Navigating the digital landscape: Relationship between type of internet use and well-being among students—A mediation and moderation analysis

Abstract

Background:

The amount and frequency of internet use are often used to forecast its pros and cons, with mixed findings. After COVID-19, technological advances and sociological upheavals have shown the internet’s ability to support numerous activities, necessitating a change in internet use. Beyond internet use frequency, it has been advised to examine why people use it and how it affects their well-being. Hence, we aimed to investigate the various mediators and moderators in the relationship between the quality of internet use (type of activity) and well-being. Further, we investigated whether user attributes such as initial age and frequency of internet use moderate the influence of internet use quality on problematic internet use, loneliness, and well-being.

Methods:

131 students in Manipur who were 18 years old were included in this cross-sectional study. Self-rated measures of the Adult Internet Usage questionnaire, Internet Addiction Test, and Mental Health Continuum-Short Form assessed internet use, dependence, and well-being.

Results:

Loneliness mediates the association between well-being and internet behaviors related to personal use-self-actualization (BootSE = .53, BootLLCI = .10, BootULCI = 2.18), cultural use-belonging (BootSE =.57, BootLLCI = -2.23, BootULCI = -.02). Average online hours moderate economic use-education, loneliness, and well-being (BootSE =.13, BootLLCI = -.55, BootULCI = -.04).

Conclusion:

To conclude, the effectiveness of internet use affects well-being differently depending on the type of activity. After a certain point, even seemingly good internet activities may have diminishing returns.

INTRODUCTION

Studies on internet use have pointed out that the diversity of internet use poses a challenge for researchers trying to quantify and categorize the use and study the behavior. The latest developments in internet use following the pandemic have only further challenged this.[1] Although there are compelling arguments for how problematic internet use (PIU) is detrimental to well-being, there is also an equally compelling argument for how the internet can improve well-being.[2,3] There is no unanimous call on internet use and well-being. Recent studies show that the general effects of the internet depend on the type of use, showing differential effects on hedonic and eudemonic well-being.[4] Current research and policies on internet use often emphasize the amount and frequency of use in assessing its risks and benefits and in diagnosing internet addiction.[5,6] Government guidelines typically advocate for regulating screen time.[7] Consequently, treatments for PIU mainly aim to decrease usage hours.[8] However, there’s little agreement on evaluating the quality of screen time. Online culture is changing, and information is more readily available, complicating the debate over whether internet use affects mental health.[9] The shift is that the specific activities conducted or made available through the internet may hold an influence rather than internet use per se.[10] Not all online activities have the same effect, positive or negative.[11] Even with healthy internet activities, overuse can lead to problems.[12] Research indicates that the broader concept of PIU is best understood by examining specific online behaviors, such as gaming, shopping, social media use, and viewing pornography, which can lead to problematic use.[13] PIU is more common in individuals who browse aimlessly or use the internet for socializing and entertainment instead of information-seeking.[14,15] However, digital technology was also used in the COVID-19 health delivery system to reduce social isolation and lockdown-induced loneliness during the epidemic.[16] Internet social activities have been shown to help alleviate loneliness and enhance well-being.[17] This suggests that some social internet uses benefit users.[18] Using the internet creatively introduces both challenges and opportunities. While innovative apps can foster self-expression and personal growth, they may also pose difficulties for certain users.[19] However, short video addiction, caused by more recent popular internet platforms, negatively impacts learning motivation and well-being.[20] Further, the pandemic led to a surge in digital communication in younger individuals.[21] Excessive online communication can harm well-being, yet innovative mobile apps supporting physical and mental health show the internet’s great potential.[19] This pointed out that future studies on well-being and online need to be seen in the light of meaningful online activities, i.e., the context.

To complicate internet use, users’ pre-existing attributes have been significant determinants.[19] Younger individuals are more prone to PIU.[22] Studies show that females and older users use the internet more for specific goals and less for diverse activities. Notably, certain online behaviors like viewing pornography among the young and streaming media among older users are linked to problematic use within those specific age groups.[23] Internet use positively correlates with user qualities like living alone.[24] High-status users often use the internet to boost their capital, while low-status users typically seek entertainment, regardless of socioeconomic differences.[11]

Previous studies have seldom explored the relationship between how much and how well people use the internet, underscoring the importance of grasping the complexities of internet usage. As more work, education, and research move online, the simplistic notion that the amount of internet use directly predicts addiction may require reevaluation. There’s a growing call to understand the reasons behind internet use and to adapt treatments for PIU accordingly.[25]

In light of these, we aimed to investigate the relationship between the quality of internet use (type of activity) and well-being. We hypothesize that PIU and loneliness will mediate the relationship between the quality of internet use and well-being. In addition, user characteristics (such as initial age of usage and quantity, or frequency, of the internet) would moderate the effect of quality of internet use on PIU, loneliness, and well-being while maintaining gender as a covariate.

METHODS

Participants

The present study is a cross-sectional study. Participants (N = 131) were students continuing their education at colleges and universities in Manipur. Exclusion criteria were not kept to increase generalizability. Convenient sampling was adopted, and data was collected through Google Forms owing to the ongoing COVID-19 pandemic. Information for informed consent was provided in the Google form. The mean age was 21.78 years (age range 18-32 years, SD = 3.36); the majority were female (n = 83; 64.4%) and undergraduate students (n = 90; 86.7). On average, the participants reported that they first started using the internet at 15.49 years, with an average usage of 6.59 hours daily.

Measures

Apart from socio-demographic and specific internet use-related user attributes questions, the Google form included measures on the quality of internet use, PIU, well-being, and loneliness. Frequency of use in terms of average daily hours spent online and the average daily internet data consumption was collected to measure the quantity of internet use. Attempts were made to reduce common method bias through procedural remedies.[26] Along with maintaining the respondent’s anonymity, the form instructed that there were no right or wrong answers. This was in an attempt to reduce evaluation apprehension. Also, certain measure items were improved by using a few modest word changes to fit the current sample and including examples for complex sentences. After these minor revisions, the scale was forwarded to three subject experts for content validity.

User attributes

Apart from user demographics, information such as living arrangements, access to type, and various internet-supported devices were also collected. Living arrangements marked whether they were living with parent(s), living alone, living with a partner with no children, living with a partner and children, or living with non-relatives. Access to internet-supported devices such as a computer (desktop/laptop), smartphones, tablet e-readers (e.g., iPad, Kindle, etc.), game consoles, and wearables (e.g., Smartwatches, Fitbit, and Apple watch) were rated on a scale of several times a day, daily, weekly, monthly, or never.

Quality of use

The shorter version of the self-rated measures of the Adult Internet Uses questionnaire[27] was adapted for the suitability of the present study to measure the quality of use. The scale measures the quality of use in four fields with subscales within each field—Personal use (Health and lifestyle, Leisure, and Self-actualization), Social use (Formal networks and Personal networks), Cultural use (Belonging and Identity), and Economic use (Education, Property, and Finance). Questions under social media creativity were added to accommodate the new activities possible on online platforms. The questions were related to making status updates and stories and creating video content through video platform apps. The adapted scale’s reliability was high (Cronbach’s Alpha = .92). Responses are rated on a scale of 1-7, ranging from Never, Less than once a month, Monthly, Weekly, Daily, Several times per day, and I do not know.

PIU

Only the total score of the Internet Addiction Test[28] was used in the model to measure PIU. It is a 20-item scale to measure dependence on the internet based on statements describing the last month. Responses are rated on a 5-point scale ranging from 0 (not applicable) to 5 (Always). The scores indicate either a normal level of internet use (0 = 30), mild level of internet addiction (31-49), moderate level of internet addiction (50-79), or severe dependence upon the internet (80-100). The adapted scale’s reliability was high (Cronbach’s Alpha = .93).

Well-being

The mental health continuum-short form[29] was adapted for the suitability of the present study to measure well-being in the model. It is a 14-item scale measuring dimensions of well-being where responses are provided on a 6-point Likert scale ranging from 0 (never) to 5 (every day). The model made use of only the continuous scoring of the scale. The scale’s reliability was high (Cronbach’s Alpha = .92).

Loneliness

The continuous scoring of De Jong Gierveld’s 6-item loneliness scale[30] was used to measure loneliness. This is a short form of the De Jong Gierveld 11-item loneliness. Responses are generally provided in three categories—yes, more or less, and no. The scores are accorded differently for positive items and negative items.

Ethics

This paper is part of a doctoral study. Ethical approval of the study has been obtained from the Institutional Human Ethics Committee vide letter no. MZU/HEC/2023/016.

Data analysis

Apart from descriptive analysis, we used multiple correlations to determine the relationships between the variables under study. We used Cohen’s classification[31] to interpret multiple correlations. SPSS version 25 was used for the data analysis.

All tests for mediation and moderation were computed using PROCESS macro v4.1.[32] PROCESS Model 4 was used to test our first hypothesis of mediation. Conditional process analysis using PROCESS Model 10 was performed to test for moderation. This analysis combines mediation and moderation analysis, intending to describe the conditional nature of how the variables in the study affect the others. We ran multiple sets of the analysis (PROCESS Model 10) to test the independent variable’s direct and indirect effects on the dependent variable, conditional on two moderators simultaneously. As illustrated in Figure 1, PROCESS Model 10 has three regression sub-models testing whether the quality of internet use (X) on well-being (Y) is mediated by internet addiction (M1) and loneliness (M2) while taking into account possible interaction with another set of variables quantity of use (W) and initial age of use (Z), all the while keeping gender as a covariate. There are a total of 11 types of quality of use (Personal use—Health and lifestyle, Leisure, Self-actualization; Social use—Formal networks, Personal networks; Cultural use—Belonging, and Identity; Economic use—Education, Property, Finance; social media creativity) being specified. In both the models, to estimate all the direct and indirect effects of the number of Independent variables, the PROCESS model was executed 11 times for each model, each time putting one of the Qualities of internet use as X in the model and the remaining types of quality of use as covariates to rule out any spurious associations. We used 5,000 bootstrap samples and a 95% confidence interval for determining the analysis. All independent variables and both moderators were mean-centered before analysis.

Figure 1.

Conceptual Diagram of the Moderated Mediation Model

RESULTS

Preliminary analysis

Table 1 details the participants’ descriptive scores of the variables in the study. There was also a significant difference in internet dependency between males and females (t_69.792 = 3.06, P < .005). The average female reported IAT scores lower by 10.94 than the average male. There was a statistically significant weak positive correlation [see Table 2] between social use-personal network, culture use-belonging, and well-being, r = .24, N = 131, P < .01; r = .17, N = 131, P < .05. While there was a statistically weak negative correlation between personal use-leisure and well-being, r = -.18, N = 131, P < .05). We also found a statistically significant moderate positive correlation between culture use-belonging, culture use-identity, economic use-property, and internet addiction, r = .37, N = 131, P < .01; r = .35, N = 131, P < .01; r = .32, N = 131, P < .01 respectively. While, there was a statistically weak positive correlation between personal use-health and lifestyle, social use-formal network, economic use-finance, social media creativity, and internet addiction, r = .27, N = 131, P < .01; r = .21, N = 131, P < .01; r = .28, N = 131, P < .01 respectively. There was also a statistically significant weak negative correlation between well-being and loneliness, r = -.28, N = 131, P < .01).

Table 1.

Descriptive Data of the Variables in the Study

Variables		M		SD		Variables		M		SD
Personal use—Health and lifestyle		21.78		3.36		Economic use—property		2.13		1.23
Personal use—Leisure		4.22		0.94		Economic use—finance		1.86		0.97
Personal use—self-actualization		3.58		1.15		Social media creativity		2.69		1.02
Social use—formal network		2.88		1.06		Well-being		38.14		14.22
Social use—personal network		3.84		1.05		Internet addiction		18.28		0.49
Cultural use—belonging		2.47		0.98		Loneliness		4.03		1.60
Cultural use—identity		2.94		1.16		Average hour of use		6.59		3.11
Economic use—education		3.19		1.10		Initial age of use		15.48		2.97

Table 2.

Correlations of different variables studied

Variables				1		2		3		4		5		6		7		8		9		10		11		12		13		14		15		16
1		Personal use- Health & lifestyle		1
2		Personal use- Leisure		0.29**		1
3		Personal use- self-actualisation		0.29**		0.41**		1
4		Social use- formal network		0.40**		0.31**		0.57**		1
5		Social use- personal network		0.27**		0.29**		0.41**		0.49**		1
6		Cultural use- belonging		0.36**		0.16		0.32**		0.44**		0.35**		1
7		Cultural use- identity		0.20*		0.30**		0.38**		0.54**		0.32**		0.58**		1
8		Economic use- education		0.35**		0.30**		0.40**		0.42**		0.42**		0.44**		0.49**		1
9		Economic use- property		0.31**		0.17		0.24**		0.29**		0.14		0.49**		0.35**		0.31**		1
10		Economic use- finance		0.26**		-0.03		0.22*		0.37**		0.03		0.41**		0.37**		0.12		0.44**		1
11		Social media creativity		0.42**		0.17		0.31**		0.47**		0.44**		0.46**		0.42**		0.42**		0.40**		0.25**		1
12		Well-being		0.09		-0.18*		-0.05		0.11		0.24**		0.17*		0.01		0.08		0.06		0.15		0.17		1
13		Internet addiction		0.27**		0.07		0.09		0.21*		0.04		0.37**		0.35**		0.14		0.32**		0.25**		0.28**		0.00		1
14		Loneliness		0.07		0.02		-0.16		0.04		-0.12		0.06		0.02		-0.16		-0.05		-0.02		-0.03		-0.28**		0.16		1
15		Average hour of use		0.08		0.08		0.15		0.19*		0.05		0.15		0.27**		0.12		0.07		0.02		0.14		-0.12		0.29**		0.03		1
16		Initial age of use		-0.22*		0.01		-0.12		-0.22*		-0.03		-0.13		-0.23*		-0.11		-0.06		-0.05		-0.10		-0.06		-0.25**		-0.07		-0.15		1

*P<.05, **P<0.01

Mediation analysis

The initial mediation model analysis [see Figure 2] reported that personal use—self-actualization did not predict internet addiction (b = -1.03, SE = 1.68, P = .541). However, personal use—self-actualization created a decrease in loneliness (b = -.35, SE =0.16, P = .024) which consequently influences well-being (b = -2.58, SE =0.77, P = .001). Personal use—self-actualization also significantly (negative) directly predicted well-being (b = -2.73, SE = 1.32, P = .041). Specific indirect effects were further examined and the results supported that loneliness mediates between personal use—self-actualization and well-being, and reported as statistically significant (IND2: Personal use-self-actualization → Loneliness → Well-being [indirect effect] =.92, BootSE =.53, BootLLCI =.10, BootULCI = 2.18). However, the first indirect effect analyzing the effect of internet addiction between personal use—self-actualization and well-being was reported as statistically not significant (IND1: Personal use-self-actualization → Internet addiction → Well-being [indirect effect] =.33, BootSE =.14, BootLLCI = -.23, BootULCI =.37). The total indirect effect of the two specific indirect effects (IND1 and IND2) were also tested and found to be statistically significant (total IND1 and IND2 [indirect effect] = -.94, BootSE =.55, BootLLCI =.08, BootULCI = 2.23). After examining specific indirect effects, we compared the two indirect paths using multiple contrast tests. The contrast testing found that the two mediation pathways are significantly different (IND1 vs IND2 [effect] = -.07, BootSE =.04, BootLLCI = -.18, BootULCI = -.00).

Figure 2.

Parallel Mediation Models Presenting the Direct and Indirect Effects of Quality of Use a) Economic Use-Education (a₁= -1.03, b₁= -.03, a₂= -.35*, b₂= -2.58*, c’ = -2.73*) b) Cultural Use-Belonging on Well-Being (a₁ = 4.21*, b₁= -.03, a₂= -.38*, b₂= -2.58*, c’ = -1.88). Note: * P < .05

Cultural use-belonging was also reported to be a significant (positive) predictor of internet addiction (b = 4.21, SE = 2.03, P = .040), although internet addiction had no significant effect on well-being (b = 1.8751, SE = 1.60, P = .245) when controlling for the type of use. Similarly, the results also support that using the internet activity related to cultural use-belonging was a significant (positive) predictor of loneliness (b = .38, SE = .19, P = .046) which subsequently (negative) predicted well-being (b = -2.58, SE = 0.07, P = .001). Specific indirect effects were examined, and although internet addiction did not mediate between Cultural use-Belonging and well-being (IND1: Cultural use-Belonging → Internet addiction → Well-being [indirect effect] = -.13, BootSE =.34, BootLLCI = -.95, BootULCI =.43), the results supported that loneliness mediates between Cultural use-Belonging and well-being, and reported as statistically significant (IND2: Cultural use-Belonging → Loneliness → Well-being [indirect effect] = -.98, BootSE = .57, BootLLCI = -2.24, BootULCI = -.02). The total indirect effect of the two specific indirect effects (IND1 and IND2) were also tested and found to be statistically significant (total IND1 and IND2 [indirect effect] = -.08, BootSE =.62, BootLLCI = -25, BootULCI = -.04). Finally the results supported a statistically insignificant contrast effect comparing IND1 vs IND2 (IND1 vs IND2 [effect] = -.06, BootSE =.05, BootLLCI = -.03, BootULCI =.15).

Continuing further, even though personal use-leisure significantly (negative) predicted well-being (b = -3.05, SE = 1.42, P = .03), there was no mediation by either internet addiction (IND1: personal use-leisure → Internet addiction → Well-being [indirect effect] = -.02, BootSE =0.16, BootLLCI = -.35, BootULCI = 0.35), or loneliness (IND2: personal use-leisure → Loneliness → Well-being [indirect effect] = -.38, BootSE = .42, BootLLCI = -1.34, BootULCI = .35). Similarly, even though social use-personal network significantly (positive) predicted well-being (b = 3.67, SE = 1.42, P = 0109), there was no mediation by either internet addiction (IND1: social use-personal network → Internet addiction → Well-being [indirect effect] = -.072, BootSE = .204, BootLLCI = -.328, BootULCI =.55), or loneliness (IND2: social use-personal network Loneliness Well-being [indirect effect] = -.73, BootSE =.48, BootLLCI = -.05, BootULCI = 1.84).

Conditional process analysis

The model test results reported that Economic use-education decreased loneliness (b = -0.37, SE =0.15, P = .016), which increased the well-being (b = -2.58, SE =0.80, P = .002) of the user. The interaction term (Economic use-education × Average hours) was significant in the model (b = .08, SE =.04, P = .041). The index of partial moderated mediation shows that the effect of loneliness between economic use-education and well-being is being moderated by average hours (IPMM = -.22, BootSE =.13, BootLLCI = -.55, BootULCI = -.04). This indicates moderated mediation in the mechanism represented by Economic use-education loneliness well-being. On further probing [see Figure 3], it was indicated that those who reported lesser than average quantity of internet use (average hours of internet use) experienced a greater effect of economic use-education on loneliness, as compared to those who reported average hours or higher than average hours of use of the internet. The influence decreases as the average hours of internet use keep increasing.

Figure 3.

Graph Showing Simple Slope of predicted values of Y (Loneliness) and X (Economic Use-Education) Across Levels of Average Hours of Internet Use. Note: EUEAVG: Economic use-education average score; DJLTS: Loneliness test score

Further, the interactions (person use-health and lifestyle × age of initial use) were found to be significant on Well-being (b = 1.17, SE = .52, P = .025) in the model, meaning that the effect of personal use-health and lifestyle on well-being depend upon initial age of internet use [Figure 4]. However, no evidence exists for moderated mediation by the initial age of internet use. Average hour is a significant positive predictor of the effect of quality of internet use on internet addiction (b = 1.33, SE = .48, P = .006). However, there were no significant moderations on the effect of internet addiction on the quality of use and well-being. Loneliness is also a significant negative predictor of well-being, controlling for quality of use (b = -2.68, SE = .80, P = .001), although there was no evidence for a main or indirect effect. There was also a positive direct effect of the social use-personal network (c’ = 3.50, SE = 1.42, P = .016) and a negative direct effect of both personal use-leisure (b = -30, SE = 1.43, P = .040) and personal use-self-actualization (b = -2.69, SE = 1.34, P = .071) on well-being. However, we did not find evidence for moderated mediation, as the indirect effects were insignificant. Further, personal use-self-actualization had a direct negative effect (b = -.36, SE =.16, P = .026) on loneliness, while cultural use—belonging (b = .40, SE =.19, P = .039) and Social use-formal network (b = .42, SE =.20, P = .037) had a positive effect on loneliness.

Figure 4.

Graph Showing Simple Slope of predicted values of Y (Well-Being) and X (Personal Use-Health and Lifestyle) Across Levels of Age of Initial Internet Use. Note: PUHLAVG: Personal use-health and lifestyle average score; MHCTS: Well-being total score

DISCUSSION

This research endeavor was conducted with the purpose of investigating the effects of PIU, loneliness, and user characteristics on the relationship between the quality of internet use and well-being. Quality of use, specifically personal use-leisure, personal use-health and lifestyle, personal use-self-actualization, and social use-personal network, directly impacted well-being more than other activities. Also, we found that loneliness mediates the relationship between the quality of internet use and well-being. Specifically, activities that inherently lower loneliness, such as personal use-self-actualization and economic use-education, positively impacted well-being. Furthermore, we found moderated mediation by the quantity of internet use. Also, there was a significant interaction between the initial age of usage and use for personal leisure and well-being.

As current works of the literature suggest, we found that internet use does not directly affect well-being when controlled for quality of use and user attributes. Instead, our findings indicate that the activity engaged was of interest to whether the use will impact well-being. Engaging in certain internet activities was found to have a more direct impact on well-being than others. This is consistent with recent studies showing that quality of use rather than just quantity is associated with a negative impact on well-being.[33] Findings suggest that some activities specifically promote well-being either directly or indirectly. Internet-based social activities that help people stay in touch with their family or friends and use the Internet for activities related to education or learning opportunities were found to increase well-being. Studies have pointed out that internet activities that ease social interactions promote well-being more than those that do not.[34] The extent of social media’s positive and negative impact on well-being depends on several factors, including user characteristics and the extent and nature of internet use.[35] Previous findings have supported that internet activities related to social use promote social engagement and reduce loneliness, thereby increasing well-being over time.[36] Another activity that was associated with decreased well-being was personal use of leisure. These are activities that specifically relate to those undertaken by the user on their own as a form of entertainment. Such activities have been consistently linked with poor well-being and higher internet addiction.

Contrary to the literature, there was no correlation between PIU and well-being in our study. Our first hypothesis was rejected as PIU did not mediate the relationship between the quality of internet use and well-being.

While engaging in activities that promote self-actualization negatively affects well-being, loneliness mediates this. In the study context, self-actualization activities involve collecting information by communicating with people online about concerns or matters that are important to them, possibly improving self-esteem. Such activities were associated with a decrease in loneliness, which, in turn, increases well-being. Although internet activities that facilitate a feeling of belonging to a community did not directly impact well-being, they were found to increase loneliness, negatively impacting the user’s well-being. What is noteworthy here is that loneliness significantly (negatively) predicts well-being. So, any internet activity that lowers loneliness ultimately improves well-being, albeit it may negatively impact well-being on its own. So, our hypothesis that loneliness will mediate the relationship between quality of use and well-being is partially accepted. This is in conjunction with studies that show that the relationship between internet use and loneliness is dynamic, with the direction of the relationship determined by one’s motives for internet use.[37] Taken together, these results suggest that specific internet activities can decrease loneliness and consequently increase well-being.[38,39]

A significant interaction we found in our study was that individuals who started using the internet at a later age than average benefited more from using the internet for activities related to health and lifestyle than those who started using the internet at an early age. The former group was found to have higher well-being with increased activities involving health and fitness. However, we did not find significantly moderated mediation by the initial age of internet usage. In general, the early age of exposure and internet use has been significantly linked to heavy internet use, which is worrying.[40]

The pandemic saw an increase in the digitalization of education.[1] This includes a collaborative effort from educational institutes and governmental agencies to continue access to education seamlessly.[41] As a result, students may spend more time on the internet with activities relating to learning. This could be one of the reasons why, among the quality of internet use of interest in the study, it was the only type of internet activity significantly moderated by the average hours of use. Hence, our hypothesis that frequency of use would moderate the effect was also partially accepted. We found that using the internet for education-related activity was associated with a decreased level of loneliness, which increased levels of well-being, but this is conditional on the hours of use. This is consistent with study findings that show that even excessive use of apparent healthy internet activities can become problematic.[12] This is in line with studies showing that new activities that seemingly promote well-being and coping with loneliness have also started becoming problematic.[20,42]

Notwithstanding the relatively limited sample, this work offers valuable insights into understanding how internet use impacts well-being differently. Based on our findings, we would like to highlight the need to reassess the fundamentals of internet use and its consequences on our lives. Such would help make better predictions and better choices for healthy internet use. Understanding the interaction of the type and quantity of internet use will raise attention to changing the existing treatment protocol for PIU. A more efficient model of internet use can precisely point out behavior markers and create efficient guidelines for clinicians and users alike to deal with the looming internet addiction crisis.

Our study had certain limitations. The study’s measurement of the quality of internet use needs to be revised. The quality of internet use measurement might be strengthened by including additional activities (using AI for academics and other purposes, telehealth, especially online psychotherapy) that are becoming commonplace due to digitization. Future studies need to keep updating activities coming onto mainstream internet use. The study also needs to consider the function of digital skills. Digital skills will influence the type of online activities or type of internet use that users engage in, thereby affecting well-being. Further, as the quantity/frequency of hours spent online is self-reported, there is a risk of underreporting and over-reporting for average usage hours. There may be limited generalizability of the study results due to the sample size.

CONCLUSION

The present paper set out to explore the relationship between the quality of internet use and well-being. The study has confirmed that the quality of internet use affects well-being differently. The results showed that loneliness mediates the relationship between quality of internet use (personal use-self-actualization, cultural use-belonging) and well-being, and quantity of use (average hours of internet use) moderated the effect of loneliness between economic education and well-being. The evidence from the study suggested that internet activities that inherently lower loneliness (personal use-self-actualization, economic use-education) promote well-being more than those that do not. Further, apparently, healthy internet activities promote well-being only up to a certain level. Higher than-average use of healthy internet activities can become problematic and will no longer promote well-being. Also, the findings suggest that users who start using the internet activity at a later age experience more well-being from engaging in activities related to health and lifestyle. Therefore, studies on internet use need to be seen from more than just a frequency perspective but a holistic one that is mindful of the context and the environment.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.