Patterns of crime: age, gender, place and time

Linh Tran has been looking forward to this lesson since the first class. Theories, she says, are only as good as the facts they explain, and criminology has a small set of facts so consistent that every theory has to account for them. Professor Desjardins agrees. Crime is not spread randomly across people, places or times. It is concentrated among young men, in a small number of places and at predictable hours, days and seasons, and it has risen and fallen in long waves that few people predicted. For police these patterns are not academic. They decide where officers are deployed, when shifts overlap and which problems deserve a focused effort.

The age-crime curve

The relationship between age and crime is one of the most robust findings in all of social science. In every society and period for which data exist, offending rises sharply in early adolescence, peaks in the late teens or early twenties and declines steadily afterward. The curve is steepest for crimes such as vandalism, theft and robbery, which peak earliest, and flatter for fraud, which peaks later because it requires access to money and positions of trust. In Canada, police-reported data show the rate of persons accused of crime is highest among people aged 18 to 24, and the youth crime rate for those aged 12 to 17 was 2,694 per 100,000 in 2025, down 3.8% from the year before and far below its level of the early 2000s.

Theories you have studied explain the curve in different ways. The maturity gap and adolescent brain development explain the rise and the teenage peak. Social learning stresses the importance of peers in adolescence. Control and life-course theories explain the decline through the bonds of adulthood: work, relationships, children and the routines that come with them. Moffitt's typology explains why the curve combines a large group who offend mainly as teenagers with a small group who continue. Gottfredson and Hirschi, by contrast, argued that the curve is the same for everyone and cannot be explained by social factors at all, a claim much debated. For police, the curve means that a large share of the people they deal with are young adults, and that most will stop offending as they age if they are not pushed deeper in.

Gender and crime

The second great regularity is gender. Men commit the large majority of crimes, and the gap is widest for serious violence. In Canada, about three-quarters of persons accused of crime by police are male, and men account for roughly 90% of persons accused of homicide. Women's share is highest for offences such as shoplifting and fraud and lowest for robbery, sexual offences and homicide. When women do commit violence, it is more often against partners and family members and less often against strangers. The gender gap has narrowed somewhat over time in some offence categories, mostly minor ones, but it remains large everywhere.

Explanations draw on several perspectives. Biosocial researchers point to average differences in risk taking and physical aggression. Control theorists note that girls are typically supervised more closely than boys. Feminist scholars stress how ideas of masculinity that value toughness, dominance and risk shape men's offending, and how different pathways bring women and girls into crime. Men are also more likely to be victims of violence overall, especially violence by strangers and in public places, while women are more likely to be victims of sexual violence and violence by intimate partners. Victimization, like offending, is patterned by gender.

Crime and place

The third regularity concerns place, and it has transformed policing. In 1989 Lawrence Sherman and his colleagues studied calls to police in Minneapolis and found that about half came from just 3% of addresses. Later research by David Weisburd in Seattle, New York, Tel Aviv and many other cities found that crime is concentrated on a small share of street segments, the stretch of street between two intersections. Across cities, about 50% of crime typically occurs on roughly 4% of street segments, and about 25% on roughly 1%. Weisburd called this the law of crime concentration. The concentration is remarkably stable: the same segments stay hot year after year, even as crime rises or falls across the city.

Why are some places so much more criminogenic than others? Environmental criminologists Paul and Patricia Brantingham argued that crime occurs where the routine activity spaces of offenders, the places they live, work, shop and play and the paths between them, overlap with suitable targets. They described crime generators, places such as transit hubs and shopping centres that attract many people for legitimate reasons and so produce many opportunities, and crime attractors, places such as open drug markets and certain bars that draw people looking for criminal opportunities. Some places are risky because of how they are managed: a bar that overserves, a parking lot with no lighting or a building with an absentee landlord.

Linh maps Port Ardent's 2025 incidents onto the city's 1,450 street segments. Most segments recorded no incidents at all; a handful recorded more than a hundred. The code below sorts the segments from the busiest to the quietest and calculates how much crime the busiest segments account for.

# Port Ardent 2025: how many street segments recorded each number of incidents
table = [  # (incidents on the segment, number of segments)
    (120, 5), (90, 10), (50, 15), (30, 20), (20, 30), (11, 40),
    (7, 50), (5, 60), (4, 70), (3, 90), (2, 140), (1, 250), (0, 670),
]
segments = sum(n for _, n in table)
incidents = sum(c * n for c, n in table)
print(f"segments: {segments}, incidents: {incidents}")
cum_seg = cum_inc = 0
for c, n in table:
    cum_seg += n; cum_inc += c * n
    if c in (90, 30, 11, 4, 1):
        print(f"busiest {cum_seg:4d} segments ({cum_seg / segments:5.1%}) hold {cum_inc / incidents:5.1%} of incidents")
print(f"segments with no incidents: {670 / segments:.0%}")

segments: 1450, incidents: 5620
busiest   15 segments ( 1.0%) hold 26.7% of incidents
busiest   50 segments ( 3.4%) hold 50.7% of incidents
busiest  120 segments ( 8.3%) hold 69.2% of incidents
busiest  300 segments (20.7%) hold 85.8% of incidents
busiest  780 segments (53.8%) hold 100.0% of incidents
segments with no incidents: 46%

Port Ardent fits the law of crime concentration closely. The busiest 15 segments, about 1% of the total, account for about a quarter of all incidents; the busiest 50, about 3.4%, account for half; and almost half of all segments recorded no crime at all. The zone-level map from lesson 6 hid this. Even in zone 1, the highest-crime zone, most streets are quiet, and the crime is concentrated on a few blocks around the bus terminal, two bars and a rooming house. Focusing on those blocks is far more efficient than spreading patrols across the whole zone, which is the idea behind hot spots policing.

Crime and time

Crime also follows the clock and the calendar. Assaults and disturbances rise sharply in the evening, peak around bar closing time on Friday and Saturday nights and fall to their lowest levels in the early morning hours after 5 a.m. Residential break-ins happen most often on weekday daytimes, when homes are empty, while commercial break-ins happen mostly at night, when businesses are closed. Thefts from vehicles cluster at night in residential streets and in the daytime in commuter parking lots. Many offences, especially assaults and disorder, rise in summer, when people spend more time outdoors and in public places. Domestic violence calls rise around holidays and major sporting events.

These temporal patterns follow from routine activity theory, which you will study in the next lesson: crime happens when offenders and targets come together in the absence of capable guardians, and daily and seasonal routines decide when that happens. They matter greatly for staffing. Port Ardent's shift schedule, which you studied in the first course, sets higher night-shift minimums because calls for disorder and violence peak at night. Professor Desjardins says that the single most useful chart in a police service is often a simple table of incidents by hour and day of week, because it shows where officers are needed and when.

Patterns in online and financial crime

Online crime breaks some of these patterns and follows others. Fraud, extortion and the sharing of intimate images do not need the offender and victim to be in the same place, so street-segment maps and patrol schedules do not capture them. A scammer in another country can target thousands of Port Ardent residents in an afternoon. Even so, criminologists find familiar regularities. Victimization is concentrated: a small share of people, often older adults targeted by romance and investment scams or young people targeted by sextortion, suffer repeated attacks. Offending is concentrated too, in organized groups that run fraud as a business. And timing still matters: scam calls and messages spike around tax deadlines, holidays and major news events, when people expect messages from governments, delivery companies and charities.

The Canadian Anti-Fraud Centre, run jointly by the RCMP, the Competition Bureau and the Ontario Provincial Police, estimates that only a small share of fraud victims report to it or to police, so recorded fraud greatly understates the problem. Priya notes that her bank's internal data showed the same pattern as police data on street crime: a few accounts, merchants and devices accounted for most of the losses, which is why banks, like police, focus their attention on concentrated risk. For police services, online crime raises a practical question about deployment, since the officers best placed to respond may be analysts and investigators working with banks and platforms, not patrol officers on the street.

Long-term trends: the great crime decline

Crime rates also move in long waves. In both countries, police-reported crime rose steeply from the early 1960s to a peak around 1991, then fell for more than two decades. The national crime rate fell by roughly half between 1991 and 2014, and the Crime Severity Index fell by about a third. Crime rose again between 2015 and 2022, driven partly by fraud, shoplifting, extortion and violent offences, before falling in 2024 and 2025. Similar declines occurred in many other high-income countries in the 1990s and 2000s, which suggests that causes specific to one country cannot be the whole explanation.

Criminologists still debate the causes of the great crime decline. Candidate explanations include better security, such as vehicle immobilizers and home alarms, which made common property crimes harder, an idea called the security hypothesis; the ageing of the population; changes in policing, including hot spots and problem-oriented strategies; the decline of crack cocaine markets; economic growth; the removal of lead from gasoline; increased imprisonment, mainly in the United States, though Canada's crime decline was similar without a large rise in imprisonment; and the move of young people's leisure time indoors and online. Most researchers believe several of these combined. Linh notes that the internet may also have shifted some crime rather than reduced it, since fraud and online offences have grown as traditional property crime fell.