Does Social Media Impact Adolescent Mental Health?

Causal Effects of Exposure to Social Media on Adolescent Mental Health

The mental health of adolescents in the United States has seen a steep decline since 2011, roughly coinciding with the increasing popularity of social media and smartphones. But does social media have a causal impact on the mental health of adolescents or are concerns about the effect of social media on kids a form of public hysteria? In this study, the investigators will conduct the first field experiment in 11-14-year-olds to examine whether, how, and for whom social media harms mental health.

Study Overview

• Status: Recruiting
• Conditions: Mental Health
• Intervention / Treatment: Behavioral: Restricted Social Media

• Study Type: Interventional
• Enrollment (Estimated): 500
• Phase: Not Applicable

Contacts and Locations

• Study Contact: Name: Kostadin Kushlev, PhD; Phone Number: 804-585-4385; Email: kk1199@georgetown.edu

Study Locations

• United States
  • District of Columbia
    • Washington, District of Columbia, United States, 20057
      • Recruiting
      • Georgetown University
      • Contact: Kostadin Kushlev, PhD; Email: kk1199@georgetown.edu

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child
• Accepts Healthy Volunteers: Yes

Description

Inclusion Criteria:

1. is an adolescent who is between the ages of 10 and 14;
2. is an adolescent whose parents have decided to buy them their first smartphone;
3. is an adolescent both of whose parents/guardians consent for them to be in the study;
4. is an adolescent who assents to be in the study; 5) is an adolescent who speaks and reads English sufficiently to complete surveys and provide informed consent.

Exclusion Criteria:

1. is an adolescent who has severe developmental problems (e.g., autism, severe language delay);
2. is an adolescent who is currently or has ever been diagnosed with severe or moderately severe mental illness;
3. is an adolescent who is currently or has ever experienced suicidal ideation.

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Basic Science
• Allocation: Randomized
• Interventional Model: Crossover Assignment
• Masking: None (Open Label)

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
No Intervention: Naturalistic Social Media; In this control condition, participants will have no study-imposed restrictions on social media use.
Experimental: Restricted Social Media; Participants will have social media apps on their phones blocked for three months.	Behavioral: Restricted Social Media; Participants (N = 500) will be randomly assigned to either have no study-imposed restrictions on social media use (naturalistic social media condition) or have no access to social media apps on their phones (restricted social media condition). This manipulation will last three months, after which both groups will have no study-imposed restrictions on social media for three more months.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Revised Child Anxiety and Depression Scale (RCADS-25): Total Score	The average score varies between 0 and 4, with higher scores indicating a worse outcome.	Baseline, three months, and six months

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Revised Child Anxiety and Depression Scale (RCADS-25): Depression Subscale	The average score varies between 0 and 4, with higher scores indicating a worse outcome.	Baseline, three months, and six months
Revised Child Anxiety and Depression Scale (RCADS-25): Anxiety Subscale	The average score varies between 0 and 4, with higher scores indicating a worse outcome.	Baseline, three months, and six months

Other Outcome Measures

Outcome Measure	Measure Description	Time Frame
DSM-5 Self-Rated Level 1 Cross-Cutting Symptom Measure Youth Self Report	The average score varies between 0 and 4, with higher scores indicating a worse outcome.	Baseline, three months, and six months
DSM-5 Self-Rated Level 1 Cross-Cutting Symptom Measure Parent Report	The average score varies between 0 and 4, with higher scores indicating a worse outcome.	Baseline, three months, and six months
Upward Social Comparison (EMA)	Mediator measured with a single, face-valid item in EMA surveys: "Over the past 24 hours, how often have you compared yourself to someone who you believe is better off than you?"	Every three days throughout the study with two averages computed for 0 to 3 months and 3 to 6 months.
Offline Social Capital: Quantity of In-Person Social Interactions (EMA)	Mediator measured with a single, face-valid item in EMA surveys: "Over past 24 hours, how much time did you spend socializing in person?"	Every three days throughout the study with two averages computed for 0 to 3 months and 3 to 6 months.
Offline Social Capital: Quality of In-Person Social Interactions (EMA)	Mediator measured with a single, face-valid item in EMA surveys: "Over past 24 hours, how much did you enjoy your in-person social interactions with others?"	Every three days throughout the study with two averages computed for 0 to 3 months and 3 to 6 months.
Quantity of Online Social Interactions (Mobile Sensing)	Mediator: Quantity of digital social interactions based on the total number of words participants type on their phone keyboards each day.	Daily throughout the study with two averages computed for 0 to 3 months and 3 to 6 months.
Valence of Online Social Interactions (Mobile Sensing)	Mediator: Quality/valence of digital social interactions by measuring the balance between positive and negative words participants type on their phone keyboards. The valence of the words typed will first be scored using the Linguistic Inquiry and Word Count (LIWC) software. The sum of the valence scores of all words will produce the final score for analysis.	Daily throughout the study with two averages computed for 0 to 3 months and 3 to 6 months.
Active (vs Passive) Social Media Use (Mobile Sensing)	Active usage includes creating content and direct messaging, while passive usage includes browsing, scrolling, or watching videos. Using data from the Inertial Measurement Unit (i.e., accelerometer and gyroscope data) and keyboard input (i.e. typing or not), an Extremely Randomized Trees algorithm will classify behaviors into active or passive.	Daily throughout the study with two averages computed for 0 to 3 months and 3 to 6 months.
Passive (vs Active) Social Media Use (Mobile Sensing)	Active usage includes creating content and direct messaging, while passive usage includes browsing, scrolling, or watching videos. Using data from the Inertial Measurement Unit (i.e., accelerometer and gyroscope data) and keyboard input (i.e. typing or not), an Extremely Randomized Trees algorithm will classify behaviors into active or passive.	Daily throughout the study with two averages computed for 0 to 3 months and 3 to 6 months.
Time Spent in Specific Social Media Apps (Mobile Sensing)	Time participants spend in specific social media apps as assessed unobtrusively by mobile sensing.	Daily throughout the study with two averages computed for 0 to 3 months and 3 to 6 months.

Collaborators and Investigators

• Sponsor: Georgetown University
• Collaborators: National Institutes of Health (NIH)

Study record dates

Study Major Dates

• Study Start (Actual): September 29, 2024
• Primary Completion (Estimated): November 1, 2027
• Study Completion (Estimated): February 1, 2028

Study Registration Dates

• First Submitted: August 23, 2023
• First Submitted That Met QC Criteria: September 16, 2023
• First Posted (Actual): September 22, 2023

Study Record Updates

• Last Update Posted (Actual): June 17, 2025
• Last Update Submitted That Met QC Criteria: June 13, 2025
• Last Verified: June 1, 2025

More Information

Terms related to this study

• Keywords: depression; mental health; anxiety; social media
• Other Study ID Numbers: 1R01MH135467 (U.S. NIH Grant/Contract)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: YES
• IPD Plan Description: Demographic, clinical (self-reports and parent reports), neurocognitive (gradCPT), and passive mobile sensing data (accelerometer, GPS, keyboard input, app usage) will be collected and made available on NIMH Data Archive (NDA). All data will be de-identified prior to deposition to NDA. To request access to the data, researchers will use the standard processes at NDA, and the NDA Data Access Committee will decide which requests to grant. All code for the analyses-including statistical analysis, data transformations, and data cleaning-will be shared on the Open Science Framework (OSF). The code can be found by searching for "Kostadin Kushlev" on the Open Science Framework.
• IPD Sharing Time Frame: All data will be deposited to NDA starting 12 months after the award begins and will be deposited every six months thereafter following the usual NDA data submission dates.
• IPD Sharing Access Criteria: To request access to the data, researchers will use the standard processes at NDA, and the NDA Data Access Committee will decide which requests to grant. The standard NDA data access process allows access for one year and is renewable. The protocol and code will be available on OSF to anybody with access to the internet. De-identified data associated with each publication will also be made available on OSF by the publication date of each paper.
• IPD Sharing Supporting Information Type: STUDY_PROTOCOL; SAP; ICF; ANALYTIC_CODE

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No