Original Research

Psychological barriers to technology acceptance in South African workplaces: Investigating socio-economic status as a moderator

Mariella Noriega Del Valle, Karolina Łaba, Claude-Hélène Mayer

Received: 18 June 2025; Accepted: 16 July 2026; Published: 31 Aug. 2026

Copyright: © 2026. The Authors. Licensee: AOSIS.
This work is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license (https://creativecommons.org/licenses/by/4.0/).

Abstract

Orientation: Digital transformation in South African organisations hinges on employee acceptance, yet psychological barriers may inhibit technology uptake. Socio-economic inequalities may further shape engagement with cloud-based technology.

Research purpose: To determine whether socio-economic status (SES) moderates the relationship between psychological barriers and technology acceptance.

Motivation for the study: Given South Africa’s socio-economic disparities, psychological and socio-economic factors were examined to better understand workplace technology resistance.

Research approach/design and method: A quantitative, cross-sectional survey was conducted with 436 South African employees who regularly used cloud-based technologies at work. Measures included the Psychological Barriers to Technology Acceptance Tool (PBTAT). Descriptive, correlational and moderation analyses were conducted.

Main findings: Psychological barriers reduced technology acceptance (β < 0, p < 0.05), while SES was non-significant. Lack of psychological safety and low frustration tolerance significantly predicted technology acceptance (both p < 0.01). Self-efficacy, organisational trust and psychological availability were non-significant moderators (p > 0.05).

Practical/managerial implications: Psychological barriers should be addressed across the workforce. As SES did not alter their effect, training initiatives should be inclusive rather than tailored by socio-economic background.

Contribution/value-add: The study extends the Technology Acceptance Model by integrating five psychological barrier constructs into a South African workplace framework. It introduces and preliminarily validates the 62-item PBTAT and shows that psychological barriers inhibit technology acceptance regardless of SES.

Keywords: technology acceptance; psychological barriers; socio-economic status; psychological safety; organisational trust; self-efficacy; frustration tolerance; South Africa.

Introduction

Technological innovation in the workplace has accelerated with the rise of Industry 4.0 systems, including cloud-based technology that enables organisations to store data remotely, facilitate real-time collaboration, and scale digital operations across geographically dispersed workforces (Bhatt, 2024). Successful adoption of such technologies depends not only on technical implementation but also on employee acceptance (Kaya & Dikmen, 2024; Marabelli & Lirio, 2024; Zahoor et al., 2022).

In South Africa, a uniquely stratified socio-economic context persists because of historical inequalities, creating a pronounced digital divide. These disparities continue to shape how employees engage with new technologies (Khoza et al., 2024; Roberts et al., 2021). Lower socio-economic status (SES) groups often face additional barriers to technology use, such as limited access to infrastructure, reduced digital literacy and fewer upskilling opportunities (Bvuma & Marnewick, 2020; Lethoko, 2024; Spaull & Van der Berg, 2020). Rogers’ (2003) Diffusion of Innovations theory posits that those most in need of new technologies are often the last to adopt them. Indeed, empirical studies show that individuals with lower income and education levels tend to adopt digital tools more slowly (Chancel et al., 2022; World Bank, 2023).

In the South African context, apartheid-era educational and economic legacies have deepened the divide. Only 13% of black-headed households have fixed internet access at home, compared to substantially higher rates among other racial groups (Freedom House, 2024). These persistent inequalities highlight the importance of contextualising workplace technology acceptance within SES realities.

Despite the evident need for context-sensitive models, many technology acceptance theories remain overly cognitive. The Technology Acceptance Model (TAM) developed by Davis (1989) focuses on perceived usefulness and ease of use as primary determinants of user adoption. Subsequent extensions (TAM2, TAM3) maintained this emphasis, helping predict behavioural intentions towards technology by examining system-related perceptions (Venkatesh & Bala, 2008; Venkatesh & Davis, 2000). While parsimonious and widely applied, TAM has been criticised for its limited consideration of emotional, cultural and contextual variables (Dalle et al., 2024). Specifically, TAM does not account for emotions such as anxiety, fear of failure or technostress, nor does it incorporate contextual factors such as organisational climate, cultural values or systemic inequality (Lee et al., 2025; Tarafdar et al., 2011). Alternative frameworks have attempted to address these gaps: The Unified Theory of Acceptance and Use of Technology (UTAUT; Venkatesh et al., 2003) introduced social influence and facilitating conditions, while Oreg (2006) centred resistance to change as a psychological disposition. Nevertheless, these extensions remain predominantly cognitive and do not fully account for the emotional and contextual dimensions that shape technology acceptance, particularly in diverse, non-Western settings (Sabiteka et al., 2025).

Recent scholarship has argued that such ‘system-centric’ approaches overlook crucial psychological and social dynamics (Kučević et al., 2024; Lee et al., 2025; Übellacker, 2025). This is particularly salient in countries like South Africa, where cultural diversity and systemic inequality influence workforce readiness (Khoza et al., 2024). Western-derived acceptance frameworks may not fully capture the lived realities of employees in developing contexts, necessitating more localised and nuanced models (Sabiteka et al., 2025).

South African studies echo this view. Roestenburg (2021) found that in social work settings, technology acceptance was shaped more by socio-cultural and emotional factors than by system functionality. Similarly, Khoza et al. (2024) emphasised the need to adapt acceptance models to reflect systemic disadvantage and varying psychological experiences. These insights align with principles in Industrial and Organisational Psychology, which foreground the role of emotion, perception and organisational climate in shaping employee behaviour (Coetzee & Veldsman, 2022).

This study therefore builds on a growing consensus that psychological barriers should be integrated into technology acceptance frameworks, particularly in unequal societies. Psychological barriers are defined here as emotional, cognitive or attitudinal impediments that reduce an individual’s willingness to engage with new technologies (Roberts et al., 2021). Examples include fear of failure, stress, resistance to change or perceived lack of support (Oreg, 2006; Rehman et al., 2021; Van den Heuvel et al., 2015). In the South African workplace, such barriers may be intensified by contextual stressors related to inequality, workload or past educational access.

The research is positioned at the intersection of three critical dimensions: Psychological readiness, socio-economic context and workplace technology adoption. In doing so, it seeks to expand dominant models by incorporating the often-overlooked psychological and contextual factors that influence real-world technology uptake.

Literature review

Psychological barriers to technology acceptance

Psychological barriers refer to internal emotional, cognitive or attitudinal factors that impede an individual’s willingness to engage with new technology (Roberts et al., 2021). These may include resistance rooted in fear, perceived incompetence, anxiety about disruption or low confidence (Oreg, 2006; Rehman et al., 2021; Van den Heuvel et al., 2015). In the workplace, such barriers are often activated during technological change, particularly when employees anticipate failure, judgement or increased workload (Golgeci et al., 2025; Pichlak, 2016; Valtonen & Holopainen, 2024). Five barrier constructs were identified in earlier phases of this research as particularly relevant in the South African context: (1) psychological safety, (2) self-efficacy, (3) frustration tolerance, (4) organisational trust and (5) psychological availability (Noriega Del Valle et al., 2024). These constructs also reflect themes found in broader literature on technology adoption and change readiness (Gonzalez & Kanitz, 2023).

Psychological safety

This construct describes the extent to which individuals feel safe to experiment, ask for help, or make mistakes without fear of negative consequences. A psychologically unsafe environment fosters risk aversion and reduces willingness to engage with new systems (Choi & Ji, 2015). Rehman et al. (2021) and Andersson et al. (2020) demonstrated that psychological safety in the workplace correlates with lower resistance and greater innovation. More recently, Jin and Peng (2024) confirmed that team psychological safety significantly enhances innovative performance through open communication, while Zadow et al. (2023) found that a psychosocial safety climate predicts creativity, innovation and work performance in digital environments. In the South African workplace, where hierarchical dynamics may inhibit voice, the absence of such safety may substantially hinder technology engagement.

Self-efficacy

Self-efficacy is defined as the belief in one’s ability to perform a task successfully (Bandura, 1986). Technological self-efficacy influences how employees approach unfamiliar tools, with higher self-efficacy predicting stronger acceptance and lower avoidance behaviours (An et al., 2022; Wang et al., 2024). Venkatesh and Davis (2000) included self-efficacy in TAM extensions, linking it to perceived ease of use. Pan (2020) demonstrated that individuals confident in their digital capabilities were more open to adopting e-learning platforms. In this study, employees with low self-efficacy often described heightened anxiety during new software rollouts. This aligns with Molino et al. (2020), who emphasised that self-efficacy is a key resource in technology engagement and can be enhanced through training and support. In the South African context specifically, Khoza et al. (2024) demonstrated that technology self-efficacy is a significant predictor of technology acceptance and work engagement, reinforcing its relevance as a barrier construct in this study.

Frustration tolerance

Frustration tolerance is a broader psychological trait that refers to an individual’s capacity to withstand discomfort, setbacks and obstacles without becoming overwhelmed or disengaging (Harrington, 2005). In the context of workplace technology adoption, low frustration tolerance manifests as impatience, stress or early withdrawal when engaging with unfamiliar or complex systems. While technostress dimensions such as techno-complexity and techno-overload are relevant triggers (Tarafdar et al., 2011), frustration may equally stem from uncertainty about one’s role, perceived lack of control over technological change, or insufficient organisational support during implementation (Oreg, 2006; Valtonen & Holopainen, 2024). Heidenreich et al. (2016) found that satisfied employees resisted innovation if they anticipated ongoing frustration. In this study, frustration tolerance was operationalised using the Frustration Tolerance subscale of the Psychological Barriers to Technology Acceptance Tool (PBTAT; Noriega Del Valle et al., 2024), a validated instrument developed specifically for the South African organisational context. Training that builds coping strategies and resilience can enhance tolerance, while transparent communication and organisational support may reduce frustration triggers (Lifshitz et al., 2022; Valtonen & Holopainen, 2024).

Organisational trust

Trust in an organisation’s motives for implementing technology significantly affects employee engagement. A lack of trust may arise if employees suspect ulterior motives (e.g. surveillance or job cuts). Gefen et al. (2003) extended TAM to include trust, showing its importance in conditions of uncertainty. In low-SES environments, trust may also depend on peer validation, as shown by Ebrahim and Van den Berg (2024), who found that in informal businesses, trust in digital tools grew through social proof. Transparent communication, participatory implementation and positive examples can thus strengthen trust and promote acceptance (Ezeudoka & Fan, 2024). Recent scholarship has further highlighted that digital trust, defined as employees’ confidence in their organisation’s intentions and competence in managing digital tools, is a prerequisite for sustained technology engagement and a key enabler of successful digital transformation (Capestro et al., 2024).

Psychological availability

Psychological availability refers to the individual’s capacity to engage emotionally and cognitively at work. Kahn (1990) defined it as the readiness to invest personal resources in a role. In the context of technology adoption, low availability may stem from emotional exhaustion, mental overload or competing priorities. Tripathi and Bajpai (2021) observed that digital workplace cultures often assume continuous availability, which may not be realistic in high-stress environments. Shi (2022) found that higher psychological availability among students was associated with greater openness to experimentation and digital innovation. Molino et al. (2020) also linked psychological capital (including resilience and energy) to stronger technology engagement. While understudied, this construct may help explain resistance in otherwise competent individuals who feel mentally depleted. Recent research on digital workplace demands has confirmed that cognitive and emotional overload, driven by hyperconnectivity, information overload, and constant availability expectations, significantly depletes employees’ psychological resources, reducing their capacity to engage with new technologies (Marsh et al., 2024).

By articulating these five dimensions, the study aims to enrich existing TAMs and offer a diagnostic framework that organisations can use to identify and address internal resistance to digital change. This study integrates established cognitive frameworks of technology acceptance with psychological and socio-economic dimensions to develop a more context-sensitive model suited to South African workplaces.

Traditional models, particularly the TAM by Davis (1989), conceptualise acceptance through perceived usefulness and ease of use. Extensions such as TAM2 and TAM3 (Venkatesh & Bala, 2008; Venkatesh & Davis, 2000) have retained this cognitive foundation, focusing primarily on how perceptions shape intention. While influential, these models largely neglect emotional and contextual variables. Critics argue that they underestimate the human experience of change, particularly in non-Western or stratified settings (Dalle et al., 2024; Kučević et al., 2024).

In contrast, the proposed framework emphasises psychological barriers which may operate independently of, or in tandem with, perceived usefulness and ease of use. These barriers are particularly salient in diverse, historically unequal environments such as South Africa, where employees may bring vastly different life experiences and resources to the workplace; as a result of this, SES is introduced as a potential moderator, with the assumption that resource disparities may amplify or mitigate the effects of psychological resistance. Literature on digital inequality (e.g. Rogers, 2003; Roberts et al., 2021) suggests that adoption is shaped not only by attitudes but by access, confidence and cultural exposure.

Figure 1 presents the proposed theoretical framework, which integrates the TAM (Davis, 1989) with five psychological barrier constructs as additional predictors of technology acceptance. The framework further proposes that the relationship between psychological barriers and technology acceptance is moderated by SES, as illustrated by H2 and its sub-hypotheses below.

FIGURE 1: Proposed theoretical framework of psychological barriers, socio-economic status and technology acceptance.

Based on the above literature and proposed theoretical framework, the following hypotheses were formulated:

H1: Psychological barriers will negatively predict technology acceptance among employees.

H2: SES will moderate the relationship between psychological barriers and technology acceptance.

Five sub-hypotheses are specified:

H2a: SES moderates the relationship between low self-efficacy and technology acceptance.

H2b: SES moderates the relationship between low frustration tolerance and technology acceptance.

H2c: SES moderates the relationship between lack of psychological safety and technology acceptance.

H2d: SES moderates the relationship between lack of organisational trust and technology acceptance.

H2e: SES moderates the relationship between lack of psychological availability and technology acceptance.

These hypotheses draw on prior research suggesting that individuals with fewer resources or less digital exposure may be more susceptible to psychological impediments in the context of organisational technology change.

Research design

This study adopted a quantitative, cross-sectional survey design situated within a pragmatist paradigm. As the third phase of a broader mixed-methods project, it aimed to test whether SES moderates the relationship between psychological barriers and technology acceptance. The approach was explanatory and correlational, relying on validated instruments to test variable relationships based on insights from the earlier qualitative phase.

Participants and sampling

The sample included n = 436 South African employees working in organisations that had recently implemented or were implementing cloud-based workplace technologies. A non-probability convenience sampling strategy was used. Participants were recruited via professional networks, corporate mailing lists and social media platforms (LinkedIn, Facebook).

Inclusion criteria required participants to: (1) use cloud-based technology in their work (e.g. customer relationship management systems, enterprise software) and (2) be proficient in English to ensure comprehension of survey items.

Demographically, 77.3% identified as female and 22.5% as male; 51.8% were aged between 26 and 41 years. The sample included predominantly white people (73.9%), with the remainder comprising black African people (13.7%), Indian people and/or Asian people (6.6%) and Coloured people (5.2%). Most respondents held tertiary qualifications (75.2%) and occupied skilled or managerial roles, indicating a largely middle- to high-SES profile.

Measures

Psychological Barriers to Technology Acceptance Tool

The 62-item PBTAT (Noriega Del Valle, 2025) was developed and validated specifically within the South African organisational context as part of a sequential mixed-methods doctoral study. The instrument was constructed through an iterative process that included a qualitative phase to identify culturally relevant psychological barrier constructs, followed by exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) to establish its psychometric properties. Exploratory factor analysis supported a five-factor solution explaining 66% of total variance, with a Kaiser-Meyer-Olkin measure of sampling adequacy of 0.96 and significant Bartlett’s test of sphericity (χ²(1891) = 21 291.09, p < 0.001). Confirmatory factor analysis confirmed the five-factor structure with acceptable goodness-of-fit indices Tucker-Lewis Index [TLI] = 0.86, Comparative Fit Index [CFI] = 0.87, Root Mean Square Error of Approximation [RMSEA] = 0.06, Standardised Root Mean Square Residual [SRMR] = 0.05. All five subscales demonstrated excellent internal consistency (Cronbach’s α ranging from 0.93 to 0.96). As the PBTAT is currently in press, the present study constitutes part of the ongoing validation of this instrument within its target context, contributing preliminary evidence of its applicability in technology-intensive South African workplaces. The five dimensions assessed included lack of psychological safety, assessed through items such as: ‘I worry I’ll be punished for mistakes I make on new systems’; lack of organisational trust, exemplified by statements such as ‘I do not trust my organisation’s reasons for introducing new technology’; lack of psychological availability, reflected in items such as ‘I’m usually too mentally drained to learn a new technology at work’; low self-efficacy, captured through statements like ‘I doubt my ability to use new workplace technologies effectively’; and low frustration tolerance, represented by items such as ‘If a new system does not work smoothly right away, I lose patience’. Items were rated on a five-point Likert scale (1 = strongly disagree to 5 = strongly agree), with high scores indicating stronger presence of the barrier.

Technology Acceptance Questionnaire

A 16-item scale based on Davis (1989) and Chin et al. (2008) was used to assess technology acceptance across four constructs. Perceived usefulness was measured through items such as ‘Using new workplace technology improves my job performance’. Perceived ease of use included items like ‘Learning to operate new systems is easy for me’. Attitude towards use was assessed with items such as ‘I have a positive attitude towards using new workplace technologies’. Behavioural intention to use was measured with items like ‘I intend to continue using new technology in my job’. All items were rated on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). The scale demonstrated high internal consistency (α = 0.94), and a composite technology acceptance score was computed for analysis.

Socio-Economic Status

Socio-Economic Status was determined using three indicators: education level, occupational category and income bracket, following standard SES classification frameworks (Van Aardt et al., 2016). Race was collected for descriptive purposes but excluded from SES computation. As a result of limited representation in the low-SES category (n = 1), SES was dichotomised into middle SES and high SES for moderation analysis.

Ethical considerations

Prior to data collection, ethical approval was obtained from the University of Johannesburg’s research ethics committee (Ethics Clearance Number: IPPM-2022-606 D). All participation was voluntary. Participants were fully informed of the study’s purpose, their right to withdraw at any time without consequence, and guarantees of confidentiality and anonymity. No personal identifying information was collected. A mandatory consent field on the survey landing page confirmed that participants had read the information sheet and agreed to participate before proceeding.

Procedure

The survey was administered online via Google Forms and distributed through professional networks, corporate mailing lists, and social media platforms (LinkedIn and Facebook). The Google Forms document contained the information sheet and consent form on its landing page, ensuring participants were informed of the nature, purpose and objectives of the study prior to completion. Socio-economic status-related items were placed at the end of the survey to reduce priming effects. The survey link remained open for 3 months in 2024. A total of 459 responses were received, of which 10 incomplete responses were removed, yielding a final sample of N = 436.

Data analysis

All analyses were conducted using IBM SPSS Statistics along with the PROCESS macro v4.3 (Hayes, 2022). Model 1 was employed to test whether SES moderated the relationship between psychological barriers and technology acceptance. Psychological barriers, both as a total composite and as five individual subscales (psychological safety, self-efficacy, frustration tolerance, organisational trust and psychological availability), served as independent variables. The dependent variable was the overall technology acceptance score, computed from the TAM questionnaire. Socio-economic status was treated as a categorical moderator (0 = middle SES; 1 = high SES) and dummy-coded accordingly.

Following best practices for moderation with categorical moderators (Hayes, 2018), predictor variables were not mean-centred. Each model was estimated in two stages: the first included the main effects of the psychological barrier and SES, while the second introduced the interaction term (Psychological Barrier × SES). These models allowed for examination of conditional effects across SES groups using the pick-a-point method at each SES level.

The significance level was set at α = 0.05, and all tests were two-tailed. Effect sizes (R²) were reported to indicate the proportion of variance explained by each model. Missing data were handled using listwise deletion. All assumptions were tested and met. Residual plots did not reveal heteroscedasticity, and variance inflation factors were well below the threshold of concern (VIFs < 2.5), indicating no multicollinearity.

Results

Psychometric properties

Prior to the moderation analysis, the psychometric properties of the measures were established to confirm that constructs were measured consistently and accurately. For the PBTAT, EFA supported a five-factor solution accounting for 66% of total variance (Kaiser–Meyer–Olkin [KMO] Measure of Sampling Adequacy = 0.96; Bartlett’s χ²(1891) = 21 291.09, p < 0.001). Confirmatory factor analysis validated the five-factor structure with acceptable goodness-of-fit indices (TLI = 0.86, CFI = 0.87, RMSEA = 0.06, SRMR = 0.05). All five PBTAT subscales demonstrated excellent internal consistency, with Cronbach’s alpha coefficients ranging from 0.93 to 0.96 (Psychological Safety α = 0.94; Organisational Trust α = 0.96; Frustration Tolerance α = 0.94; Self-Efficacy α = 0.95; Psychological Availability α = 0.93). For the TAM instrument, internal consistency was equally strong across all subscales (Perceived Usefulness α = 0.95; Perceived Ease of Use α = 0.92; Attitude Toward Use and Behavioural Intention α = 0.92), with an overall composite score α = 0.94. All coefficients exceeded the recommended threshold of 0.70 (Nunnally & Bernstein, 1994), confirming the reliability of both instruments prior to inferential analyses. Table 1 presents the descriptive statistics and Cronbach’s alpha coefficients for all PBTAT and TAM subscales.

Descriptive statistics demonstrated that employees reported a moderate degree of psychological resistance (barriers) to technology acceptance. On a five-point Likert scale (1 = strongly disagree, 5 = strongly agree), the overall mean barrier score was M = 2.8 (SD = 0.9). Among the five dimensions, low frustration tolerance had the highest mean (M = 3.0), followed closely by lack of psychological safety. Psychological availability exhibited the lowest barrier mean, consistent with the profile of a motivated workforce. Technology acceptance scores were generally high, with a mean of M = 3.98 (SD = 0.67). An inverse correlation was observed between overall psychological barriers and technology acceptance, r = −0.40, p < 0.001, indicating that higher psychological resistance was associated with lower levels of acceptance.

Socio-economic status comparisons revealed that high-SES participants reported slightly higher acceptance scores than their middle-SES counterparts (mean difference ≈ 0.15), although this difference did not reach statistical significance (p = 0.10). Similarly, high-SES participants reported marginally lower barrier scores (M = 2.7) compared to middle-SES employees (M = 2.9). The only significant SES-related difference emerged for self-efficacy: High-SES individuals reported significantly fewer self-efficacy concerns than middle-SES respondents (p < 0.05).

Table 2 presents the Pearson correlations between the psychological barrier subscales and technology acceptance measures. The correlation matrix confirms significant negative relationships between all five PBTAT subscales and the TAM overall composite score, with self-efficacy (r = −0.55) and psychological availability (r = −0.51) showing the strongest inverse associations. Among the TAM subscales, perceived ease of use demonstrated the strongest negative correlations with the barrier constructs. While some inter-subscale correlations were moderate to strong, variance inflation factors across all moderation models were well below the threshold of concern (VIFs < 2.5), confirming the absence of multicollinearity.

Moderation analysis (overall model)

The first moderation model examined the overall psychological barrier score as the independent variable and SES as a moderator. This regression model was statistically significant, F(3, 370) = 52.51, p < 0.001, explaining 33.5% of the variance in technology acceptance (R² = 0.335). A significant main effect was observed for psychological barriers (B = −0.66, p = 0.005, 95% CI [−1.13, −0.20]), suggesting that higher levels of perceived barriers significantly reduced technology acceptance.

Thus, H1: Psychological barriers negatively predict technology acceptance among employees was confirmed.

The main effect of SES was, however, not significant (B = −0.005, p = 0.981), nor the interaction term (B = 0.04, p = 0.681). These results indicate that SES did not significantly moderate the relationship between psychological barriers and technology acceptance.

Figure 2 displays the parallel regression slopes for each SES group. The negative relationship between barriers and acceptance was nearly identical across groups, reinforcing the absence of an interaction effect.

FIGURE 2: Relationship between psychological barriers and technology acceptance by socio-economic status group.

Dimension-specific analyses

To explore whether particular psychological barriers were differentially moderated by SES, separate moderation models were estimated for each barrier subscale.

For low self-efficacy, the overall model was significant, F(3, 419) = 27.90, p < 0.001, R² = 0.170. The main effect of low self-efficacy on acceptance was marginal (B = –0.35, p = 0.071). Socio-economic status and the interaction term were both non-significant.

For lack of organisational trust, the model was also significant, F(3, 411) ≈ 25.50, p < 0.001, R² ≈ 0.160, with a near-significant negative main effect (B = −0.31, p = 0.082). Neither SES nor the interaction term were significant.

For low frustration tolerance, a robust negative relationship with technology acceptance emerged (B = −0.51, p = 0.002), with the model accounting for 20.8% of the variance, F(3, 409) = 27.74, p < 0.001. The SES interaction was again non-significant (B = 0.09, p = 0.160).

For lack of psychological safety, this barrier also significantly predicted reduced acceptance (B = −0.41, p = 0.036), with the model demonstrating strong explanatory power, F(3, 413) = 45.40, p < 0.001, R² = 0.310. Socio-economic status and interaction terms were not significant.

For lack of psychological availability, the model showed moderate explanatory power, F(3, 412) = 31.98, p < 0.001, R² = 0.274, but the main effect of psychological availability was non-significant (B = −0.37, p = 0.127), as were SES and interaction terms.

Across all five models, the SES × Barrier interaction terms were statistically non-significant (p > 0.05). Hence, Hypotheses 2a, 2b, 2c, 2d, and 2e were rejected. The moderation analysis results for each psychological barrier dimension are summarised in Table 3.

Discussion

This study makes three distinct contributions to the technology acceptance literature. Firstly, it extends the TAM by demonstrating that psychological barriers, specifically frustration tolerance and psychological safety, are significant predictors of technology acceptance, operating alongside and independently of TAM’s traditional cognitive constructs of perceived usefulness and ease of use. This finding supports calls for more psychologically grounded and contextually sensitive acceptance frameworks, particularly in unequal societies such as South Africa (Khoza et al., 2024; Sabiteka et al., 2025). Secondly, this study provides empirical validation of the PBTAT as a reliable and valid instrument for measuring psychological barriers in South African organisational contexts, thereby contributing a locally developed diagnostic tool to the field. Thirdly, the findings challenge the assumption that socio-economic background is a primary driver of technology resistance among employed workers. Instead, once employees are embedded in technology-enabled roles, psychological readiness appears to emerge as the more salient predictor regardless of SES group.

The present study examined whether SES moderates the relationship between psychological barriers and technology acceptance in South African workplaces. While psychological barriers significantly predicted reduced technology acceptance, SES did not moderate these relationships. The influence of psychological constraints, such as frustration, fear and mistrust, appeared consistent across both middle- and high-SES employees. This finding supports the view that internal psychological factors, including cognitive overload and fear of failure, can impede technology adoption regardless of socio-economic background. In line with literature on technostress (Tarafdar et al., 2011) and resistance to change (Oreg, 2006), these constraints appear to exert a pervasive effect across occupational groups.

Among the five measured dimensions, low frustration tolerance and lack of psychological safety demonstrated the strongest negative associations with technology acceptance. Employees who become easily discouraged by technical complexity or errors may avoid engaging with digital systems, which aligns with previous studies linking techno-overload with disengagement (Heidenreich et al., 2016). Similarly, employees who fear being judged or penalised for mistakes may resist engaging with unfamiliar systems. This finding aligns with Edmondson’s (2018) conclusion that blame-free organisational cultures foster innovation and readiness for change. These findings further support the extended theoretical framework of this study, which integrates TAM with psychological barriers. While TAM traditionally emphasises perceived usefulness and ease of use (Davis, 1989), the strong influence of frustration tolerance and psychological safety suggests that internal emotional responses also play a critical role in shaping acceptance. This supports calls for expanded models that incorporate emotional and cognitive-affective barriers alongside traditional cognitive appraisals (Gonzalez & Kanitz, 2023; Venkatesh & Bala, 2008).

The remaining psychological variables demonstrated comparatively weaker relationships with technology acceptance, though these findings remain theoretically relevant. The weaker influence of self-efficacy may reflect the composition of the sample, which consisted predominantly of skilled professionals already embedded in technology-intensive roles and therefore likely to possess a baseline level of digital confidence. Similarly, the weaker relationship observed for organisational trust may reflect relatively high levels of institutional trust among professionally employed participants. Psychological availability also did not emerge as a strong predictor, possibly because participants reported sufficient cognitive and emotional resources to engage with workplace technology. Nevertheless, these variables may assume greater importance in contexts involving high technological demand, organisational instability or lower levels of digital exposure (Molino et al., 2020).

Although SES did not moderate the relationship between psychological barriers and technology acceptance, this finding warrants careful interpretation. The sample was relatively homogeneous, comprising predominantly middle- to high-SES employees in skilled, technology-intensive roles, which may have constrained the detection of moderation effects. In addition, participants shared a common level of digital exposure because of their work environments, potentially minimising SES-based differences in technology acceptance. It is also possible that SES functions as a distal rather than proximal variable, shaping access to digital exposure, training opportunities and support resources rather than directly influencing attitudinal responses once employees are embedded in technology-enabled roles. This interpretation is consistent with Rogers’ (2003) diffusion theory and research highlighting how educational inequality shapes digital readiness in South Africa (Spaull & Van der Berg, 2020). The findings further suggest that psychological barriers may represent relatively universal human experiences that transcend socio-economic background, consistent with self-determination theory, which posits that psychological needs such as competence and relatedness operate across demographic groups (Deci & Ryan, 2000).

Despite the absence of SES moderation effects, psychological barriers accounted for a substantial proportion of variance in technology acceptance, emphasising the practical relevance of integrating psychological constructs into acceptance models. These findings suggest that organisations introducing new technologies should prioritise psychologically informed interventions that strengthen employees’ frustration tolerance, confidence and sense of psychological safety. Training initiatives should therefore accommodate learning curves, normalise difficulties associated with new technologies, and foster supportive organisational climates that encourage experimentation without fear of judgement or failure.

In conclusion, this study demonstrates that psychological barriers play a central role in workplace technology acceptance, whereas SES appears to exert a less direct influence once employees are embedded in technology-enabled environments. The findings reinforce the importance of expanding technology acceptance frameworks beyond system-level characteristics to include internal psychological processes that shape employee readiness for technological change.

Implications

This study contributes to an expanded understanding of technology acceptance by integrating psychological and contextual dimensions. The significant impact of psychological barriers reinforces the argument that acceptance models should incorporate human-centric constructs. Contrary to expectations, SES did not moderate these relationships, supporting the proposition drawn from the Theory of Planned Behaviour (Ajzen, 1991) that psychological variables predict technology-related intentions across diverse populations, while background factors such as SES may serve more as antecedents than direct moderators. Future theoretical frameworks should therefore treat SES as shaping preconditions for acceptance, such as digital access and training quality, rather than as a moderator of attitudinal predictors.

From a practical standpoint, interventions aimed at enhancing technology adoption should be implemented broadly across the workforce, without assumptions that only specific demographic groups require support. Rather than segmenting by SES or other background variables, organisations would benefit from diagnosing specific psychological barriers within teams prior to digital rollouts using a validated instrument such as the PBTAT. Tailored strategies can then be applied, increasing transparency around system rationale for teams low in trust, or offering mentorship and hands-on guidance for groups low in self-efficacy. The non-significant SES moderation effect also highlights the importance of ensuring low-SES employees have equitable access to technology-intensive roles in the first place, as this study’s sample under-representation suggests a broader societal barrier that organisations can help address through inclusive onboarding and digital skills development.

Limitations and future research

Several limitations should be acknowledged. Firstly, the sample was skewed towards middle- to high-SES employees in skilled, technology-intensive roles, with minimal representation of lower-SES individuals. The reliance on an online survey and cloud-based technology as the study context may have structurally excluded employees in manual, entry-level or informal sectors. Future research should purposefully recruit from more socio-economically diverse occupational backgrounds.

Secondly, the cross-sectional design limits causal inference. The direction of influence between psychological barriers and technology acceptance cannot be firmly established, and reverse causality remains possible. Longitudinal designs would better capture how these relationships evolve during and after technology implementation.

Thirdly, contextual factors such as workload, organisational change history and leadership style were not measured. These variables may interact with or mediate the effect of psychological barriers and should be incorporated in future studies.

Fourthly, the SES index relied on education, income and occupational category, which may not fully capture the lived socio-economic experience in South Africa, where racialised inequality and geographic disparities persist. Multidimensional SES indicators would strengthen future investigations.

Finally, this study focused on technology acceptance intentions rather than actual usage behaviour or performance outcomes. Future research should extend the framework to examine whether psychological barriers also inhibit sustained use and productivity following implementation

Conclusion

This study advances understanding of technology acceptance by demonstrating that psychological barriers, particularly low frustration tolerance and lack of psychological safety, significantly inhibit technology acceptance among South African employees, regardless of socio-economic background. The non-significant SES moderation effect suggests that once employees are embedded in technology-enabled roles, internal psychological readiness is a more salient predictor of acceptance than demographic background. These findings support the integration of psychological constructs into TAMs and highlight the universal nature of certain psychological needs within organisational change contexts.

For researchers, this study signals a reorientation from broad demographic assumptions towards more granular psychological profiling when studying technology adoption. For practitioners, it emphasises the need to embed psychological support structures, including training that builds resilience, leadership that models psychological safety, and diagnostic tools such as the PBTAT, into change management strategies. In contexts of rapid digital transformation and widening inequality, fostering psychological readiness may prove a universally applicable lever for successful technology implementation.

Acknowledgements

This article is based on research originally conducted as part of Mariella Noriega Del Valle’s doctoral thesis titled ‘Measuring Technological Acceptance In Organisations: A Psychological Barriers Perspective’, submitted to the Department of Industrial Psychology and People Management, College of Business and Economics, University of Johannesburg in 2025. The thesis is currently unpublished and not publicly available. The thesis was supervised by Karolina Łaba and Claude-Hélène Mayer. The thesis was reworked, revised and adapted into a journal article for publication. The author confirms that the content has not been previously published or disseminated and complies with ethical standards for original publication.

This article is based on data from a larger study. Another article was published from the same thesis. The article focusing on ‘Unlocking technology acceptance among South African employees: A psychological perspective’ has been published in the SA Journal of Industrial Psychology, Volume 50.

Competing interests

The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.

CRediT authorship contribution

Mariella Noriega Del Valle: Conceptualisation, Investigation, Writing – original draft. Karolina Łaba: Supervision, Writing – review & editing. Claude-Hélène Mayer: Methodology, Supervision, Writing – review & editing. All authors reviewed the article, contributed to the discussion of results, approved the final version for submission and publication, and take responsibility for the integrity of its findings.

Funding information

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Data availability

The data that support the findings of this study are available from the corresponding author, Mariella Noriega Del Valle, upon reasonable request.

Disclaimer

The views and opinions expressed in this article are those of the authors and are the product of professional research. They do not necessarily reflect the official policy or position of any affiliated institution, funder, agency, or that of the publisher. The authors are responsible for this article’s results, findings, and content.

References