Population aging is accelerating worldwide, with the proportion of people aged 60 years and older projected to nearly double from 12% in 2015 to 22% by 2050 [1]. Meanwhile, with the rapid development of information and communication technologies (ICTs), digital technologies have become increasingly embedded in older adults’ everyday lives, mediating routine activities such as banking, shopping, health management, and social interaction and making digital competence increasingly comparable to foundational literacy [2,3]. The first-level digital divide, which refers to inequalities in access to digital technologies [4], has narrowed substantially with the widespread diffusion of internet-enabled devices and improved connectivity [5-7]. In China, for example, the number of internet users aged 60 years and older increased from 119 million in December 2021 to approximately 173 million in December 2025, an increase of approximately 45.4% over 4 years [8,9]. Nevertheless, older adults remain disproportionately disadvantaged in digital access, with an internet penetration rate of 53.7%, compared with 80.1% for the population as a whole [8]. In addition, despite the narrowing first-level digital divide, substantial inequalities persist at the second and third levels, reflected in disparities in how older adults use digital technologies and in the digital skills and competencies needed for effective and meaningful engagement [10,11]. Indeed, older adults’ online engagement remains markedly skewed toward instrumental uses (IUs; eg, online services and communication) rather than hedonic uses (HUs; eg, entertainment, gaming, and social media) [11-13]. For instance, in the United Kingdom, 58% of older internet users (ages 65‐74 years) shop online and 65% bank online, whereas only about 32% play online games [14].

Barriers to ICT adoption among older adults are multifaceted, encompassing health-related limitations, inadequate digital skills, resource constraints such as cost and connectivity, insufficient support networks, and psychological factors [15]. While many of these barriers primarily limit older adults’ opportunities or capacity to engage with technology, psychological factors may exert a more immediate influence on adoption decisions by shaping how older adults perceive their technological capabilities, evaluate the benefits and demands of technology, and respond to difficulties encountered during use. Accordingly, psychological barriers—particularly technophobia—are salient for older adults. Technophobia is commonly defined as a fear or anxiety about technology, and it is associated with delayed or hesitant adoption, especially among individuals with limited digital literacy or inaccurate beliefs about technology [11,16]. Although scholars have called for conceptualizing technophobia as multidimensional, much of the literature continues to treat it as a unidimensional construct, limiting fine-grained insight into how distinct fears map onto different forms of technology use [16-19]. Moreover, while prior research has examined demographic correlates of technophobia and shown that structured interventions (eg, training and intergenerational mentoring) can alleviate it [20,21], evidence remains scarce on whether and how spontaneous day-to-day technical support—more typical of naturalistic settings—shapes technophobia within a broader sociotechnical context.

To address these gaps, this study examines how daily technical support (DTS) influences older adults’ instrumental versus hedonic technology use through multiple fear-related dimensions of technophobia. We further test whether these indirect effects vary by gender, given documented gender differences in support-seeking motivations and technology-related stereotypes [22]. By clarifying which dimensions of technophobia are most responsive to everyday support—and for whom—this research advances digital divide scholarship and informs more targeted, age-friendly support and training strategies.

We recruited Chinese adults aged 55 years and older through online platforms and offline communities in Nanjing, China. Inclusion criteria were (1) age of ≥55 years, (2) sufficient cognitive ability to complete the questionnaire independently or with assistance, and (3) the ability to understand the questionnaire content. The target sample size was determined based on recruitment feasibility during the predefined data collection period. To further evaluate whether the achieved sample size was adequate for the proposed moderated mediation analyses, we conducted a sensitivity analysis using G*Power 3.1 (Franz Faul). Because the hypothesized involved first-stage moderation, the analysis focused on the incremental variance explained by the interaction term in a linear multiple regression model (fixed model, R² increase). Each first-stage regression included 3 predictors and 9 covariates, resulting in a total of 12 predictors. With the final sample of 212 participants, α=.05, and power=.80, the study was capable of detecting an incremental interaction effect of approximately f²=.037 or larger. Thus, the achieved sample provided sufficient sensitivity to detect relatively small moderation effects in the proposed model.

A total of 230 questionnaires were collected between June and October 2024, thus yielding 212 valid responses for analysis after data screening. No missing data were present in the final dataset, as all items were set as mandatory in the online questionnaire, while trained volunteers assisted participants with questionnaire completion during offline data collection.

This study was approved by the Research Ethics Committee of the School of Aging Services and Management, Nanjing University of Chinese Medicine (number 20240523) and was conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent before participation. No personally identifiable information was collected, thereby protecting participant privacy and confidentiality. Participants received either RMB 7 (approximately US $1) or a small gift of equivalent value as compensation for their participation.

All measures were adapted from validated scales and translated via forward-backward translation to ensure linguistic and cultural validity for Chinese older adults. Full item wording, outer loadings, reliability and validity statistics for all constructs are reported in Table 1. DTS was assessed using 3 items with 5-point ordered category and 6-point Likert-type response formats; scores from the 6-point scale were linearly rescaled to a 5-point metric for comparability, with higher scores indicating greater DTS. Instrumental and hedonic technology use were each measured using 2 items based on the frequency of use over the past week, with responses recorded on 5-point scales and higher scores indicating more frequent use. Technophobia was measured using Sinkovics et al [19] 13-item scale with a 6-point Likert-type agreement scale. Reverse-coded items were recoded prior to analysis, with higher scores indicating greater technophobia. All constructs were modeled as latent variables in the subsequent analyses. Covariates included age, marital status, residence, education, economic status, functional status, subjective health, number of children, and social activity participation, with their coding and categories reported in Table 2. Gender was coded as 0 for men and 1 for women.

Procedural and statistical controls were implemented to address potential common method bias that might arise from self-reported data. Procedurally, a pilot test with older adults improved item clarity and reduced bias. The questionnaire items were phrased in neutral language to avoid leading responses. Statistically, Harman single-factor test was performed via exploratory factor analysis with all measurement items. The results indicated that no single factor accounted for more than 50% of the total variance. Interconstruct correlation coefficients ranged from 0.154 to 0.814 with all values being below the 0.90 threshold [67]. Variance inflation factors (VIFs) for all constructs were between 1.054 and 3.261, which were below the critical value of 3.3 [68], thus eliminating severe multicollinearity.

Figure 2 illustrates the final partial least squares structural equation modeling (PLS-SEM) model with standardized path coefficients.

PLS-SEM was conducted in SmartPLS version 4.1.1.4 (SmartPLS GmbH) to estimate the measurement and structural models examining how DTS and technophobia shape older adults’ technology engagement. PLS-SEM is well suited to this study because it accommodates complex models with multiple latent constructs, is robust to nonnormal survey data, and performs well with relatively small samples. Therefore, model parameters were first obtained using the partial least squares (PLS) algorithm. Statistical inference for all hypothesized effects—including direct paths, the moderation and indirect (mediated) effects—was then performed via a nonparametric bootstrapping procedure with 5000 resamples. For the moderated mediation hypothesis, conditional indirect effects were computed at different levels of the moderator, and their significance was evaluated using 95% bias-corrected and accelerated (BCa) bootstrap CIs. Effects were considered significant when the corresponding 95% CI did not include zero. This approach is consistent with recent applications of moderated-mediation designs examining sex or gender-differentiated pathways in health-related outcomes [69,70].

Participants’ demographic characteristics are reported in Table 2. The sample (N=212) had a mean age of 66.96 (SD 6.96; range 55‐86) years and was predominantly women (142/212, 68.4%). Most respondents lived in urban areas (203/212, 95.8%) and were married or cohabiting (192/212, 92.9%). In addition, 68.9% (146/212) reported having volunteer service experience, indicating active social participation among the sample. Regarding family structure, most reported having 1 or 2 children. Overall health was generally moderate, with over 40% (95/212) reporting no or minimal limitations in daily activities and most rating their health as fair or good. Educational attainment was relatively low, with nearly 60% (126/212) reporting fewer than 12 years of schooling and about one-fifth having completed higher education. The digital devices used by participants varied, with smartphones being the most commonly owned device (187/212, 88.2%), followed by smart televisions (48/212, 22.6%) and smartwatches or wearable devices (26/212, 12.3%). Regarding DTS, participants reported support from both same-age peers and younger adults, with 25.9% (55/212) reporting more than 10 same-age technical supporters and 26.4% (56/212) reporting more than 10 younger-adult technical supporters; only 12.7% (27/212) and 14.6% (31/212), respectively, reported having no such supporters.

Table 1 reports factor loadings, reliability, and convergent validity. Most outer loadings exceeded the recommended threshold of 0.708, ranging from 0.661 to 0.916, indicating adequate indicator reliability [71]. Convergent validity was supported, with Cronbach α ranging from 0.668 to 0.929, composite reliability (CR) from 0.832 to 0.946, and average variance extracted (AVE) from 0.508 to 0.779; all CR values exceeded 0.70 and AVE values exceeded 0.50 [72]. Discriminant validity was established using the Fornell–Larcker criterion and Heterotrait-Monotrait ratio (HTMT). As shown in Table 3, the square roots of AVE exceeded interconstruct correlations, and all HTMT ratios were below 0.90 [67].

Using a cross-sectional survey of community-dwelling older adults and PLS-SEM for hypothesis testing, the current research found that DTS lowers the technology-inconvenience dimension of technophobia, and this reduction subsequently promotes instrumental—though not hedonic—technology use, especially among older men. By contrast, DTS was not significantly related to the personal-failure or human-machine ambiguity dimensions, but higher human-machine ambiguity was indeed linked to lower IU. These results refine theory by specifying a dimension-specific and gendered mechanism, and offer practical guidance to bridge the digital divide.

The first finding of this research is that not all fears under the umbrella of technophobia equally predict older adults’ actual technology use. Older individuals’ instrumental (vs hedonic) use of technology was significantly influenced by their fear of inconvenience and by human-machine ambiguity but not by their fear of personal failure. Our finding is consistent with other research that has shown that extrinsic motivations, especially technology usefulness, often outweigh intrinsic motivations in determining the usage of involuntary technologies [73,74]. On the other hand, the finding from the dimension of human-machine ambiguity aligns with Smrke et al [27] findings, that older adults who are uneasy about technology’s role tend to stay away from new health technologies.

Fear of personal failure did not significantly predict either instrumental or hedonic technology use among older adults in our study. One possible explanation is that our survey captured a general, self-reported belief about technological competence rather than a contextually primed perception [75]. Research on age stereotypes suggests that the activation of negative self-stereotypes is often context dependent and more likely to occur when salient age-related cues are present [76]. As a result, older adults may not explicitly endorse negative self-evaluations in a generic survey setting, which could attenuate observed associations with actual technology use.

The key finding is that the effect of DTS on technophobia varies by dimension and gender. Daily support significantly reduced fear of inconvenience but did not significantly change fear of personal failure or human-machine ambiguity. These latter 2 dimensions are more closely related to internal psychological concerns than to the instrumental difficulties captured by fear of inconvenience. Previous research has indicated that fear of personal failure is often rooted in internalized age stereotypes about older adults’ technological incompetence [6] and typically requires the development of self-efficacy through structured learning [77]. The nonsignificant association observed in the current study differs from findings reported in some intervention studies [53,54]. One possible explanation is that older adults may respond differently to different forms of technical support [53,54]. Informal, immediate help may function primarily as instrumental problem-solving, whereas structured programs may be more capable of changing underlying confidence and motivation [74,78]. Similarly, human-machine ambiguity reflects broader attitudinal or ideological concerns about technology and its potential to replace or reduce human interaction [42]. Because TS typically targets immediate operational issues rather than these deeper beliefs, its lack of a significant association with human-machine ambiguity is understandable.

More importantly, this research uncovered that the effect of DTS on reducing the fear of inconvenience is more salient among older men than among older women. Our research resonates with existing findings, that men are more likely to address the instrumental aspect of social support than women, thus the extrinsic motivation-related fear is affected [59]. However, our research did not yield significant gender differences in the effect of technical support on fear of personal failure [63,64]. This outcome may also support the point proposed above that older adults generally seek technical support primarily to solve problems (ie, extrinsic motivation) rather than to obtain internal psychological relief (ie, intrinsic motivation). Notably, additional analyses provided no evidence that gender moderated the overall associations between DTS and either instrumental or hedonic technology use. This pattern suggests that gender may function more specifically as a boundary condition of older adults’ psychological responses to technical support rather than as a general moderator of its behavioral association with technology use.

On the other hand, DTS has a significantly positive effect on older adults’ IU of technology but no significant effect on HU. This finding is consistent with prior observations that many older adults view technology, such as mobile apps, primarily as a tool for social engagement rather than a source of entertainment [79]. Therefore, DTS, by increasing technology accessibility and decreasing frustration, demonstrates the concrete utility of technology in everyday tasks [12]. Meanwhile, many older people have well-established leisure routines outside of digital technology; even with daily assistance available, they may not be motivated to explore new entertainment technologies [79].

The gender-specific finding demonstrated that DTS boosts instrumental technology use by reducing fear of inconvenience to a greater extent among older men than among older women. This result is in line with the established TAM; that is, technical support serves as an external environmental factor that affects one’s perception toward technology per se, which is further a key driver of technology adoption [34]. It bridges 3 dimensions of technophobia and 2 classic focal predictors of technology use (ie, technology usefulness and ease of use), as well as incorporates gender as a moderator. As such, it generalizes TAM into older adults and enables a nuanced understanding of how the social environment can influence older men and women’s technology adoption behavior differently.

Our findings suggest a shift in how interventions to promote technology use among older adults are designed. Older adults are often stereotyped as technophobic and lacking confidence in their ability to use technology, and interventions have therefore tended to emphasize intrinsic empowerment (skills training and confidence building). Our findings indicate that more instrumental forms of support may also be important, particularly for technologies that have become difficult to avoid in everyday life. By reducing the inconvenience, effort, and practical burden associated with technology use, DTS may help older adults recognize the functional value of technology and facilitate IU. Accordingly, interventions could place greater emphasis on readily accessible, task-oriented assistance, such as just-in-time support, simplified workflows, default configurations, and age-friendly interface design with larger fonts and icons, simplified navigation, and fewer steps required to complete common tasks [80].

The stronger indirect association observed among men further suggests that older adults may not benefit from the same form of technical support to the same extent. Rather than treating gender itself as the basis for intervention, this finding highlights the importance of identifying the specific barriers that users experience and tailoring support accordingly. For individuals whose technology use is particularly constrained by perceived inconvenience, immediate and task-focused assistance may be especially effective, whereas those whose difficulties stem from other sources of technophobia may require different forms of support, such as structured learning, confidence-building, or greater reassurance about human-technology interaction. Gender may therefore serve as 1 contextual indicator of heterogeneity in how technical support is experienced, but intervention decisions should ultimately be guided by users’ specific needs and sources of technological fear [22]. Such a barrier-specific approach may provide a more precise and sustainable way to promote technology engagement among older adults.

Despite these findings, our research is not without limitations. First, the survey sample was small and skewed toward urban older adults, thus limiting the generalizability of our findings to diverse socioeconomic and rural populations. As the sample was drawn from one country, the magnitude and relative salience of these associations may also vary across cultural and social settings. Second, the cross-sectional design precludes strong causal inference about the proposed relationship between variables. Third, measures relied on self-reports, which are susceptible to recall and social desirability biases. For instance, older men may underreport technophobia because of gendered norms about technological competence. Future research should use larger and more diverse samples, including cross-cultural comparisons, and longitudinal designs. Complementing surveys with objective behavioral indicators, such as smartphone or app usage logs and passively sensed interaction traces, could also provide fine-grained evidence on how daily support translates into reduced technophobia and sustained technology use. Future studies could further examine how emerging forms of online and real-time technical support can translate these findings into practice. As older adults increasingly engage with digital technologies as everyday users, platforms could potentially use user profiles and interaction data to provide more timely and tailored support, including gender-sensitive prompts or assistance. Finally, we failed to find a significant predictor of older adults’ HU of technology; future research can focus on this specific technology domain.

This study examined how DTS relates to older adults’ instrumental and hedonic technology use, with fear-specific technophobia as a mediating mechanism and gender as a moderator. TA was associated with lower fear of inconvenience, which in turn was associated with greater instrumental technology use, and this indirect effect was stronger among men than women. No comparable pathway was found for HU. Theoretically, these findings advance a more differentiated understanding of technophobia by highlighting the importance of examining specific sources of technological fear rather than treating technophobia as a unitary barrier. The absence of a comparable pathway for HU further suggests that different forms of technology use may be shaped by distinct psychological mechanisms and therefore warrant more use-specific investigation. Practically, the findings underscore the value of moving beyond one-size-fits-all approaches to digital inclusion by matching support strategies to the particular barriers older adults experience. Efforts to reduce technophobia should therefore complement generic skills training with targeted assistance that addresses specific sources of fear, practical barriers that can be alleviated through daily support, and differences in how older adults psychologically respond to such support. Such a more targeted approach may help improve the effectiveness of interventions designed to promote sustained and meaningful technology engagement in later life.