December 2025
Screen time and subjective well‑being
Insights from a few countries worldwide
With people spending several hours each day on smartphones, computers, and other connected devices (i.e. for work, social interaction, entertainment, health monitoring, and more), digital engagement plays an increasingly important role in shaping daily life. While digital tools can enhance connection, convenience, and access to information, higher levels of personal screen time influence well-being outcomes.
This analysis explores these dynamics using recent cross-country poll data for 14 countries (Australia, Brazil, Canada, France, Germany, India, Italy, Japan, Korea, Mexico, the Netherlands, South Africa, the United Kingdom, and the United States), collected in early 2025 in collaboration with Cisco as part of the OECD Digital Well-being Hub.
The empirical analysis uses binary logistic regression models to examine the relationship between digital engagement and poor well-being outcomes. Digital engagement has complex effects on well-being. Individuals who spend more than 5 hours per day on screens for personal purposes show markedly higher odds of poor well-being outcomes, compared to those who use screens moderately for 1-3 hours daily. However, lifestyle and socio-economic factors are more predictive than screen time alone: sleep deprivation, financial hardship, and low physical activity remain the strongest predictors of poor well-being.
Certain groups appear more vulnerable to extended screen time. When long screen use is combined with factors such as loneliness or unemployment, the likelihood of poor mental well-being increases further, highlighting the importance of context in understanding the impact of digital engagement.
Digital technologies have become integral to daily life, transforming how people work, communicate, access services, and spend leisure time. As digital engagement becomes increasingly embedded in everyday routines, concerns are rising over its impacts on subjective well-being. Early studies focused on excessive or problematic use, but recent research has shifted toward examining habitual digital behaviors and their nuanced effects across different populations.
Evidence increasingly shows that the relationship between screen time and subjective well-being is complex and contingent on various factors. While high or unbalanced screen time is consistently associated with lower mental health outcomes (Santos et al., 2024[1]; Madhav, Sherchand and Sherchan, 2017[2]), moderate and purposeful use may support well-being, particularly when fostering meaningful social connections or activities (Przybylski and Weinstein, 2017[3]; Twenge and Campbell, 2018[4]). The “Goldilocks hypothesis” posits that digital use benefits well-being only within a moderate range. Research also highlights how effects vary by age, gender, and socioeconomic status (Lee and Zarnic, 2024[5]), and that social media use, in particular, may be detrimental when emotionally intense or displaces real-life social interaction (Riehm et al., 2019[6]; Gao et al., 2020[7]). Mobile and computer use on weekdays and weekends has been associated with higher depression risk of adults, even when adjusting for lifestyle factors (Zhang et al., 2022[8]), and excessive computer use can independently predict depression and anxiety (de Wit et al., 2011[9]).
This policy brief contributes to this growing evidence base by analyzing newly collected, cross-country data from 14 countries in 2025 (ranging from OECD members to large emerging economies) as part of the OECD Digital Well-being Hub project in collaboration with Cisco. The dataset uniquely captures screen time and the use of digital technology, alongside subjective well-being indicators: the WHO-5 Well-Being Index, the OECD measures of life satisfaction and eudaimonia.
This analysis also considers loneliness, a psychosocial factor increasingly acknowledged as an important area of well-being (OECD, 2025[10]), and how it may interact with digital exposure to influence subjective well-being. By exploring the role of perceived social disconnection in shaping the effects of screen use, the brief aims to deepen understanding of digital engagement. This approach offers policy-relevant insights into how digital routines and psychosocial factors together affect well-being across diverse populations.
In collaboration with Cisco, the OECD WISE Centre conducted a poll to collect insights on individuals’ experiences with digital technology, as featured in the Digital Well-being Hub. The poll was carried out in collaboration with Cisco, however, the poll is also available as a crowd-sourcing tool which has allowed the collection of data in real time. The survey was conducted from 17 to 24 February and from 10 to 17 March 2025, yielding statistically valid responses from 14 611 individuals across 14 countries. The country sample was selected to reflect a broad range of socio-economic and cultural contexts, including both OECD and non-OECD members, including Americas: Canada, United States; East Asia and Oceania: Australia, Japan, Korea; Africa: South Africa; Europe: France, Germany, Italy, the Netherlands, United Kingdom; Latin America: Brazil, Mexico; South Asia: India. Each country is represented by a sample of just over 1 000 respondents, except for India, which includes 1 500 respondents. All participants completed the full 20-question poll as well as the accompanying demographic and profile questions of the Digital Well-being Poll.
In the sample analysed, most individuals report moderate to high levels of subjective or mental well-being, while a sizeable segment continues to experience low well-being. The Poll uses three subjective well-being indicators: the WHO-5 Well-Being Index, the OECD life satisfaction measure, and the OECD eudaimonia measure, which assesses the perceived meaningfulness of life (see Figure 1 and Box 2 for OECD Guidelines on Measuring Subjective Well-being (2013[11])):
The WHO-5 Well-Being Index distribution peaks between 70 and 80, covering the full 0-100 range. This indicates generally positive mental well-being while also capturing a subset of individuals potentially at risk of poorer mental health outcomes.
Life satisfaction scores, measured on a 0-10 scale, are left-skewed and peak between 7 and 8, suggesting that while most people are relatively satisfied with their lives, a substantial number report medium to low satisfaction. These findings are broadly consistent with the OECD How’s Life? data, which reports an average score of 7.4 across 35 countries.
Eudaimonia shows a similar left-skewed distribution, peaking between 7 and 8, and implying that most people find their lives worthwhile, though again with a significant portion reporting diminished meaning and purpose.
Proportion of respondents
Note: The WHO-5 Well-being Index is a brief, self-reported questionnaire developed by the World Health Organization. It includes five positively worded statements reflecting mood and daily functioning over the past two weeks, each rated on a scale from 0 (at no time) to 5 (all of the time), with the total score converted to a scale from 0 to 100. Subjective well-being measures of life satisfaction and eudaimonia are derived from responses on a 0-10 scale to the questions “How satisfied are you with your life as a whole these days?” and “Overall, to what extent do you feel the things you do in your life are worthwhile?” respectively.
Source: OECD poll on digital well-being, https://www.oecd.org/en/blogs/2024/11/oecd-poll-on-digital-well-being.html.
Measuring well-being is essential for assessing societal progress, and there is growing recognition that subjective well-being is a key component alongside social and economic indicators. To support this, the OECD developed comprehensive Guidelines on Measuring Subjective Well-being as part of its Better Life Initiative, launched in 2011 to track progress across eleven domains of well-being. They represent the first international effort to standardize the collection, analysis, and reporting of subjective well-being data, including life evaluations, emotional experiences, and “eudaimonic” measures of psychological well-being; highlighting the importance of such data for policymaking and highlight the critical role of national statistical agencies. Additionally, the Guidelines on Measuring Subjective Well-being provide practical recommendations and prototype survey modules for use by national and international organisations.
Source: OECD (2013[11]), OECD Guidelines on Measuring Subjective Well-being, OECD Publishing, Paris, https://doi.org/10.1787/9789264191655-en.
While the majority of people report moderate to relatively high levels of subjective well-being, notable variations are observed among different population groups, such as:
Older adults (56 years and above) report the highest levels of well-being across all well-being indicators. For the WHO-5 index, 23% of older adults score in the 70-80 range, a larger share than any other age group. Life satisfaction and eudaimonia also show strong concentrations in the upper end of the 7-9 score range among older adults. Middle-aged adults (36-55 years old) are more represented in the lower ranges of life satisfaction (0-4).
Women report lower levels of subjective well-being across all three measures of well-being. They appear more frequently in the lower score ranges for both the WHO-5 index and life satisfaction, and to a lesser extent, eudaimonia. A higher proportion of women fall into the 20-40 range of the WHO-5 index compared to men, and more women report life satisfaction scores below 4.
Personal screen time patterns generally show an inverted U-shaped relationship (referred by Przybylski and Weinstein (2017[3]) to as the “Goldilocks effect”) with all three subjective well-being measures (Figure 2). Well-being peaks at low levels (typically 2 hours daily), and declines with longer screen time, while this does not mean causality. The WHO-5 index peaks with up to 1 hour of screen time, though variation is relatively high, and falls to a low of 51 among the highest usage group. Life satisfaction follows a similar curve, with the highest levels reported at 1-3 hours daily and steeper declines beyond the 5-hour threshold. Individuals with no screen time report the lowest levels of satisfaction, however, with wider variation and representing merely about 1% of the poll sample. Findings for eudaimonia follow this pattern as well, peaking at 1-2 hours and declining thereafter.
The OECD Digital Well-being Poll includes a simple measure of professional screen time data, but its analysis is possibly more complex (Box 3). Unlike personal use, professional screen time may reflect job demands rather than individual’s pure preferences. Professional screen time also, varies by occupation, and excludes non-working groups (which represent a considerable share of the sample in the Poll). Without accounting for job type or autonomy, results are hard to interpret. These limitations make professional screen time a less relevant indicator, while personal screen time offers a clearer measure of digital engagement.
The average scores of the well-being measure by screen time category for personal use
Note: Personal screen time refers to the average amount of time spent using screens for personal use per day in the past week, including on personal computers, laptops, tablets, mobile phones, televisions and gaming consoles. The vertical bars show standard errors.
Source: OECD poll on digital well-being, https://www.oecd.org/en/blogs/2024/11/oecd-poll-on-digital-well-being.html.
As digital technologies increasingly permeate both work and personal life, distinguishing between professional and personal screen time is essential when assessing the impact of digital engagement on subjective well-being. Professional screen time, that is, digital screen use for work-related purposes, often reflects occupational demands and structural conditions (e.g. office-based and manual labour roles). In contrast, personal screen time involves discretionary use, such as streaming, gaming, and social media. This behavioural distinction is critical. While personal screen time reflects individual choices and leisure activities, professional screen time may well be compulsory and structured along professional norms.
Descriptive statistics indicate that professional screen time to some extent shows an inverted U-shaped relationship with subjective well-being, similar to personal screen time, but with subtler gradients. For example, those working more than 5 hours daily on screens report notably lower WHO-5 scores, resembling the trends observed only in high personal screen time categories beyond 3 hours per day (Figure 3). Life satisfaction and eudaimonia indicators show similar patterns, but with smaller margins. These findings indicate that excessive professional digital exposure, like its personal counterpart, may correlate with lower subjective well-being.
However, several caveats should be considered when interpreting professional screen time effects. First, the observed results may conflate occupational categories as these data do not control for job type, sector, or working autonomy levels. High screen time may characterize both highly paid IT professionals and screen-reliant administrative staff. Second, professional screen time cannot be uniformly applied across the population as retired, unemployed, or informal workers fall outside this category entirely. Given these limitations, personal screen time remains a more valid and interpretable indicator of digital engagement’s influence on subjective well-being.
The average scores of the well-being measure by screen time category for professional use
Note: Professional screen time refers to the average amount of time spent using screens for professional use per day in the past week, including on personal computers, laptops, tablets, mobile phones, televisions and gaming consoles. The vertical bars show standard errors.
Source: OECD poll on digital well-being, https://www.oecd.org/en/blogs/2024/11/oecd-poll-on-digital-well-being.html.
3.1 Predicting the odds of screen time harming subjective well-being
Building on the recent literature, the Poll data are analyzed using a binary logistic regression model to identify which population groups are more likely to experience low well-being outcomes associated with personal screen use. This approach, commonly used in clinical and preventive health research, focuses on the probability of individuals falling below critical well-being thresholds rather than examining average levels across the sample. The analysis concentrates on three dichotomous measures of well-being (i.e. treated as yes-or-no outcomes):
Poor mental well-being, defined as WHO-5 index below 50 consistent with the threshold proposed by Topp et al. (2015[12]) and used in several other empirical studies (e.g. Madhav et al. (2017[2]); Santos et al. (2024[1])).
Deprivation in life satisfaction, defined as a score of 4 or below, consistent with the threshold used in the OECD How’s Life? report (2024[13]).
Deprivation in eudaimonia, defined as a score of 4 or below, consistent with the threshold used in the OECD How’s Life? report (2024[13]).
The central question is whether the amount of time that people spend online for leisure (i.e. excluding school or work-related use) affects their likelihood of falling into one of these low well-being categories. To reflect real-life usage patterns and align with other studies (e.g. Grøntved et al. (2015[14]); Twenge and Campbell (2018[4])), personal screen time is grouped into the following categories: under 1 hour, 1-3 hours, 3-5 hours, and over 5 hours per day. Based on observed distributional patterns in the poll data, the categories of “Up to 1 hour” and “None” were merged, as were “1-2 hours” and “2-3 hours”, while “3-5 hours” and “more than 5 hours” were retained as distinct categories.
Two binary logistic regression models are estimated for each subjective well-being measure: (1) a baseline model excluding the loneliness variable, and (2) an extended model including lonelines1 and its interaction terms with screen time as regressors. The base model focuses on the total effect of screen time on well-being, which does not explicitly consider loneliness. The second, extended model explicitly considers loneliness, assuming that the link between well-being and screen time is influenced also by loneliness among other factors; and examines how the relationship between screen time and well-being changes in relation to loneliness. This addition can help inform on potentially moderating effect of loneliness on well-being, however, the models do not determine any causal relationship between the well-being and screen time variables. Both models control for the same demographic characteristics (age, sex, education, and country), sleep duration, physical activity, and financial difficulty, and include interaction terms between screen time and employment status. The results from the binary logit models are not sensitive to alternative empirical specifications, found to be consistent the ordinary least squares and quantile regression analyses that were performed with the same data and specifications.
The analysis contributes to the literature in three distinct ways. First, it covers 14 OECD and non-OECD countries worldwide using representative data samples from the Digital Well-being Hub’s poll, ensuring comparability of data across countries. Second, it employs standardized measures of subjective well-being, either developed or endorsed by the OECD. Third, the data collection conducted in March 2025 offers up-to-date insights into individuals' digital experiences and well-being.
3.2 What the data reveal
The empirical analysis highlights the association between personal screen time and subjective well-being, assessed across three measures: poor mental well-being (measured by the WHO-5 Index), low life satisfaction, and low eudaimonia (i.e. sense of life purpose). While common demographic factors such as sleep duration, physical activity, and financial difficulty remain important, the findings show that higher levels of personal screen time are consistently associated with poorer well-being outcomes.
Individuals who spend more than five hours per day on personal screen use are significantly more likely to report poor well-being outcomes compared to those with moderate use (1-3 hours per day). Figure 4 presents the odds ratios from the logistic regression models. In the base model, which captures the total effect of screen time, they show 44% higher odds of low mental well-being (WHO-5), 47% higher odds of low life satisfaction, and 62% higher odds of low eudaimonia. In the extended model that explicitly includes loneliness and its interaction terms, the increased odds are 28%, 32%, and 47% respectively, suggesting that loneliness may moderate the relationship between screen time and well-being.
In the base model, which captures the total effect of screen time, the relationship is non-linear: individuals with low personal screen use show worse outcomes, with 45% higher odds of low life satisfaction and 41% higher odds of low eudaimonia. This pattern aligns with previous research (e.g. Przybylski and Weinstein (2017[3])), suggesting both digital underuse and overuse may be linked to reduced well-being, depending on the context. More favorable well-being outcomes are observed among individuals with moderate personal screen use (1-3 hours daily), suggesting a possible link between balanced digital engagement and well-being. Those using screens for 3-5 hours daily still show elevated risks (i.e. 22% higher odds of low mental well-being (WHO-5) and 32% for low eudaimonia in the base model), but lower than the risks associated with more than 5 hours of daily use.
Note: In this graph, values further to the right on the x-axis indicate a stronger association with poor or deprived well-being status. A value of 1 represents parity with the reference group (screen time: 1-3 hours). The horizontal lines extending from each estimated point represent 95% confidence intervals and the colour of the estimates indicates their level of statistical significance. Two models are estimated: (1) a base model and (2) an extended model explicitly including loneliness and its interaction terms with screen time. These models include the same set of control variables and are displayed side by side in the graph using markers of different shapes.
Source: Estimation based on OECD poll on digital well-being, https://www.oecd.org/en/blogs/2024/11/oecd-poll-on-digital-well-being.html.
The regression models also incorporate interaction terms (screen time with employment status, and screen time with loneliness), demographic factors and loneliness as explanatory variables. Figure 5 presents the odds ratios of selected regressors which have statistically significant effects in at least one model. The results suggest that both loneliness and employment status heighten the risks associated with high screen time, while loneliness on its own substantially undermines subjective well-being. Consistent with findings from Chen, Sun and Zhuang (2022[15]) and Santos et al. (2024[1]), sleep deprivation, low levels of physical activity, and financial stress appear as significant risk factors for poor mental well-being and life satisfaction. Age, gender and education attainment also show generally significant associations with well-being outcomes. The important findings include:
Loneliness and its interaction with high screen use (included in extended model only): Individuals who report feeling lonely have 2.2, 5.5 and 5.3 times higher odds of experiencing poor mental well-being (WHO-5), low life satisfaction, and low eudaimonia, respectively, compared to those who are not lonely. Among those with more than five hours of screen time, loneliness further raises the odds of poor mental well-being.
Interaction between employment status and high screen time: Among unemployed individuals, high screen time is associated with a 1.8 times higher likelihood of poor mental well-being in both the baseline and extended models, and a slight increase in the odds of low life satisfaction. For students, high screen time is linked to a marginally significant 1.5 times increase in the odds of poor mental well-being.
Sleep duration: Sleeping less than 4 hours raises the odds of low well-being by 1.5-2.0 times across all three well-being measures, even after accounting for loneliness.
Physical activity: Those with at least 8 walking sessions per week have 21-22% lower odds of poor well-being in both models.
Financial difficulty: Struggling to make ends meet raises odds of low well-being by 1.4-2.3 times, with some mediation via loneliness.
Education: People with upper-secondary education or less face 24-49% higher odds of poor well-being.
Age: Older adults (66+) have 28-34% lower odds of poor well-being, slightly reduced when loneliness is included.
Note: In this graph, values further to the right on the x-axis indicate a stronger association with poor or deprived well-being status. A value of 1 represents parity with the reference group (screen time: 1-3 hours). The horizontal lines extending from each estimated point represent 95% confidence intervals and the colour of the estimates indicates their level of statistical significance. Two models are estimated: (1) a base model without the loneliness variable, and (2) a model that includes loneliness as a regressor. These models include the same set of control variables and are displayed side by side in the graph using markers of different shapes. Since the loneliness variable is included only in the second model, the interaction term with loneliness is represented by a single estimate for each measure of subjective well-being.
Source: Estimation based on OECD poll on digital well-being, https://www.oecd.org/en/blogs/2024/11/oecd-poll-on-digital-well-being.html.
Understanding the quality of digital engagement is key to explain why similar screen time levels yield different well-being outcomes. Distinguishing further between passive and active use across social platforms and purposes (such as entertainment, networking, or information seeking) can reveal critical differences in how digital experiences affect users’ subjective well-being.
Time series data are critical to understanding whether digital experiences produce lasting effects or temporary disruptions in well-being; as well as to conclude on the causal relationship between screen time and subjective well-being. Repeated gathering of poll data would allow differentiation between short-term fluctuations and longer-term trends, offering insights into specific digital behaviours over time and across changing digital landscapes.
Future research could further explore the questions above, allowing for a deeper understanding of the causal cognitive and emotional mechanisms behind the relationship between digital engagement and poor subjective well-being, as well as identifying protective and mitigating factors that policies and digital technologies may seek to embed and promote.
References
[15] Chen, Z., J. Sun and W. Zhuang (2022), “Combination of physical activity and screen time on life satisfaction in adults: A cross-sectional survey”, Frontiers in Psychology, Vol. 13, https://doi.org/10.3389/fpsyg.2022.962520.
[9] de Wit, L. et al. (2011), “Are sedentary television watching and computer use behaviors associated with anxiety and depressive disorders?”, Psychiatry Research, Vol. 186/2-3, pp. 239-243, https://doi.org/10.1016/j.psychres.2010.07.003.
[7] Gao, J. et al. (2020), “Mental health problems and social media exposure during COVID-19 outbreak”, PLoS ONE, Vol. 15/4, https://doi.org/10.1371/journal.pone.0231924.
[14] Grøntved, A. et al. (2015), “A prospective study of screen time in adolescence and depression symptoms in young adulthood”, Preventive Medicine, Vol. 81, pp. 108-113, https://doi.org/10.1016/J.YPMED.2015.08.009.
[5] Lee, J. and Z. Zarnic (2024), “The impact of digital technologies on well-being: Main insights from the literature”, OECD Papers on Well-being and Inequalities, No. 29, OECD Publishing, https://doi.org/10.1787/cb173652-en.
[2] Madhav, K., S. Sherchand and S. Sherchan (2017), “Association between screen time and depression among US adults”, Preventive Medicine Reports, Vol. 8, pp. 67-71, https://doi.org/10.1016/J.PMEDR.2017.08.005.
[10] OECD (2025), Social Connections and Loneliness in OECD Countries, OECD Publishing, Paris, https://doi.org/10.1787/6df2d6a0-en.
[13] OECD (2024), How’s Life? 2024: Well-being and Resilience in Times of Crisis, OECD Publishing, https://doi.org/10.1787/90ba854a-en.
[11] OECD (2013), OECD Guidelines on Measuring Subjective Well-being, OECD Publishing, https://doi.org/10.1787/9789264191655-en.
[3] Przybylski, A. and N. Weinstein (2017), “A Large-Scale Test of the Goldilocks Hypothesis: Quantifying the Relations Between Digital-Screen Use and the Mental Well-Being of Adolescents”, Psychological Science, Vol. 28/2, pp. 204-215, https://doi.org/10.1177/0956797616678438.
[6] Riehm, K. et al. (2019), “Associations Between Time Spent Using Social Media and Internalizing and Externalizing Problems Among US Youth”, JAMA Psychiatry, Vol. 76/12, pp. 1266-73.
[1] Santos, R. et al. (2024), “The Associations Between Screen Time and Mental Health in Adults: A Systematic Review”, Journal of Technology in Behavioral Science, https://doi.org/10.1007/s41347-024-00398-7.
[12] Topp, C. et al. (2015), “The WHO-5 well-being index: A systematic review of the literature”, Psychotherapy and Psychosomatics, Vol. 84/3, pp. 167-176, https://doi.org/10.1159/000376585.
[4] Twenge, J. and W. Campbell (2018), “Associations between screen time and lower psychological well-being among children and adolescents: Evidence from a population-based study”, Preventive Medicine Reports, Vol. 12, pp. 271-283, https://doi.org/10.1016/J.PMEDR.2018.10.003.
[8] Zhang, Y. et al. (2022), “The relationships between screen time and mental health problems among Chinese adults”, Journal of Psychiatric Research, Vol. 146, pp. 279-285, https://doi.org/10.1016/J.JPSYCHIRES.2021.11.017.
The OECD Digital Well-being Hub with Cisco: https://www.oecd.org/en/data/tools/digital-well-being-hub.html
How’s Your Digital Well-being? https://oecdstatistics.blog/2025/02/26/hows-your-digital-well-being/
Contacts
For more information contact us: wellbeing@oecd.org
OECD Centre on Well-being, Inclusion, Sustainability and Equal Opportunity (WISE)
www.oecd.org/wise
← 1. Including loneliness enables a more nuanced interpretation of how screen time relates to well-being. Loneliness is commonly linked to lower well-being (OECD, 2025[10]), however, its association with screen time remains ambiguous and directionally unclear. While benefits from greater social connectedness may help to alleviate loneliness, digital engagement can worsen social isolation (Przybylski and Weinstein, 2017[3]).