The association appeared repeatedly among adults from 2014 to 2024, but the findings do not show that social media use causes depression.
Researchers analyzed repeated cross-sectional surveys of U.S. adults aged 18 to 70 collected each year from 2014 through 2024. They examined whether social media use was associated with a yes-or-no measure of self-reported depression and whether it helped classify depression alongside demographic information and other media-use measures.
The models showed a repeated association between higher social media use and depression classification. Social media use had a neutral contribution to the models at lower levels, then shifted to a positive contribution at approximately 150 minutes per day.
What the surveys showed
The study combined 11 surveys with a total of 183,000 adult participants. For each year, the researchers trained a balanced random-forest model to classify self-reported depression. The models had accuracies ranging from 0.71 to 0.75 and area-under-the-curve scores ranging from 0.65 to 0.73.
Analysis of how individual factors contributed to the models showed that social media use shifted from a neutral contribution to a positive one at about 150 minutes per day. This pattern appeared repeatedly across the 11 datasets, indicating a stable association between higher social media use and self-reported depression classification.
Why the pattern matters
The repeated pattern across surveys and years adds evidence that the relationship between social media use and depression is not limited to a single dataset or age group. The approximate 150-minute level gives researchers a specific exposure range to examine in future studies.
However, the results do not establish whether social media use contributes to depression, whether depression leads people to use more social media, or whether other factors influence both. The researchers say longitudinal and experimental work is needed to determine whether high exposure levels represent a clinically meaningful risk.
Evidence and caveats
This was an analysis of 11 repeated cross-sectional surveys, not a study that followed the same people over time or randomly assigned their social media use. The depression measure was self-reported and binary, meaning it recorded whether participants reported depression rather than providing a clinical diagnosis or measuring symptom severity.
The models classified depression using social media use, demographics and other media-use variables, so their performance does not by itself demonstrate a causal relationship. The reported threshold is also a model-based shift in social media use’s contribution to depression classification, not proof of a specific cutoff at which depression begins.
