From lesson 1, readable now without an account: From reports to a defined prevention problem
Imagine a transit facility receiving three messages in one week: “There are thefts everywhere here,” “I no longer feel safe by the cycle parking,” and “My child's bicycle disappeared on Tuesday evening.” These messages matter, but they supply different kinds of information. The first is a broad judgment needing a defined scope and evidence. The second reports fear, which matters for design and evaluation, but is not an incident count. The third describes an event that can be checked. On its own, it does not establish a recurring problem or identify who took the bicycle. A prevention specialist separates these layers before deciding what to measure or change.
An event is a particular occurrence with a time, place, behavior, or harm, recorded in a facility log, official report, or account. A recurring problem is a pattern of sufficiently similar events and harms that warrants an organized prevention response. An indicator is a partial measure, such as the number of eligible reports in a month. It is not the pattern itself: reporting may rise when residents gain confidence in a reporting channel. A witness's use of the word “theft” is not a final legal finding. Legal definitions and procedures differ by jurisdiction; this program teaches prevention planning within the specialist's authority.
Build a problem card by naming the observed behavior before describing a group of people. Record the place, period, unit of count, affected people, source of each observation, and remaining unknowns. “The neighborhood is unsafe” gives no behavior to track, intervention site, or baseline. “Recorded bicycle losses near the station's east entrance between 18:00 and 22:00 over four weeks” is more useful. It remains incomplete until we know how duplicates were removed and how many bicycles were parked. A tight definition protects the decision: changing entrance lighting should not be evaluated against an unrelated neighborhood wide count.
In our fictional facility file there are 18 messages over four weeks. Six repeat earlier messages. Three describe fear without property loss, two describe bicycle scratches, and seven concern bicycle losses near the east entrance. The team checks the seven loss messages and finds six independent events because two accounts describe the same bicycle on the same day. The current count is six recorded loss events, not 18 confirmed crimes. Four of the six occurred between 18:00 and 22:00. That alone does not prove evenings are riskier: use of the entrance and the number of parked bicycles may differ by time. Obtain an exposure denominator before calculating or comparing rates.
A defensible provisional card reads: “The facility log records six independent bicycle loss events near the east entrance over four weeks; four occurred in the evening. Legal classification, the number of bicycles exposed, and changes in reporting remain unverified. The team will examine storage and entrance management and improve reporting without identifying a suspect.” Keep three fear reports outside the event count. They may support a separate measure of perceived safety if collected without exposing respondents. An account of an ongoing threat or immediate danger leaves routine pattern analysis and follows the facility's local emergency referral process.
Why write down unknowns? They can reverse the intervention choice. If losses involve bicycles left without secure storage, access to storage may be worth testing. If locked bicycles are lost inside an area controlled by the facility, inspect place management. If messages rose after a reporting campaign, the message series alone cannot demonstrate an increase in incidents. These are testable hypotheses, not blame assigned to people who experienced harm and not proof of a single cause. Problem oriented work moves from scanning to analysis, response, and assessment. Community participation helps set priorities, but it does not authorize a specialist to investigate individuals or publish personal accounts.
Before using the card, ask whether another person could reproduce the incident count from its inclusion rules, whether the place and period connect to an action the organization can take, and whether observed facts have been separated from hypotheses and referral needs. If any answer is no, revise the rule or collect the missing information. The output is a small working file: a behavioral definition, an account to event conversion table, a provisional indicator with limits, and a short list of data needs for the next lesson.