Function-Driven Heterogeneity of Digital Tools and Limits of Single-Mediation Models in Research on Older Adults’ Mental Health: Reply to the Comments of Abel and Amelia and Further Thoughts [Response to Letter]

Dear editor

Recently, we received comments from Perez-Gonzalez and Lopez-Pelaez¹ on our article. Our article analyzed the digital use behaviors of older adults with chronic diseases using 2023 data from the China Longitudinal Aging Social Survey (CLASS) and proposed that social isolation plays an important mediating role in the relationship between internet and smart device use and the mental health of older adults with chronic diseases. During the analysis, however, we identified a finding that ran counter to our research hypothesis: not all smart devices help reduce social isolation; in contrast, smart sleep-monitoring devices appeared to aggravate it.²

Upon discovering this finding, we reflected on it in the Discussion and Limitations sections from two perspectives. First, the complexity of the digital technology construct gives rise to heterogeneity in its effects. The four indicators of digital technology use—internet access, internet-based medical care, smart wristbands, and smart sleep-monitoring devices—mainly involve two domains: internet use and smart device use. The effect of internet use on mental health and the mediating pathway of social isolation in this relationship are well established; however, the mediating effects of smart device use differ across devices. Smart sleep-monitoring may aggravate social isolation, possibly because older adults with chronic diseases and concurrent sleep problems are more inclined to purchase and use such devices,² a possibility consistent with the comments of Perez-Gonzalez and Lopez-Pelaez.¹ Second, the variable relationships examined here are correlational rather than causal. Unlike causal analyses, correlation analyses based on cross-sectional data may fail to account for other confounding variables.

Perez-Gonzalez and Lopez-Pelaez noted that the effects of specific digital technologies on mental health are heterogeneous and that social isolation, as a single mediating pathway, cannot adequately explain the influence of digital technology on mental health. They further argued that the conceptual heterogeneity of digital technology imposes considerable limitations on a single mediating-pathway hypothesis. We consider this observation pertinent. As digitalization sweeps across the globe, digital technology is becoming increasingly integrated into daily life, and its meaning is becoming increasingly complex.

As Perez-Gonzalez and Lopez-Pelaez pointed out, conceptual heterogeneity renders a single mediating-pathway hypothesis inappropriate, even though we employed only the concept of “use of the internet and smart devices” rather than “digital technology.” Building on this discussion, we further contend that conceptual heterogeneity originates from functional heterogeneity; the concept itself is not inherently heterogeneous. From an operationalization perspective, conceptual operationalization is a unidirectional and irreversible process, yet it determines the quality of measurement and thereby influences the estimation of relationships between variables. The unexpected findings in our article primarily stem from the fact that the mediating variables we introduced produced heterogeneous pathways in the relationship between the independent and dependent variables. From this perspective, future research should either further narrow the scope of the concept or avoid overly narrow hypotheses involving a single mediating mechanism.

The relationship between digital technology and health is becoming increasingly complex and multilayered. On the one hand, digital technology use can improve health, including mental health;³ on the other hand, health status influences digital technology use,⁴ and both relationships have been supported by numerous studies. Research has found that individuals in better health tend to choose digital devices or technologies that maintain or enhance health, whereas those in poorer health tend to use devices or technologies that alleviate disease symptoms.⁵ From this standpoint, digital technologies and devices have different instrumental meanings as health-promoting tools: for healthy individuals, digital devices signify a healthy lifestyle, whereas for individuals in poorer health, they help combat disease or unhealthy conditions. Compared with physical health, the relationship between mental health and digital technology is more complex, and this complexity may stem from three sources. First, the effect of digital technology on mental health is closely related to physical health.⁶ Second, digital device use is subject to the digital divide,⁷ which is influenced by antecedent variables. Third, digital devices may induce technology dependence, which in turn exerts negative effects on mental health.⁸ Accordingly, the effects of digital technology on mental health are not always positive, and its mediating pathways are more intricate.

We thank Perez-Gonzalez and Lopez-Pelaez for their comments and for the reflection they provoked. To better understand the relationship between digital technology and mental health, future research may proceed in two directions. First, a systematic analysis of digital technologies and devices and their effects could adopt meta-analytic methods to define digital technology and its subtypes, thereby directly addressing the root of the conceptual heterogeneity of digital technology. Second, the specific contexts in which the causal relationship between digital technology and mental health operates should be systematically examined. The contexts in which the mediating pathways of digital technology’s psychological effects take effect are specific and distinctive. Experimental studies or longitudinal tracking data can distinguish among different contexts, and the underlying logic can be analyzed through a typology of these effect contexts.

Disclosure

The authors report no conflicts of interest in this communication.