Age–Crime Trajectories, Birth Cohorts & Moffitt’s Hypothetical Curve

Birth cohorts = samples defined by year of birth; data are directly collected from individuals or their families, not drawn from official police, court, or administrative records.
- Methods used: community surveys, structured interviews, victimization questionnaires.
- Goal: obtain the largest possible volume of crime‐related information by asking the people involved.
Key implication: these data give a person‐centred view of offending, complementing record-based studies.

Classic curve shows that participation in crime/antisocial behavior rises in early adolescence, peaks in the late teens, and declines thereafter.
Aggregated curves treat the entire sample as one group, so they mask within-person heterogeneity (i.e.
individual trajectories).

Terry Moffitt’s work (to be revisited later) illustrates how the prevalence of antisocial behavior changes across the life course.
- Prevalence = proportion of the population committing any antisocial act at a given age.
- Key pattern: prevalence declines steadily with age—people are less likely to offend as they grow older.
Critical insight:
- An aggregate prevalence curve cannot reveal who is offending at each age.
- Two (non‐exclusive) possibilities:
1. The same individuals offend early and persist.
2. Early starters desist, and a different set (later starters) takes over.
Moffitt’s hypothesis: childhood‐onset offenders are most likely to persist, whereas later-onset offenders tend to have shorter careers.

From the aggregate curve we cannot infer an individual’s criminal career length, onset, or peak.
Need for statistical methods that disaggregate the curve into trajectory groups whose within-group members show similar longitudinal patterns.

Uses techniques (e.g., finite‐mixture or latent‐class growth models) to sort individuals into distinct criminal-career patterns.
Each trajectory group has:
- A unique shape (onset, peak age, desistance rate).
- Members who are more similar to each other than to the overall sample.

Dark “blurple” line (Lowest‐level offenders)
- Very low involvement; occasional acts during the teens; near-zero beyond the early $20\text{s}$ .
Blue line (Early childhood onset, early desistance)
- Rapid rise in early/mid childhood; peak during the teen years; sharp drop by the early $20\text{s}$ .
Magenta/Cerise line (Late‐adolescent peakers)
- Peaks at about $19$ ; mirrors the classic age–crime curve but still tapers into the early $30\text{s}$ .
Red line (Early‐adult high peakers)
- Highest peak around $21$ ; gradual decline yet still elevated well into the $30\text{s}$ .
Yellow line (Chronic high‐rate offenders)
- Rapid ascent to a very high level; maintains high offending through the $30\text{s}$ .

The depicted trajectories stop at roughly $30$ – $35$ .
Open research questions:
- What happens beyond $30$ , $40$ , $60$ ?
- Montreal longitudinal studies have followed cohorts longer, but those results are not shown in this figure.

Amy Adams slide (mentioned but not detailed here) also highlighted childhood‐onset persistence, aligning with Moffitt’s view.
Birth‐cohort data bridge the gap between official records and self-report measures, enriching analysis of criminal careers.

Understanding distinct trajectories assists in:
- Tailoring early-intervention programs to childhood‐onset chronic offenders.
- Allocating resources efficiently by distinguishing low-risk intermittent offenders from persistent high-rate offenders.
Misinterpreting aggregate curves risks overgeneralizing policy responses.

Self-report birth-cohort studies provide granular insight into offending not captured by official statistics.
Aggregate age–crime curves hide meaningful individual differences; trajectory analysis uncovers them.
Moffitt’s framework suggests that age of onset is a powerful predictor of criminal-career length and intensity.
Further longitudinal data (beyond the $30$ s) are crucial to complete the life‐course picture of crime.