Reels to Remembrance: Attention Partially Mediates the Relationship Between Short-Form Video Addiction and Memory Function Among Youth

Short-form video (SFV) platforms (e.g., TikTok, Instagram Reels, YouTube Shorts) deliver rapid, algorithmically curated content that encourages continuous engagement through vertical feeds, hyper-personalized recommendations, and viral challenges. These design features may increase risk for addictive use patterns and raise concerns about cognitive consequences. Prior research has largely examined broader internet and social media addiction, with mounting evidence linking problematic digital media use to diminished attention. Given the theoretical interdependence of attention and memory—where attention filters and facilitates encoding into memory—this study posits that SVA may impair attention and thereby compromise memory function. The study’s aims were: (1) to explore the prevalence and patterns of SVA among youth; (2) to identify correlates of SVA; and (3) to test whether attention mediates or moderates the relationship between SVA and memory function, focusing on Arab youth, a highly engaged yet underrepresented population with elevated mental health risks.

Existing literature documents problematic social media and internet use and their associations with adverse psychosocial and cognitive outcomes. Studies link higher frequencies of internet use with reduced attention and potential neurodevelopmental changes in attention-related brain regions. Research on excessive media use implicates deficits in executive attention. However, SFV platforms present unique risks due to short, bingeable content and powerful recommendation algorithms that maximize engagement via rewards (likes, comments) and participatory features (challenges, lip-syncing). Despite rapid SFV adoption, few studies directly assess SVA’s cognitive effects, particularly on memory. Theoretical accounts emphasize attention as a selective, capacity-limited gateway essential for encoding into memory; thus, attention deficits may degrade memory performance. This study addresses gaps by focusing specifically on SFV platforms and examining both attention and memory among Arab youth.

Design and setting: A national cross-sectional survey was conducted in Jordan, which has a predominantly young population. Data collection occurred from 30 December 2022 to 28 February 2023 across 25 higher-education institutions selected via a weighted sampling strategy to represent urban, suburban, and rural regions based on Ministry of Higher Education enrollment and population density data. Participants: Of 1051 returned questionnaires, 1029 were valid for analysis (missing data rate 2.1%; no imputation performed). Measures: (1) Short-Form Video Addiction Scale (Mu Honglei et al.), assessing unsuccessful reduction attempts, study interference, social impact, and agitation when not using SFV apps on a 5-point Likert scale; higher scores indicate greater SVA (Cronbach’s α = 0.73). (2) UGC Perception Scale (five items on immersion/impact; 5-point scale; α = 0.72). (3) Prospective and Retrospective Memory Questionnaire (PRMQ), assessing frequency of memory errors across eight factors with 5-point responses; α = 0.92. (4) Attention Control Scale (ATTC; 20 items; 4-point scale) with focusing and shifting subdomains; α = 0.82. All instruments were translated to Arabic using Chapman and Carter’s procedures. Additional covariates included institution, year of study, major, GPA, daily smartphone and SFV usage, daily studying hours, physical activity, and mental/general health history, and psychological help-seeking. Analysis: Descriptive statistics characterized SVA prevalence and platform use. Predictors of SVA were examined using linear modeling. Mediation of the SVA–memory relationship by attention was tested via structural equation modeling (SEM), controlling covariates selected from the predictor model (p < 0.10). Moderation by attention was examined using multigroup SEM (median split of attention score: ≥49 vs <49) comparing unconstrained vs constrained models. Analyses used IBM SPSS 29.0.1 and AMOS 29.0.0. The SEM included 33 parameters; sample size (n=1029) exceeded rule-of-thumb requirements (parameters × 20 = 660). Model fit indices included χ², RMSEA, GFI, IFI, TLI, and CFI.

Sample and usage patterns: Of 1029 valid respondents, 31% were male; 56% were in years 1–2; 45.2% in medical sciences; 20% in engineering/tech; 21% in arts/humanities. Mean SVA score was 13.15 (SD 3.61; range 4–20). Distribution: 50% scored >13 and 25% >16. Platform use: Instagram 73.4%, TikTok 45.9%, Facebook 43.6%, YouTube 30.0%, Snapchat 12.8%, Twitter 4.5%, Likee 1.1%, other 2.7%. Daily smartphone use >8 h reported by 18.8%. In response to the December 2022 TikTok ban, 15.0% felt annoyed and 6.5% extremely annoyed. Predictors of SVA: Model explained 48.1% of variance (adjusted R²=0.474). Higher SVA scores were associated with female gender (F=9.09, p=0.003), younger year of study (F=4.04, p=0.045), lower GPA (F=6.16, p=0.013), fewer study hours (F=7.67, p=0.006), and strongly with UGC perception (F=766.12, p<0.001). Study major, psychiatric diagnosis, and receiving psychological help were not significant predictors. Mediation: Attention partially mediated the SVA–memory relationship with excellent fit (χ²(12)=14.11, p=0.05; RMSEA=0.03; GFI=0.99; IFI=0.99; TLI=0.98; CFI=0.99). Direct effect: SVA → memory γ=0.19, p<0.001. Indirect effect through attention: α×β = (−0.31)×(−0.31)=0.10, p<0.001. Moderation: No evidence that attention moderated SVA’s effect on memory; unconstrained vs constrained models did not differ significantly (Δχ²=11.00, Δdf=8, p=0.201). The SVA → memory path was similar in high-attention (β=0.23, p<0.001) and low-attention (β=0.24, p<0.001) groups.

The study demonstrates widespread engagement with SFV content among Jordanian youth and identifies key correlates of higher SVA, including female gender, earlier academic years, lower GPA, and fewer study hours. Crucially, attention partially mediates the relationship between SVA and memory, aligning with cognitive theory that attention supports encoding and memory performance. However, the absence of moderation indicates that the detrimental association between SVA and memory is relatively consistent across different attention capacities. These findings suggest that while baseline attention influences memory pathways, SVA’s impact on memory is broadly pervasive. The results underscore the need for interventions that manage attention and reduce problematic SFV use, including digital literacy initiatives, academic support, and mental health services. Policy considerations should balance effective regulation with ethical concerns, recognizing differing governance models and enforcement challenges across contexts.

This study contributes evidence that short-form video addiction is common among Jordanian university students and is associated with cognitive risks, particularly concerning attention and memory. Attention partially mediates the SVA–memory link, whereas moderation by attention was not supported, indicating a consistent adverse relationship across attention levels. A comprehensive response is recommended: (1) digital literacy programs to promote balanced media consumption; (2) campus-based mental health and academic support services; and (3) policy interventions and platform-level features that mitigate excessive use. Future research should employ longitudinal and experimental designs to clarify causal pathways, incorporate objective digital behavior metrics, and evaluate targeted attention-management and usage-reduction interventions across diverse cultural settings.

The cross-sectional design precludes causal inference and temporality assessment among SVA, attention, and memory. Self-reported measures (including SVA, attention, memory, and media use) may introduce recall and social desirability biases. The median split used for multigroup moderation may reduce sensitivity to nuanced moderation effects. Although the sample was large and geographically diverse within Jordan, generalizability beyond Jordanian university students may be limited. Unmeasured confounders (e.g., sleep quality, comorbid psychopathology, academic stress, or specific platform features) could influence observed relationships.