Temporal Associations of Screen Time and Anxiety Symptoms Among Adolescents

Abstract

There are indications that screen time is associated with symptoms of anxiety in adolescents, but from a longitudinal perspective, the magnitude, specificity, and temporal precedence in this association are relatively under-explored. To address this gap, over the course of 4 years, we annually surveyed adolescents assessing their usage of various types of screen time per day and symptoms of anxiety.

Methods

We analyzed data from the Co-Venture Trial, a longitudinal randomized trial assessing the efficacy of a personality-targeted drug and alcohol prevention program named Preventure. A total of 3,826 seventh graders (47% female, mean age = 12.8 years, SD = 0.4 years) were recruited and studied annually from 7th grade through 10th grade, with informed assent and parent consent under the approval of the Sainte-Justine’s Hospital Institutional Review Board. Screen time was measured by how much time per day one spends playing video games (on a computer, cell phone, game console), using social media (Facebook, Twitter, or other social networking sites), watching shows or movies (on television or the computer), and on the computer engaging in other activities. The amount of screen time was operationalized into four categories: 0 to 30 minutes, 30 minutes to 1 hour 30 minutes, 1 hour 30 minutes to 2 hour 30 minutes, and 3 hour 30 minutes or more. Anxiety symptoms were measured using the Brief Symptom Inventory–Anxiety subscale, ² asking participants to indicate, on a scale from 0 (not all) to 4 (very much) to what extent they experienced seven symptoms of anxiety during the past 12 months (e.g., feeling tense, experiencing spells of terror or panic, and being suddenly scared for no reason). Gender and socioeconomic status (SES) were also assessed. In total, 3,659 participants (of 3,826) had complete data for measures of interest and were included in the analyses.

Multilevel linear modeling in R was used to assess the relationship between screen time and anxiety. The model included random intercepts and slopes at the individual and school level. The time parameter was coded as each yearly assessment. Our model evaluated the predictive role of three time-varying predictors for each of the four types of screen time: average of each screen time over 4 years (between-person effects), changes in each screen time over 4 years compared to a person’s mean (within-person effects), and screen time use the year before compared with the person’s mean use (lagged effects).

Results

Analyses included exposure/nonexposure to the trial as covariate, as well as gender and SES, which did not significantly affect the associations between screen time and symptoms of anxiety (Table 1). The significant within-person effect for social media use indicated that a 1-hour increase in average time spent using social media in a given year was associated with a 0.21 unit (on a scale from 0 to 28) increase in the severity of anxiety symptoms (95% CI, 0.14 to 0.28) within that same year. Significant, but more modest within-person relationships were revealed for television viewing and computer use on anxiety: 0.15 unit increase (95% CI, 0.08 to 0.22) and 0.11 unit increase (95% CI, 0.04 to 0.18), respectively. We found no significant within-person association between anxiety and video gaming, and there was no evidence of lasting effects of any form of screen time on anxiety if increases in screen time were not sustained.

Table 1.

Estimated Parameters for a Multilevel Model Assessing the Association Between Screen Time and Anxiety Symptoms.

	Anxiety Symptoms
Predictors	Estimate	Std. Er	Pr(>|t|)	95% CI
Intercept	2.93	0.21	0.000	2.61 to 3.43
Time	0.07	0.02	0.002	0.03 to 0.12
Socioeconomic status	−0.03	0.02	0.206	−0.09 to 0.02
Sex	−1.53	0.12	0.000	−1.76 to −1.30
Intervention	−0.08	0.31	0.782	−0.70 to 0.52
Social media between-person	0.43	0.05	0.000	0.32 to 0.54
Social media within-person	0.21	0.03	0.000	0.14 to 0.28
Social media lagged	−0.11	0.06	0.052	−0.23 to 0.01
Video gaming between-person	−0.01	0.05	0.789	−0.12 to 0.09
Video gaming within-person	0.01	0.03	0.624	−0.04 to 0.07
Video gaming lagged	0.06	0.05	0.273	−0.05 to 0.16
Television between-person	−0.12	0.05	0.041	−0.24 to −0.01
Television within-person	0.15	0.03	0.001	0.08 to 0.22
Television lagged	−0.06	0.07	0.421	−0.19 to 0.08
Computer use between-person	0.42	0.07	0.000	0.28 to 0.57
Computer use within-person	0.11	0.03	0.002	0.04 to 0.18
Computer use lagged	−0.05	0.07	0.406	−0.18 to 0.07

Note. Significant associations are marked in bold font.

Discussion

Our results indicate that anxiety symptoms within a given year may become more severe when social media use, television viewing, and/or computer use increase within that same year. These are interesting findings considering another recent publication reporting within-person associations between depression, social media and television, but not computer use, among Canadian adolescents, which were explained by decreases in self-esteem. ³ Computer use appears to be uniquely associated to increases in anxiety, potentially in relation to using the computer for homework activities, but this needs further investigation. Finally, the findings suggest that the association between screen time and anxiety was concurrent and did not last if screen time reduced in intensity.

The rising rates of adolescent anxiety and depression in the Western World ⁴ have been attributed to the excessive amount of time that adolescents, nowadays, spend in front of digital screens. ⁵ The present study indicates that increases in time spent using social media, television, and computers are associated with increases in anxiety. These worrisome findings do not only ask for further investigation concerning potential explanatory models of this association but also stress the need for the Canadian scientific community and creators/distributors of digital products to join forces in order to better understand the potential role of digital technologies on mental health. Equally, parents and clinicians are requesting guidance on how to manage the impact of digital devices in child and adolescent mental health. Common measurement tools and more innovative research designs are needed, including ones that help us to better understand how specific types of content and the algorithmic nature of some technologies are potentially exacerbating internalizing symptoms in adolescents.