Overview

Background

Social media use among youth populations is widely discussed in both research literature and contemporary news media, particularly with reference to high rates of usage: 97% of youth in the US use social media daily and 46% use it “constantly” (Pew Research Center, 2023). Additionally, internalizing symptoms (e.g., depression, anxiety) may be increasing in youth populations, prompting an obvious question: Is social media use negatively affecting adolescent mental health?

The answer to this question is more complicated than it appears at first glance. Firstly, even if social media use is negatively impacting youth mental health, it is still unclear why. Could it be the amount young people use social media? Could it be the way they use it (e.g., “addictive use,” compulsive use)? Could it be the content that they engage with? There is also the question of causality. Does social media use predict increased internalizing symptoms, does increased internalizing symptoms predict social media use, or is it some combination of both?

Furthermore, while it is easy to ponder the consequences of social media, what can be said about its potential benefits? Many adolescents turn to social media for social connection and emotional support (TaeHyuk Keum et al., 2023). As we consider these questions, it becomes evident that there is still so much to learn about the relationship between social media and mental health. In our study, we aim to gain a deeper understanding of the potential connection between social media and internalizing symptoms by examining the bi-directional relationship between these factors.

Study Overview

This study involves a longitudinal cohort of 1,000 adolescents. To qualify, participants must pass a consent quiz (which tests their comprehension of the study requirements), live in the USA, be aged 13-20 years old, use social media, and not be actively suicidal. 

Over the course of the 6 month study period, participants are asked to do three main tasks:

  1. Complete three longer surveys: one at baseline, and then again at the 3- and 6-month mark. These three surveys capture demographic and mental health information.
  2. Complete ecological momentary assessments (EMAs) twice daily. 
  3. Regularly use a special keyboard that tracks their typing.

The EMAs and special keyboard are made available to participants via a mobile app called Effortless Assessment Research System (EARS). Following their completion of the baseline survey, and subsequent publication for the study, participants complete a welcome call with a research assistant during which they download EARS. For the next six months, they use EARS to complete their daily EMAs and upload data by using the EARS keyboard. 

While the three long surveys capture a big picture understanding of how a participant's mental health fluctuates across the study period, the EMAs afford us a day-to-day look into their general mental health symptoms. Then, using natural language processing (NLP), we can analyze the text data from the EARS keyboard to identify patterns in language generation (e.g., use of cognitively distorted (CD) language). Together, we can use this information to understand how certain language patterns may relate to mental health symptoms. This will also allow us to address Aim 1 of our study—modeling the bidirectional association between CD language and internalizing symptoms.

When participant complete the three longer surveys, they also have the option to provide their social media handles to s (i.e., for Instagram, TikTok, and/or Twitter/X). Using a data scraper, we will be able to analyze the prevelance of CD language on each profile, as well as basic engagement metrics (e.g., number of likes, comments, and shares). This will allow us to address Aim 2 of our study—establishing the contagion of CD-content vs. non-CD content on social media. With the results of Aim 1 and 2, we will be able to address Aim 3—determine the bidirectional association between cognitively distorted (CD) language and internalizing symptoms.