Risk factors and missing persons: advancing an understanding of ‘risk’

The paper addresses how ‘risk’ for going missing should be conceptualized and studied, noting that missingness is a persistent global issue with limited, inconsistent scholarship. Prior studies identify various correlates—often labeled risk factors—across demographics, psychopathology, and environment, but findings are mixed and typically treated as fixed, decontextualized predictors. The study’s purpose is to clarify risk terminology, distinguish true risk factors from concomitants or consequences, and to move beyond listing correlates toward understanding dynamic mechanisms and pathways through which factors relate to missingness. This is important for developing accurate police risk assessments and effective prevention policies, especially given evidence that risk varies by population and context and that multiple, intersecting factors likely contribute to missing episodes.

The review organizes reported correlates of missingness into demographic, psychopathological, and environmental domains. Demographic: female and male sex/gender have each been associated with missingness in different contexts; Indigenous identity; and young age are frequently implicated. Data constraints limit examination of other demographics. Physical and cognitive disabilities (e.g., dementia/Alzheimer’s), and medical conditions (e.g., diabetes, eating disorders) have been linked to missingness; trauma exposure (physical/sexual/domestic abuse) appears to increase risk. Psychopathological: diagnosed and undiagnosed mental health problems (e.g., schizophrenia), suicidal ideation, self-harm, and substance use/abuse are consistently associated with missingness, particularly in mental health settings. Environmental: unemployment or being out of the labor force, financial strain and limited opportunities, relationship breakdown and family conflict/rejection, life strains and immediate social pressures, homelessness or transient living, and sex trade work are repeatedly cited. The literature hints at interrelations among factors (e.g., intersections of high-risk categories; transient lifestyles linking homelessness and sex trade work). However, many studies are cross-sectional, limiting causal inference and clarity on temporal ordering.

The study is a narrative review and conceptual application. Steps: (1) Review international literature on missing persons to collate reported correlates commonly labeled risk factors; (2) Apply Kraemer et al.’s (1997, 2001) risk factor classification system (fixed markers, variable risk factors, causal risk factors; distinction from concomitants and consequences) to clarify terminology and status of factors; (3) Use the MacArthur framework to outline and exemplify five pathway types by which variables can relate to missingness: independent risk factors, overlapping risk factors, proxy risk factors, mediators, and moderators; (4) Provide illustrative examples from the missing persons literature for each pathway type; (5) Discuss implications for research design (emphasizing temporal precedence, longitudinal needs, interaction terms) and for police practice and policy. No primary data are analyzed; rather, the paper synthesizes prior findings and maps them onto a theoretically grounded framework to propose testable pathways and a common vocabulary.

• Many variables commonly labeled ‘risk factors’ for missingness lack evidence of temporal precedence and are better classified as concomitants or consequences; confidence in causal claims is weak due to predominant cross-sectional designs and limited data retention for longitudinal analysis.
• Fixed markers (e.g., sex/gender, race/ethnicity) can identify higher-risk groups but cannot be manipulated to change outcomes; variable risk factors may change over time but often lack evidence that changing them alters missingness; few factors meet criteria to be called causal risk factors.
• Risk likely operates through multiple pathways that cross categorical boundaries: (a) independent risk factors (rarely demonstrated robustly); (b) overlapping factors reflecting shared constructs (e.g., financial strain and lack of employment opportunities as economic vulnerability; homelessness and sex trade work as indicators of transient lifestyles); (c) proxy risk factors (e.g., sex/gender or Indigenous identity may proxy underlying mechanisms like resource accessibility or economic vulnerability); (d) mediators explaining how one variable affects missingness; and (e) moderators specifying for whom/under what conditions effects occur (e.g., coping style moderating effects of life strain, relationship conflict, trauma on missingness).
• The framework suggests combining overlapping indicators into stronger constructs (e.g., transient lifestyles) and disaggregating proxies to locate underlying mechanisms.
• Standard practice of regressing missingness on numerous candidate predictors without modeling mediation/moderation or considering temporal order risks misleading conclusions and null results due to unmodeled heterogeneity.
• Consistent terminology and theoretically driven models of pathways can improve risk assessments, identification of low- vs. high-risk groups, and prevention strategies.

By clarifying risk terminology and mapping reported correlates onto Kraemer/MacArthur categories, the paper reframes the central question from identifying standalone predictors to understanding mechanisms and pathways linking correlates to missingness. This addresses inconsistencies in prior findings by recognizing that effects differ across populations, contexts, and interacting variables. The proposed pathway lens (independent, overlapping, proxy, mediator, moderator) explains how multiple vulnerabilities (e.g., economic strain, transient lifestyles) can compound risk and why some demographic markers function as proxies for structural mechanisms. The significance lies in guiding researchers toward designs and analyses that establish temporal precedence where feasible, incorporate interaction terms to test moderation, and evaluate mediation to uncover mechanisms, thereby improving interpretability and applicability of findings. For practice, fixed markers can aid triage while variable mechanisms (e.g., coping style) suggest targets to mitigate the impact of other risk exposures, refining police risk assessments and informing prevention and social policy.

The paper contributes a common language and conceptual framework for studying risk in missing persons research, applying Kraemer/MacArthur principles to classify factors and articulate plausible risk pathways. It concludes that many purported risk factors are likely concomitants, consequences, or proxies, and that missingness arises from multiple intersecting pathways rather than single causes. The author calls for: (1) consistent, precise use of risk terminology; (2) theoretically driven models that test mediation and moderation; (3) where possible, longitudinal or temporally sensitive designs to establish precedence; (4) aggregation of overlapping indicators into robust constructs (e.g., transient lifestyles) and disaggregation of proxies to identify underlying mechanisms; and (5) pragmatic use of fixed markers for identification alongside modifiable variables for intervention. Future research should prioritize pathway-based analyses, improve data collection on demographics and contextual factors, and develop prevention strategies targeting mechanisms (e.g., coping interventions, reducing economic/structural vulnerabilities).

• Predominance of cross-sectional, observational studies in the field prevents establishing temporal precedence and causal inference, complicating classification of variables as true risk factors versus concomitants or consequences.
• Limited availability and retention of police data (often ≤5 years) hinder longitudinal designs.
• The Kraemer/MacArthur framework was developed largely in contexts amenable to randomized or longitudinal evaluation, making its stringent criteria challenging to meet in missing persons research.
• Existing datasets often lack comprehensive demographic and contextual variables, constraining exploration of additional risk mechanisms and subgroup differences.
• Examples provided are illustrative; absence of primary data analysis means proposed pathways require empirical testing.