16			.16
R² (Conditional)	.95			.96			.96			.96			.97
AIC	74.46			76.48			76.48			78.34			79.40
BIC	87.56			91.45			91.45			95.18			99.98

Note. Reference category for Gender (1 = Man); reference category for Time (1 = Pre); reference category for Immersion Level (1 = desktop VR).

Findings

Self-Efficacy Impact

Our research shows significant SE increases following VR-based training, with women experiencing greater gains, closing the initial gender gap. These results align with prior findings that training can especially benefit groups with lower initial confidence (Baker et al., 2007; Mackay & Parkinson, 2010). The lack of difference between IVR and desktop VR conditions extends Bandura’s (1977) theory, suggesting that structured mastery experiences—rather than immersion—primarily drive SE development. Both VR formats provided progressive achievement opportunities, demonstrating that Bandura’s framework remains applicable across different immersive technologies.

Gender Stereotype Impact

Our examination of GS outcomes revealed more complex patterns. Women showed significant mitigation in GS following training, while men’s scores remained stable, aligning with research on stereotype threat neutralization (Spencer et al., 2016). The immersion level played a more decisive role in GS outcomes than in SE development. Participants in the IVR condition showed significant mitigation in GS, while those in desktop VR showed no detectable change. These findings suggest that training programs aimed at addressing deep-rooted social cognitive biases may require different technological approaches than those focused primarily on enhancing self-perceptions and confidence judgments. When workforce diversity is a priority, training designers might consider investing in more immersive technologies to effectively challenge entrenched stereotypes.

Limitations

Despite these promising findings, our study has limitations. Our university student sample may not represent typical crane operators, and our small sample size (n = 24) requires cautious interpretation of these preliminary results. Future work should include industry participants, examine long-term effects, and explore how attitudinal changes translate into career behaviors.

Conclusion

This study investigated how VR-based training influences SE and GS in tower crane operation, with a focus on gender and immersion level differences. The results suggest that while SE can develop through different immersion levels, the transformative potential for addressing deeper attitudinal barriers may require the enhanced immersion that only IVR provides. For training designers seeking not only to build technical confidence but to cultivate truly inclusive workforces, IVR offers distinctive advantages that warrant investment.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by National Science Foundation (NSF) under Award Number CMMI-2222730. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NSF.

ORCID iDs

Yi Wu

Mohammad H. Kazemi

References

Akinlolu

Omotayo

Ebohon

O. J.

Ekundayo

(2023). An analysis of the mediational effect of socio-cultural factors on student’s career choice outcomes in the built environment: Applying the socio-cognitive career theory. Education Sciences, 13(10), 1035. https://doi.org/10.3390/educsci13101035

Ausburn

L. J.

Martens

Washington

A. L.

Steele

Washburn

(2009). A cross-case analysis of gender issues in desktop virtual reality learning environments. Journal of Industrial Teacher Education, 46(3), Article 6. https://ir.library.illinoisstate.edu/jste/vol46/iss3/6

Baker

Krause

Yaşar

Ş.

Roberts

Robinson-Kurpius

(2007). An intervention to address gender issues in a course on design, engineering, and technology for science educators. Journal of Engineering Education, 96(3), 213–226. https://doi.org/10.1002/j.2168-9830.2007.tb00931.x

Bandura

(1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. https://doi.org/10.1037/0033-295X.84.2.191

Bandura

(2006). Guide for constructing self-efficacy scales. In Pajares

Urdan

(Eds.), Self-efficacy beliefs of adolescents (Vol. 5, pp. 307–337). Information Age Publishing.

Buttussi

Chittaro

(2018). Effects of different types of virtual reality display on presence and learning in a safety training scenario. IEEE Transactions on Visualization and Computer Graphics, 24(2), 1063–1076. https://doi.org/10.1109/TVCG.2017.2653117

Cha

Peck

T. C.

(2025). The impact of avatar clothing on the Proteus effect and stereotype threat. In Iwai

Nedel

Peck

T. C.

Popescu

(Eds.), 2025 IEEE Conference on Virtual Reality and 3D User Interfaces (VR) (pp. 296–306). IEEE. https://doi.org/10.1109/VR59515.2025.00054

Chang

Luo

Walton

Aguilar

Bailenson

(2019). Stereotype threat in virtual learning environments: Effects of avatar gender and sexist behavior on women’s math learning outcomes. Cyberpsychology, Behavior, and Social Networking, 22(10), 634–640. https://doi.org/10.1089/cyber.2019.0106

Cyrus Rezvanifar

Amini

. (2020). Self-efficacy versus gender: Project-based active learning techniques in biomedical engineering introductory computer programming courses. Journal of Biomechanical Engineering, 142(11), Article 111004. https://doi.org/10.1115/1.4047924

10.

Deloitte. (2024, January 4). 2024 engineering and construction industry outlook. Deloitte United States. https://www2.deloitte.com/us/en/insights/industry/engineering-and-construction/engineering-and-construction-industry-outlook-2024.html

11.

Devine

P. G.

(1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56(1), 5–18. https://doi.org/10.1037/0022-3514.56.1.5

12.

Francis

Archer

Moote

DeWitt

MacLeod

Yeomans

(2017). The construction of physics as a quintessentially masculine subject: Young people’s perceptions of gender issues in access to physics. Sex Roles, 76(3), 156–174. https://doi.org/10.1007/s11199-016-0669-z

13.

Francis

E. R.

Bernard

Nowak

M. L.

Daniel

Bernard

J. A.

(2020). Operating room virtual reality immersion improves self-efficacy amongst preclinical physician assistant students. Journal of Surgical Education, 77(4), 947–952. https://doi.org/10.1016/j.jsurg.2020.02.013

14.

Heilman

M. E.

(2012). Gender stereotypes and workplace bias. Research in Organizational Behavior, 32, 113–135. https://doi.org/10.1016/j.riob.2012.11.003

15.

Home Builders Institute. (2024). Construction Labor Market Report (Fall 2024). https://www.nahb.org/advocacy/workforce-development-and-careers-in-construction

16.

Jimenez

P. P.

Pascual

Espinoza

Martin

S. S.

Guidi

(2021). Pedagogical innovations with a gender approach to increase computer programming self-efficacy in engineering students. In Klinger

Kollmitzer

Pester

(Eds.), Proceedings of the 2021 IEEE Global Engineering Education Conference (EDUCON) (pp. 322–328). IEEE. https://doi.org/10.1109/EDUCON46332.2021.9453933

17.

Jost

J. T.

Banaji

M. R.

(1994). The role of stereotyping in system-justification and the production of false consciousness. British Journal of Social Psychology, 33(1), 1–27. https://doi.org/10.1111/j.2044-8309.1994.tb01008.x

18.

Kaluthantirige

Silva

Hewage

Kaur Gil

S. P.

Gill

(2023). Construction labor shortage, challenges, and solutions: A survey-based approach. Proceedings of International Structural Engineering and Construction, 10(1), Article CON-16. https://doi.org/10.14455/ISEC.2023.10(1).CON-16

19.

Mackay

Parkinson

(2010). Gender, self-efficacy and achievement among South African technology teacher trainees. Gender and Education, 22(4), 445–462. https://doi.org/10.1080/09540250802467935

20.

Makransky

Petersen

G. B.

Klingenberg

(2020). Can an immersive virtual reality simulation increase students’ interest and career aspirations in science? British Journal of Educational Technology, 51(6), 2079–2097. https://doi.org/10.1111/bjet.12954

21.

Mikropoulos

T. A.

Natsis

(2011). Educational virtual environments: A ten-year review of empirical research (1999–2009). Computers & Education, 56(3), 769–780. https://doi.org/10.1016/j.compedu.2010.10.020

22.

Muddassir

Zayed

Ali

A. H.

Elrifaee

Abdulai

S. F.

Yang

Eldemiry

(2025). Automation in tower cranes over the past two decades (2003–2024). Automation in Construction, 170, Article 105889. https://doi.org/10.1016/j.autcon.2024.105889

23.

Pande

Jepsen

P. M.

(2025). Science lab safety goes immersive: An ecological media-comparison study with gender analyses assessing iVR’s learning effectiveness. Research and Practice in Technology Enhanced Learning, 20, Article 001. https://doi.org/10.58459/rptel.2025.20001

24.

Peck

T. C.

Doan

Bourne

K. A.

Good

J. J.

(2018). The effect of gender body-swap illusions on working memory and stereotype threat. IEEE Transactions on Visualization and Computer Graphics, 24(4), 1604–1612. https://doi.org/10.1109/TVCG.2018.2793598

25.

Pinheiro

J. C.

Bates

D. M.

(2000). Linear mixed-effects models: Basic concepts and examples. In Mixed-effects models in S and S-Plus (pp. 3–56). Springer. https://doi.org/10.1007/b98882

26.

Powell

D. R.

Son

S.-H.

File

San Juan

R. R.

(2010). Parent-school relationships and children’s academic and social outcomes in public school pre-kindergarten. Journal of School Psychology, 48(4), 269–292. https://doi.org/10.1016/j.jsp.2010.03.002

27.

Schoenherr

(2024). Personalizing real-world problems: Posing own problems increases self-efficacy expectations, intrinsic value, attainment value, and utility value. British Journal of Educational Psychology, 94(2), 407–424. https://doi.org/10.1111/bjep.12653

28.

Shu

Huang

Y.-Z.

Chang

S.-H.

Chen

M.-Y.

(2019). Do virtual reality head-mounted displays make a difference? A comparison of presence and self-efficacy between head-mounted displays and desktop computer-facilitated virtual environments. Virtual Reality, 23(4), 437–446. https://doi.org/10.1007/s10055-018-0376-x

29.

Song

Kim

Ahn

Kang

(2021). Effectiveness of VR crane training with head-mounted display: Double mediation of presence and perceived usefulness. Automation in Construction, 122, Article 103506. https://doi.org/10.1016/j.autcon.2020.103506

30.

Spencer

S. J.

Logel

Davies

P. G.

(2016). Stereotype threat. Annual Review of Psychology, 67, 415–437. https://doi.org/10.1146/annurev-psych-073115-103235

31.

Spencer

S. J.

Steele

C. M.

Quinn

D. M.

(1999). Stereotype threat and women’s math performance. Journal of Experimental Social Psychology, 35(1), 4–28. https://doi.org/10.1006/jesp.1998.1373

32.

Welfare

Sherratt

Hallowell

(2021). Perceptions of construction work: Views to consider to improve employee recruitment and retention. Journal of Construction Engineering and Management, 147(7), Article 04021053. https://doi.org/10.1061/(ASCE)CO.1943-7862.0002057