The Pattern Test SOC 3333 · Crime Lab · Midwestern State University
Crime is patterned, not random, and those patterns are the reason criminological theory exists. Everything in the second half of this course is an attempt to explain the regularities on this page. Work through the five patterns, guessing before each reveal, and then take the self-check at the end, which covers everything from the first unit. There is no forum this week, so this page is a study tool rather than an assignment. Nothing you click is recorded, which also means nobody sees your quiz score but you.
The age-crime curve, and the news about it
For a century the age-crime curve was the most reliable finding in criminology. Offending rises sharply in adolescence, peaks in the late teens, and declines for the rest of life. It showed up across eras, countries, and offense types, and it created the field's deepest puzzle: most offenders stop on their own, with no arrest, treatment, or punishment.
In the FBI's 2024 arrest data, at roughly what age does the total arrest rate now peak?
The curve did not disappear. It moved, because the young end fell out from under it. Arrest rates for teenagers have collapsed. For 15 to 19 year olds, aggravated assault arrest rates fell more than 43 percent since 2006. Burglary arrests for that group fell nearly 89 percent. Larceny-theft arrests for 10 to 14 year olds fell nearly 82 percent. Meanwhile, arrest rates for people 30 and older stayed flat or rose modestly. Subtract a collapsing left side and the peak slides right.
Why this is the perfect thing to study during an exam week. It forces you to use three things from this unit at once. From Chapter 1, remember that a good theory has to be falsifiable, and a theory built on teenage peak offending just met evidence it has to answer for. From Chapter 2, remember that these are arrest data, so part of any change could be a change in policing rather than a change in behavior. From Chapter 5, remember that a correlate is not a cause and the two classic readings of the curve, maturation and cohort differences, make different predictions about exactly this. Maturation says everyone ages out, so the curve should be stable. Cohort thinking says each generation is different, so the curve can move. One of those just got better evidence.
Sources: analysis of the FBI's 2024 arrest data reporting a total arrest rate peaking near age 32, with murder arrests peaking at 18 and motor vehicle theft at 15, published in a 2025 examination of the age-crime curve. Percent declines since 2006 for ages 15 to 19 and 10 to 14 are from the same analysis. The classic late-teen peak is the long-standing finding described in criminology textbooks including Zito, Chapter 5. Retrieved August 2026.
The gender gap, and what scales it
Men out-offend women in every society for which we have data, and the gap is largest for the most serious offenses. Before you look at the numbers, put one down yourself.
Across all offenses combined, what percent of arrests in the United States are of men?
The gap scales with the seriousness of the offense. Nearly nine in ten murder arrests are of men. Larceny-theft is close to even. Embezzlement is essentially even, which is worth sitting with, because embezzlement is a crime of workplace access rather than physical force. When the offense requires a position rather than a body, the gap closes almost completely. That single comparison does more explanatory work than most sentences in this chapter.
Three explanations to keep separate, because the exam will not blend them for you. The first is opportunity: different lives create different access to different offenses, which is exactly what the embezzlement row suggests. The second is gender performance, the idea that some offending is a way of doing masculinity, which you will meet again in the subcultural theories of Chapter 9. The third is a measurement question rather than a behavioral one. The chivalry hypothesis asks whether the system arrests and charges women more leniently, which would mean part of this gap is a gap in enforcement rather than in conduct. Chapter 2 taught you to always ask that question, and here it has a name.
Source: FBI Uniform Crime Reports, Table 42, Arrests by Sex, 2019 (10,831 agencies, estimated population 229,735,355). Percent male: total all offenses 72.5, murder and nonnegligent manslaughter 88.0, robbery 84.2, burglary 79.3, motor vehicle theft 76.8, aggravated assault 76.5, larceny-theft 57.4, embezzlement 49.8. Remember Chapter 2: these are arrests, not offenses, so they measure enforcement as well as behavior. Retrieved August 2026.
Crime concentrates, and it concentrates hard
Ask where crime happens and most people answer with a neighborhood. The research says the right unit is far smaller than that: the street segment, meaning one block face between two intersections. Below is a grid standing in for 400 street segments in a city.
What share of a city's street segments do you think produce about half of all its crime?
Every square is a street segment. The highlighted ones produce roughly half the crime in the city.
This is Weisburd's law of crime concentration, and its most surprising feature is stability. In Seattle, the share of segments producing half the crime stayed inside a narrow band for sixteen years while the city's overall crime rate fell by more than twenty percent. Hot spots stayed hot. The finding has since replicated in cities of very different sizes, including suburban ones where two percent of segments produced half the crime.
What it does to theory, which is the point of this whole page. If crime concentrates this tightly at addresses, then explanations built entirely around the characteristics of offenders are incomplete, because offenders are mobile and street corners are not. Something about particular places keeps generating crime across decades and across the people who pass through them. Chapter 14 takes this seriously enough to build a whole prevention strategy on it, and asks the question this pattern implies: what happens if you treat the place instead of the person?
Sources: Weisburd's longitudinal study of Seattle street segments, which found 50 percent of crime incidents at between 4.7 and 6.1 percent of street segments across 16 years. Weisburd's law of crime concentration, with a range of 2.1 to 6.0 percent of streets producing 50 percent of crime across eight additional cities, and 0.4 to 1.6 percent producing 25 percent. The grid above is an illustration of the proportion, not a map of any real city. Retrieved August 2026.
The biggest pattern of all
The single most important fact about crime in your lifetime is that there is much less of it than there used to be. Drag through the decades and watch the homicide rate, the best-measured crime there is.
Two families of explanation, and a lesson in humility. The first family points at specific opportunity changes: engine immobilizers in cars, less cash carried in pockets, better emergency medicine turning some homicides into aggravated assaults, security technology everywhere. The second points at broad shifts in how people live: an aging population, more time indoors and online, changes in policing and incarceration. Both have evidence and neither is settled. What matters more than picking a side is noticing that almost nobody predicted the decline, and that the criminology of the early 1990s confidently expected the opposite. Every theory you meet starting in Chapter 6 should be held with that in mind.
Sources: FBI Uniform Crime Reports historical homicide rates per 100,000 for 1960 through 2019 (1960: 5.1, 1970: 7.9, 1975: 9.6, 1980: 10.2, 1985: 8.0, 1990: 9.4, 1991: 9.8, 1995: 8.2, 2000: 5.5, 2005: 5.6, 2010: 4.8, 2014: 4.4, 2015: 4.9, 2019: 5.0). FBI and BJS published estimates for 2020 through 2024 (2020: 6.5, 2022: 6.5, 2024: 5.0). Every point plotted above is a published figure. The following year is not plotted because only preliminary estimates exist for it: the FBI reported an estimated 18.1 percent decline in murder, and the Council on Criminal Justice projected a rate near 4.0, which would be the lowest ever recorded in data reaching back to 1900. Retrieved August 2026.
Self-check: Chapters 1 through 5
Ten questions in the style of the exam, drawn from the whole unit. Answer each one, then read the explanation whether or not you got it right, because the explanation is where the studying happens. Nothing is recorded and no score leaves this page.
How to read your score. If you missed questions about data sources, go back to the Chapter 2 comparison table first, because those concepts appear in every chapter after it. If you missed definition questions, the fix is the offense definitions in Chapter 3 and the vocabulary slides in each deck. If you missed the pattern questions, this page is the review. Take it again later in the week and see what changed.
The checklist every theory has to pass
This is the payoff of the whole first unit. Starting with the next chapter, every theory you meet gets graded against this list. A theory that explains one item and contradicts three is not a good theory, however satisfying it sounds.
- Why offending rises in adolescence and fades with age, and why most people stop without any intervention at all.
- Why the peak of the arrest curve appears to be moving later as youth offending collapses.
- Why men out-offend women in serious violence in every society we can measure, and why that gap nearly closes for offenses of workplace access.
- Why crime concentrates on a tiny share of street segments, and why those same segments stay hot for decades.
- Why crime follows daily and seasonal rhythms, peaking at night, after school for juveniles, and in summer for violence.
- Why immigration does not raise crime rates, contrary to what most people assume.
- Why crime fell by half after the early 1990s, rose sharply in 2020, and then fell again at record speed.
- Why unemployment and crime do not move together in the simple way folk theory expects, as the Great Recession showed.
- Why so much youthful offending happens in groups rather than alone.
Use this in two places. On the exam, a short answer question that asks you to evaluate a theory is really asking whether you can hold it up against a pattern like these. In your term paper, Part 3 asks for exactly this kind of pattern evidence about your chosen offense: who offends, who is victimized, and where and when it happens. Same skill, two assignments.
One discipline to carry with you. Every item above is a statement about rates and groups. None of them is a statement about any individual person, and none of them is a claim about a cause. Say associated with, not causes, until you have the evidence that earns the stronger word. Chapter 1 gave you the three requirements for that upgrade: association, temporal order, and non-spuriousness.