Social media reduction or abstinence interventions are providing mental health benefits – reanalysis of a published meta-analysis

A recent meta-analysis published in this journal included 27 studies that experimentally manipulated social media use and investigated their impact on mental health outcomes (Ferguson, 2024). The author concluded that social media effects were statistically no different from zero. However, this meta-analysis did not investigate potential moderating effects of length of social media reduction or abstinence interventions. We conducted a re-analysis to investigate the impact of social media reduction / abstinence intervention length on mental health outcomes. We used the information available on the OSF platform related to the original meta-analysis and excluded 7 studies because they were not reduction / abstinence interventions. We categorized studies into those with intervention lengths of less than 1 week vs. 1 week or longer, and tested for curvilinear relationships between intervention length (weeks and days as continuous variables) and outcomes by including quadratic terms. Stratified analyses indicated that interventions of less than 1 week resulted in significantly worse mental health outcomes (d=-0.168, SE=0.058, p=.004), while interventions of 1 week or longer resulted in significant improvements (d=0.169, SE=0.065, p=.01). Analyses of intervention length as continuous moderator included k=19 studies and showed significant quadratic effects for number of weeks (b=-0.022; SE=0.008; z=-2.6; p<.01), as well as number of days (b=-0.001; SE=0.000; z=-2.5; p<.05). These findings suggest that social media use reduction / abstinence interventions should have a minimum length of 1 week or longer to confer mental health benefits. An ideal intervention length may be around 3 weeks, but future research is needed to confirm this.