The standard safety canon — don’t walk alone at night, park under streetlights, trust your gut, carry pepper spray — gets repeated everywhere from corporate trainings to parental advice. Most of it is well-intentioned, and a lot of it is partially true. But it’s calibrated to a stranger-danger model of violence that doesn’t match the actual statistics on where, when, and from whom most people experience harm. The result is advice that misses the actual risks while creating a false sense of competence.
This isn’t a critique of caution. It’s a case for caution based on data rather than folklore.
The “stranger lurking” frame is mostly wrong
Most violence against adults is committed by someone the victim knows. Bureau of Justice Statistics data consistently shows that more than half of violent crimes against women, and a substantial share against men, involve current or former partners, family members, friends, or acquaintances. Sexual violence skews even more sharply toward known offenders — typically over 80%.
This shifts the risk landscape significantly. Generic safety advice tends to focus on managing public spaces, which are statistically lower-risk for most people. The under-discussed risk is in homes, relationships, workplaces, and social contexts where the people who do harm have legitimate access. Telling someone to walk to their car with keys between their fingers does nothing to address the partner who controls their finances or the colleague who follows them home.
“Trust your gut” cuts both ways
The instinct-based advice has a real basis — there’s good evidence that people pick up on threat cues subconsciously and that ignoring those cues correlates with worse outcomes. But “trust your gut” also produces enormous false-positive rates, often patterned along race and class lines. It can rationalize avoidance of stigmatized but objectively safe people while missing risks from familiar, “safe-seeming” individuals who are actually higher-risk.
The more useful version of the advice is calibrated. Pay attention to specific behavioral signals — boundary-pushing, isolation tactics, anger over minor refusals, control over your time or contacts. These predict harm far better than demographic gut reactions do. Most people who hurt others give specific, observable signals before they escalate. Training the eye to those signals is a higher-leverage skill than scanning for “shady-looking strangers.”
Self-defense tools are oversold
Pepper spray, rape whistles, and personal alarms all have a place, but their effectiveness in the situations where most violence actually happens is limited. They presume a discrete confrontation with a stranger in a public place, with enough warning to deploy a device. That scenario is real but uncommon relative to the violence that occurs from people you’ve already let close.
Tools that scale better: financial autonomy, social networks that notice and respond, knowledge of your local resources and laws, and concrete safety planning if you’re in a relationship with warning signs. These don’t fit on a wallet card, which is part of why they don’t show up in the standard advice list.
The takeaway
Safety advice should match the threat model. The real risks for most adults aren’t ambush in a parking lot but escalation in places where they feel comfortable. Plan accordingly.
Leave a Reply