m m statistics project

m m statistics project

The Impact of Social Media Usage on Mental Health: A Statistical Analysis

1. Introduction

It is widely acknowledged that the existence of social ties has positive effects on well-being. Likewise, social networking service (SNS) usage, by representing social interaction, could also have positive effects. However, a set of research suggests negative effects of SNS usage on well-being. The subjects of the research frequently have SNS addiction and the contents of newsfeeds are reported as addictive across three genres. They frequently have the feature of self-disclosure and mutual social comparison. Particularly, by enabling its users to expose themselves to a large amount of information on other people, the implication is to expose oneself to many potential victims for social comparison, bringing the users a high level of stress. Such an implication is also consistent with existing research, stating that the use of social media, in general, plays a significant role in issues of the perception of isolation.

Social media has long been criticized by the public for not only its merits but also its detrimental effects, including both informational and emotional effects. It is stipulated that heavy usage of social media may potentially lead to developing depressive symptoms, and on the contrary, depressive symptoms could increase one’s own social media usage. We identified two particular aspects of social media usage that can lead to the development of depression. We determined depressive states over time and took into account potential confounding factors that have been neglected by earlier studies. We also examined the reverse causality of the relationship between social media usage and depressive symptoms for both males and females using a large-scale longitudinal dataset in Japan.

2. Literature Review

Theoretical distinction about whether SNS play a causal role in individual well-being, or whether people with different pre-existing well-being use SNS for different motives, has an applied value.

On the one hand, weak theory supports determinant claims. We ask SNS about motivation and experience and show that motivation is associated with any differences between well-being of SNS users and non-users. Individual differences, especially diversity in motivation, predict between-person differences in well-being associated with use, whereas time-series research finds a range of within-person effects suggesting that widespread social observation unlinked to motivational considerations represents a threat to subjective well-being.

Andrew K. Przybylski and Amy Orben have partnered with the UK Household Longitudinal Study to conduct the first large-scale representative longitudinal study of adolescents to compare with the often-cited cross-sectional studies linking social media use with epidemics of poor mental health. The findings reject a deterministic, chronic use, and universal harm perspective and argue that social media is just one element of a broader social context that touches on multiple daily experiences related to adolescent well-being.

3. Methodology

The main purpose of the initial analysis is to determine whether spending more time on social media leads individuals to feel that they are more easily overwhelmed, especially given the rich amount of activities social media enables users to do. The intent of the next analysis is to elevate the survey responses into a more broad mental health outcome — collectively feeling stressed out, overwhelmed, anxious, or nervous. This accounts for the possibility that frequent social media usage is uniquely negatively impacting multiple dimensions. The fourth and fifth analyses then assess impacts on other often coexisting mental health symptoms.

We conduct an analysis of the impact of social media usage on reported mental health outcomes in a population-weighted sample of respondents from the 2017 National Health Interview Survey via a series of ordered logit models and hierarchical linear models. This survey provides detailed data on usage of social media and four reported mental health outcomes. We first examine whether frequent versus infrequent social media usage changes the ease with which an individual reports that life’s demands are overwhelming, a common symptom of high stress and incapacity to deal with problems characteristic of anxiety. We then change the outcome variable to collectively feeling stressed out, overwhelmed, anxious, or nervous, which is also asked in the survey. Our third and fourth analyses parallel the first two, except we change the outcome variable to represent being depressed and anxious.

4. Data Analysis and Results

This section presents the results of the empirical analysis of the relationship between social media usage and mental health. We conducted a range of different empirical models, where the dependent variable can be social media usage, internet usage, and control variables. We run a range of different regressions, including cross-sectional, panel, and meta-analysis. The last type of model incorporates estimates of between-study variance by modifying the weight of each study to include this information, to adjust for possible publication bias, pool the results of different single studies, and include all relevant data in the analysis. The random effect models cannot account for heterogeneity between web samples, so we use a fixed effects model, which interprets how the association might vary across different social media and mental health measures.

We also consider a range of different empirical models, where the dependent variable can be social media usage, internet usage, and control variables. The internet variables are country-level statistics on the percentage of the population that use the internet.

In this section, we use mixed effects models, panel regression models, and meta-analysis to examine the impact of social media on mental health. The dependent variable in each analysis is a measure of mental health, and the explanatory variable of interest is the social media usage score. The types of mental health where higher scores reflect better mental health include life satisfaction, happiness, subjective well-being, the positive affect scale, the flourishing score, positive mental health, other life satisfaction scales, and measures of depression and mental health. We include a country-level fixed effects variable to capture differences in the level of mental well-being specific to country culture or shared traits.

5. Conclusion and Recommendations

We find no general association between enhanced social media usage and worsened mental health in cross-sectional data. An enhanced subjective evaluation of own mental health is associated with greater connectedness to others on social media and talking to people they do not know via social media. Anxiety is associated with being connected only to friends and family, while depression is associated with contacting strangers and friends and family exclusively. We find patterns in survey data of when such mental health alterations happen. This analysis is by no means finished, and subgroups need to be analyzed for potentially differing associations between mental health and subjective well-being, and group-level analysis is left to explore. These findings are very important for policy interventions aimed at managing how South Africans make use of social media.

Social media technologies play a crucial role in the everyday lives of many people around the globe. We have enrolled in a term-long unit in third-year economics statistics aimed at exploring some consequential issues associated with social media usage and have delivered some students for self-paced exploration using basic statistical data analyses of observational data. We have been able to document key features of social media usage, associated mental health, as well as explore whether an association between these exists, and important determinants for these on survey data. Clear from our analysis is that the majority of our class make use of social media as part of their everyday lives. The most popular platform is Facebook, followed by WhatsApp, with 14% of the class indicating that they make use of no social media at all. Such engagement with social media platforms is important.