Timing Matters: Age of Onset in Online Child Sexual Solicitation and Abuse and Young Adults’ Well-Being

Technical Summary: Timing Matters: Age of Onset in Online Child Sexual Solicitation and Abuse and Young Adults' Well-Being

Problem Statement
Online child sexual solicitation and abuse (OCSSA) is a widespread phenomenon, with approximately one-third of young adults reporting exposure before age 18. While existing literature documents the adverse psychological correlates of OCSSA, the role of developmental timing across the spectrum of OCSSA subtypes remains poorly understood. Most empirical studies conceptualize exposure dichotomously (victimization vs. non-victimization) rather than examining the specific age of first onset. Furthermore, prior evidence regarding age-related gradients derives primarily from image-based sexual abuse (IBSA), leaving a gap in understanding whether comparable timing-related patterns exist for high-prevalence, solicitation-based forms. Additionally, the independent prognostic relevance of developmental timing has rarely been isolated from co-occurring offline child maltreatment (CM) and sociodemographic factors.

Methodology
The study utilized data from a nationally representative German survey of $N = 1,006$ young adults (aged 18–29). The research employed an exploratory design without a priori hypotheses regarding specific age-of-onset combinations, aiming to identify robust predictors of psychological well-being.

• Measures:
  • OCSSA: Assessed retrospectively using a 9-item screening tool covering nine subtypes (e.g., non-consensual sharing, unwanted sexualized talk, unwanted exposure to pornographic material). Age of first exposure was operationalized as single-year indicators (ages 1–18).
  • Outcome: Psychological well-being was measured using the WHO-5 Well-Being Index.
  • Covariates: Offline CM (dichotomized yes/no based on the Adverse Childhood Experiences Questionnaire), age, gender, educational attainment, and equivalized household income.
• Statistical Strategy:
  To address the high-dimensional predictor space (29 predictors including 24 age-of-onset dummy variables) and mitigate overfitting, the authors employed a dual machine learning approach:
  1. LASSO Regression: Used for automatic variable selection via L1 regularization to identify predictors with non-zero coefficients.
  2. Random Forest Regression: Used to capture non-linear relationships and rank variable importance via permutation (%IncMSE).
  3. Post-Hoc Multiple Linear Regression: Conducted on the top predictors identified by the machine learning models to provide interpretable effect estimates (unstandardized coefficients, $B$) and formal significance testing.
     Missing data were handled using K-Nearest Neighbor (KNN) imputation.

Key Results

• Prevalence: 32.3% of participants reported at least one form of OCSSA before age 18. First exposure clustered primarily in mid-to-late adolescence (ages 13–17).
• Dominant Predictors: Offline CM was the strongest predictor of reduced well-being ($B = -17.66, p < .001$) in the post-hoc model, followed by equivalence income below the poverty risk threshold ($B = -7.11, p < .001$).
• Timing-Specific Findings: Three OCSSA age-of-onset predictors were identified by both machine learning models:
  1. Unwanted exposure to pornographic/sexualized material at age 13.
  2. Unwanted sexualized talk at age 15.
  3. Unwanted exposure to pornographic/sexualized material at age 15.
• Statistical Significance: In the post-hoc multiple linear regression, only unwanted exposure to pornographic/sexualized material at age 13 remained statistically significant ($B = -11.22, p = .005$). Unwanted sexualized talk at age 15 showed a marginal association ($p = .065$). All other age-of-onset variables failed to reach statistical significance in the multivariable model.
• Convergence: The LASSO and Random Forest models converged on offline CM, income, and specific age-of-onset variables (particularly age 13 and 15) as the most important predictors, strengthening the robustness of these findings.

Key Contributions

• Developmental Timing: The study provides initial population-based evidence that the developmental timing of first exposure to specific OCSSA subtypes is associated with psychological well-being in young adulthood, independent of offline CM and sociodemographic factors.
• Granularity of Onset: By operationalizing age of onset as single-year indicators rather than dichotomized groups (e.g., before/after age 12), the study identifies discrete vulnerability windows, specifically highlighting early-to-mid adolescence (ages 13 and 15) as potentially sensitive periods.
• Subtype Expansion: The findings extend timing-related evidence beyond image-based abuse to include solicitation-based experiences (e.g., unwanted talk, unwanted exposure), suggesting that non-coercive forms of OCSSA can also be linked to long-term well-being deficits at specific developmental stages.
• Methodological Rigor: The use of a dual machine learning approach (LASSO and Random Forest) followed by post-hoc regression allows for robust variable selection in a high-dimensional space while controlling for the confounding effects of offline maltreatment.

Significance and Claims
The authors claim that their findings suggest prevention efforts should not focus solely on late adolescence, where OCSSA exposure peaks, but should begin earlier. Specifically, the association between unwanted exposure to sexualized material at age 13 and reduced well-being suggests that age-appropriate digital sexual literacy should be established before this age. The convergence of findings at age 15 indicates that mid-adolescent programs should address the broader spectrum of OCSSA, including solicitation-based experiences often minimized as less severe.

The paper modestly frames these findings as "initial population-based evidence" and "descriptive indications" rather than confirmatory causal proofs, citing the cross-sectional design and retrospective nature of the data as limitations. The authors emphasize that offline CM and socioeconomic disadvantage remain the strongest predictors of poor well-being, cautioning against treating OCSSA prevention as a stand-alone digital safety issue. Instead, they argue for integrated prevention strategies embedded within broader child protection and family support systems. The study concludes that the developmental timing of exposure interacts with neural and social maturation, making early-to-mid adolescence a critical window for intervention.