Three Things Have to Line Up · SOC 3333 Crime Lab

A puzzle that broke the standard explanation

Through the 1960s and into the 1970s the United States got richer. Unemployment fell, incomes rose, more people finished high school, more households owned cars and appliances. Every major criminological theory of the period expected crime to fall along with poverty. Before you look at what happened, commit to a guess.

The FBI reported a violent crime rate of 160.9 per 100,000 people in 1960. What do you think the rate was fifteen years later, in 1975?

200

Violent crime rate, 1960

160.9

Violent crime rate, 1970

363.5

Violent crime rate, 1975

487.8

Rates per 100,000 population. FBI Uniform Crime Reports historical series, 1960 to 1975. Property crime moved the same direction over the same years, from 1,726.3 per 100,000 in 1960 to 4,810.7 in 1975. Retrieved August 2026.

The violent crime rate roughly tripled while the economy improved. That is the puzzle Cohen and Felson set out to solve, and it is why they went looking for an explanation that had nothing to do with how many people wanted to commit crime. Read these figures with the caution Chapter 2 taught you. They come from the UCR Summary series, they are police-recorded rather than survey-measured, and reporting practices changed across this period. The rise is far too large to be an artifact of counting, but the exact size of it is not something the UCR can pin down.

Their answer: nobody was home

Cohen and Felson noticed that American daily routines had been rearranged. More women had entered paid work outside the home, so houses stood empty during the day. More people ate out, traveled, and spent time away from their own property. And a wave of new consumer goods was small, valuable, and easy to carry off. In their own data, married women's labor force participation rose 31 percent between 1960 and 1970. Prosperity did not reduce crime. It redistributed people and property across space in a way that put more unguarded targets in front of whatever offenders already existed.

The theoretical move here is worth naming. Cohen and Felson did not argue that offender motivation is unimportant. They argued that it can be treated as roughly constant and that the interesting variation lives elsewhere. That is a claim about what to study, not a claim that nobody chooses to offend. Chapter 8 onward will spend a great deal of time on the motivation side.

Cohen, L. E., and Felson, M. (1979). Social change and crime rate trends: A routine activity approach. American Sociological Review 44(4), 588 to 608.

Break the triangle

Pick a situation. All three elements start switched on, which means the crime happens. Switch any one of them off and read what changes. Pay attention to what you are actually doing when you switch an element off, because in every case it corresponds to something real that somebody could decide to do.

Motivated offender

Somebody willing to do it

Suitable target

Worth taking, and takeable

Capable guardian

Somebody whose presence would stop it

Notice that the third element runs the other way. The offender and the target have to be there for a crime to happen. The guardian has to be missing. That asymmetry is why criminologists write the third element as the absence of a capable guardian rather than the presence of an incapable one, and it is why this element is the easiest of the three for policy to reach. Adding a person to a place is a thing you can actually do.

What you did not do: you never rehabilitated anyone, never raised anyone out of poverty, and never changed anyone's values. You moved an object, added a person, or changed a routine, and the event stopped happening. That is the entire policy appeal of this theory, and it is also the source of the most common objection to it, which is that preventing an event in one place is not the same as reducing the number of people who want to offend.

Suitability in Cohen and Felson's framing has four components: value, visibility, accessibility, and inertia, meaning how hard the thing is to move. The scenarios above vary all four.

The word "capable" is doing a lot of work

A guardian in this theory is almost never a police officer. It is a neighbor on a porch, a clerk who looks up, a roommate who notices, a dog, a lit window. Cohen and Felson meant guardianship by ordinary people of one another and of each other's property, as a side effect of going about ordinary life. But saying somebody is there is not the same as saying they are capable. Danielle Reynald's research broke capability into three separate stages, and a person can fail at any one of them.

1

Available

Physically present, or close enough that presence could register. Nobody in the building, nobody on the street, nobody in the group chat.

2

Monitoring

Actually paying attention to the surroundings. Reynald found this is the stage that separates real guardianship from mere presence, because available people who are not watching prevent nothing.

3

Willing to intervene

Doing something about what was seen. Directly, or indirectly by summoning someone else.

The bars fill in once you lock in the guess below. Each bar shows the share of the people who made it through the previous stage, not the share of everyone, so the funnel narrows faster than the individual numbers look.

In Reynald's survey of 255 residents, 85 percent said they watched what happened on their street. Of those watchers, what percentage said they would intervene directly rather than call someone else or wait to see how serious it got?

50%

85%

of residents said they monitored street activity, and 15 percent said they paid no attention

80%

of the monitoring residents were willing to intervene in some manner

16%

said they would intervene directly

Among monitoring residents, 41 percent preferred indirect intervention such as calling police, 16 percent would intervene directly, and 23 percent said it would depend on how serious the incident looked. Reynald, D. M., research on guardianship in action, surveying 255 residents. Percentages as reported. Retrieved August 2026.

Read the funnel, not the individual numbers. Guardianship leaks at every stage. Most people are not there. Of the people who are there, some are not looking. Of the people who are looking, most will do something, but only a small fraction will step in themselves. The theory says a capable guardian prevents crime. Reynald's contribution is showing that "capable" is an outcome that has to survive three filters, and that the last filter is the narrowest.

Honesty note. These percentages come from surveys of residents on residential streets in a particular national setting, asking what they would do rather than observing what they did do. Treat the shape of the funnel as the finding and the exact percentages as illustrative. People reliably overestimate their own willingness to intervene when asked in the abstract, which means the real narrowing at stage three is probably sharper than these self-reports suggest.

What predicts whether somebody steps in

Reynald found a consistent set of factors: a felt sense of obligation to the place and the people in it, how the person reads the environment around them, prior training, an assessment of personal risk, whether they have anything to intervene with, and physical confidence. Neighborhood context predicted both monitoring and willingness, with residents of lower-crime, higher-income areas reporting substantially more of both. That last finding should sit uncomfortably. It means guardianship is thinnest exactly where victimization is highest.

Guardians are not the only controller

Ronald Clarke and John Eck extended the triangle by giving each of the three elements a controller. The extension matters because it names people who are standing right there and are not guardians in the original sense, but who have more leverage than anyone else in the room.

Handler

controls the offender

Someone who knows the potential offender well and can exert influence over their behavior. Parents, siblings, teachers, friends, partners. The person most likely to stop a specific individual is usually someone who came with them.

Manager

controls the place

Whoever has responsibility for behavior in that location. A bar owner, a landlord, a bus driver, a teacher, a flight attendant. Someone whose house it is.

Guardian

controls the target

People protecting themselves, their belongings, or those of family, friends, and coworkers, plus public police and private security. The original Cohen and Felson element.

The problem analysis triangle, ASU Center for Problem-Oriented Policing. Retrieved August 2026. A criminal event requires the inner three to converge while all three controllers are weak or absent.

Hold onto the handler. In the next tab, the person with the most power to stop what is happening is not a guardian at all.

Content note. This section works through a scenario in which the risk being discussed is sexual assault. It is here because it is the clearest case in criminology of guardianship failing for reasons that have nothing to do with whether anyone was present. If you would rather not work through this material, tell me and you will get an alternate scenario with no loss of credit. Support is available at any hour through the National Sexual Assault Hotline at 800-656-HOPE (4673) and at online.rainn.org, and on campus through the counseling center and the Title IX office. You are never required to disclose anything about your own experience for a grade in this course.

The same thirty seconds, six ways

It is a house party. A woman who has been drinking heavily has become unsteady and stopped tracking the conversation. A man she met earlier in the evening helps her up, puts an arm around her, and walks her down a hallway toward a bedroom. Roughly six other people see some part of this happen. The scene lasts about thirty seconds.

By the strict letter of routine activity theory, guardianship is abundant here. Six people are available. But availability was only stage one. What determines whether anything happens next is not how many people are in the house. It is how each of them defines what they just saw, and definitions are not evenly distributed. Reveal each person and notice which controller they are and where their definition sends them.

The sociology underneath this

In 1928 W. I. Thomas and Dorothy Swaine Thomas wrote a sentence that has organized this part of sociology ever since: "If men define situations as real, they are real in their consequences." The scene in the hallway has no meaning printed on it. Six people build six meanings out of it, and each of them then acts on the meaning they built, not on the scene. The teammate is not lying when he says nothing was wrong. He is reporting his definition accurately. That is what makes this hard, and it is why "there were people right there" turns out to be such a weak protection.

Thomas, W. I., and Thomas, D. S. (1928). The Child in America: Behavior Problems and Programs. New York: Knopf, page 572.

What people actually report doing

The 2019 AAU Campus Climate Survey asked 181,752 students at 33 colleges and universities about sexual assault and misconduct, including what they did when they witnessed a situation they believed could lead to a sexual assault. Guess first.

Of students who said they witnessed such a situation, what percentage said they directly intervened or interrupted what was happening?

50%

Checked in with the person affected

50.3%

Directly intervened or interrupted

37.0%

Went to a friend of one of them

27.5%

Did nothing, unsure what to do

14.0%

Did nothing, thought it was handled

10.8%

Categories are not mutually exclusive, so they do not sum to 100. AAU aggregate figures across participating institutions, 2019 AAU Campus Climate Survey on Sexual Assault and Misconduct. The same survey found that 13 percent of all students, and 26.4 percent of undergraduate women, had experienced nonconsensual sexual contact by physical force or inability to consent since enrolling. Retrieved August 2026.

Look at the two inaction categories through the lens you just built. Fourteen percent did nothing because they were unsure what to do, which is a stage-three failure of willingness. Nearly eleven percent did nothing because they concluded the person was handling it, which is not a failure of willingness at all. It is a definition. Those students were available, they were monitoring, they were willing in principle, and they decided the situation did not call for anything. That is the teammate in the hallway, and it is the single most common way guardianship fails in a room full of people.

Two findings that cut against each other

A study of 219 real arguments and assaults captured on public CCTV cameras in Amsterdam, Lancaster, and Cape Town found that in 91 percent of them at least one bystander, and usually several, did something to help. Larger groups helped more, not less, which reverses the classic bystander effect that most people were taught. Meanwhile a study of 427 women who reported a sexual assault found that a bystander had an opportunity to intervene in 23 percent of those incidents, and alcohol use by perpetrators, victims, and bystanders was widespread in exactly those settings.

These two findings are not in conflict, and the difference between them is the lesson. The CCTV study captured public street conflicts, where the situation defines itself. Someone is shouting, someone is being shoved, and every bystander reads the same scene the same way. A hallway at a party is the opposite. The scene is ambiguous by construction, most of the people watching have a relationship to one of the two parties, and everyone present has been drinking. Bystanders are not generally passive. They are unreliable precisely when the definition is contested, which is precisely when the harm is most likely.

Philpot, R., Liebst, L. S., Levine, M., Bernasco, W., and Lindegaard, M. R. (2019). American Psychologist, 74(1). Haikalis, M., Leone, R. M., Parrott, D. J., and DiLillo, D. (2018). Sexual assault survivor reports of missed bystander opportunities. Violence Against Women 24(10), 1232 to 1254. Retrieved August 2026.

Does any of it work

The National Institute of Justice runs CrimeSolutions, which grades programs and practices against the research evidence using three labels: Effective, Promising, and Ineffective. Ratings are assigned to specific outcomes rather than to a whole program, which is why the same intervention can land in two different places depending on what you measure. Below are five real findings for interventions that work on one leg of the triangle or another. Commit to an answer before you look. Some of these will not go the way you assume, and the ones that do not are the more useful ones.

All ratings and findings from CrimeSolutions, National Institute of Justice, crimesolutions.ojp.gov, drawing on the practice profiles for hot spots policing, improved street lighting, and bystander education programs, and the program profile for Green Dot. Retrieved August 2026. Ratings are periodically re-reviewed and can change, so check the site before citing any of these in a paper.

Forum: where guardianship broke

Post your initial response and then reply to two classmates. Length and deadlines are in that week's module in the LMS.

Work through the hallway scenario in the previous tab. Name the reading that came closest to your own first instinct, and say honestly why that one. Nobody is graded on which one they picked.
Using Reynald's three stages, identify the specific stage at which guardianship failed in that scene. Availability, monitoring, or willingness. Defend the choice, because the three are genuinely arguable here and your classmates will not all land in the same place.
Bystander education programs are rated Effective at raising bystander confidence, intent to help, and rejection of rape myths, but only Promising for actual helping behavior, and the Green Dot evaluation found no statistically significant effect on sexual violence victimization. Using what you worked through about definitions, offer one explanation for that gap. Do not simply say the programs need to be better.
Then take one step outside the scenario. Identify a routine setting you actually pass through in an ordinary week and name who occupies each of the three controller roles there. Handler, manager, guardian. If one of them is missing, say which.

Replies: respond to two classmates who identified a different failure stage than you did. Ask what would have to be true about the room for their stage to be the right answer.

A word on tone. This forum asks you to analyze a situation in which people did not act. Do that as criminologists. The point of the chapter is that these failures are structured and predictable, not that the people in that hallway were uniquely careless. Anyone can write a post that condemns the bystanders. The harder and better post explains why six reasonable people produced six different readings.