Policing and the law: legislation, policy and discretion

Professor Desjardins starts week seven with a story from his first year on patrol. At 2 a.m. he stopped a car weaving across the center line. The driver was a nurse coming off a sixteen-hour shift, exhausted but sober. The law allowed him to write a ticket for careless driving; policy required him to check for impairment, which he did; nothing told him what to do next. He followed her home with his lights on and told her to call a taxi next time. Was that the right decision? Amara says yes. Marco is not sure: what if she had crashed the next night? Linh asks whether a young man in an old car would have received the same kindness. Professor Desjardins says that all three reactions go to the heart of today's topic. Police officers act under law and policy, but in the space between them is discretion, and that space is where much of policing actually happens.

Where police powers come from

Police officers can do only what the law allows. Their powers come from several sources arranged in a hierarchy. At the top is the constitution, which limits what any government official, including police, may do. The Fourth, Fifth, Sixth and Fourteenth Amendments limit searches, seizures, interrogation and the treatment of the accused. Below the constitution are statutes. State criminal codes, such as the Ohio Revised Code, define most crimes and set out powers of arrest, while state traffic, liquor and mental health laws give police further powers. City ordinances add local rules on noise, parking and public order that police enforce. Then come regulations made under those statutes, such as provincial or state standards for training and equipment.

Court decisions interpreting the Constitution set many of the rules officers follow day to day. Terry v. Ohio (1968) allows a brief stop when an officer has reasonable suspicion that a person is involved in crime, and a pat-down for weapons when the officer reasonably believes the person is armed and dangerous. Miranda v. Arizona (1966) requires warnings before custodial interrogation. Graham v. Connor (1989) holds that force is judged by whether it was objectively reasonable from the perspective of a reasonable officer on the scene, and Tennessee v. Garner (1985) limits deadly force against a fleeing suspect to cases where the officer has probable cause to believe the suspect poses a significant threat of death or serious injury. Finally, each police service issues its own policies and procedures, which can be stricter than the law but never looser. A service may forbid vehicle pursuits for minor offenses, for example, even though the law allows them.

Powers of arrest and the use of force

Officers may arrest without a warrant when they have probable cause to believe a person has committed a felony, and generally for misdemeanors committed in their presence, with exceptions in many states for domestic violence and some other offenses. Probable cause means facts that would lead a reasonable person to believe a crime was committed and the person committed it. Many minor offenses are handled by citation rather than arrest; Ohio law requires officers to issue a citation instead of arresting for most minor misdemeanors. The use of force is governed by the Graham v. Connor standard, by state law and by agency policy, which increasingly requires de-escalation where feasible and a duty to intervene when another officer uses excessive force. You will study police powers in depth in a later course; for now, notice that each power comes with conditions, and that the conditions often require judgment.

What discretion is

The legal scholar Kenneth Culp Davis defined discretion this way: a public officer has discretion whenever the effective limits on his or her power leave the officer free to make a choice among possible courses of action or inaction. Police discretion operates at every level. A chief decides to make impaired driving a priority. A sergeant decides to send two cars to a call rather than one. A patrol officer decides whether to stop a car, whether to search, whether to arrest or caution, whether to write a ticket for 20 over the limit or 25 over. Discretion includes the choice not to act, which is invisible in most statistics because nothing is recorded.

Discretion exists for practical and principled reasons. Practically, there are far more offenses than police could ever enforce; full enforcement of every law would swamp the courts and turn the public against police. The law is also written in general terms and cannot anticipate every situation. In principle, justice often requires judgment. A rigid rule that treats a hungry teenager who steals a sandwich the same as a professional shoplifter produces outcomes most people consider unjust. Researchers began to take police discretion seriously in the 1950s, when the American Bar Foundation's survey of criminal justice found that officers routinely chose not to arrest in situations where the law allowed it, contrary to the official view that police simply enforced the law.

What shapes discretionary decisions

Research on how officers use discretion has found consistent patterns. The seriousness of the offense matters most: officers rarely ignore serious violence. The strength of the evidence matters, as does whether there is a complainant who wants action. In a famous set of observational studies in the 1960s, Donald Black and Albert Reiss found that officers' decisions to arrest in minor disputes depended heavily on whether the complainant wanted an arrest. The relationship between the parties matters; officers have historically been less likely to arrest when the parties are family members, which is why many jurisdictions changed their policies on domestic violence. The suspect's attitude matters: people who are hostile or disrespectful are more likely to be arrested, although later research showed that part of this effect reflects hostile people also committing more serious offenses or resisting.

Characteristics of the people involved, such as race, age, gender and class, should not matter, but research in both countries shows that they sometimes do. Some disparities reflect where police are deployed and the kinds of calls they receive; some reflect individual or institutional bias. Distinguishing the two is difficult, which is why services increasingly collect data on stops, searches and outcomes and analyze it carefully. Many states, including Ohio for some agencies, require or encourage collection of data on traffic stops by race, and several cities publish it. Organizational factors also shape discretion: the service's priorities, supervisors' expectations, performance measures, training and peer culture. An officer in a unit where everyone writes ten tickets a shift will make different choices from one in a unit focused on problem-solving.

Structuring discretion

Because unchecked discretion can lead to unfairness, police services structure it. Some decisions are taken out of officers' hands. Many states, after research in the 1980s suggested arrest could deter repeat domestic violence, adopted mandatory or preferred arrest laws for domestic violence, and Ohio law makes arrest the preferred response when an officer has reasonable grounds to believe an act of domestic violence has occurred. Later research showed mixed effects of arrest on repeat violence, a reminder that rules need evaluation. Juvenile diversion policies encourage officers to refer young people to programs rather than charging them for minor offenses. Use-of-force policies set out when different options may be used and require reporting and review. Pursuit policies weigh the danger of a chase against the seriousness of the offense. Body-worn cameras and supervisory review make discretionary decisions more visible.

Discretion cuts both ways in American law. In Castle Rock v. Gonzales (2005), the Supreme Court held that a woman whose estranged husband killed their three children had no constitutional right to have police enforce her restraining order, despite a state law that seemed to require it, because police traditionally have discretion over enforcement. Critics argued that the decision left victims without a remedy when police fail to act. Officers who use discretion to protect a friend or colleague, however, can face criminal charges for obstruction or misconduct.

Discretion and fairness: reading the data

Data can reveal patterns in discretion, but it must be read with care. Suppose Port Ardent's traffic unit notices that young drivers who are stopped receive tickets more often than older drivers. Is the unit treating young drivers more harshly? Before reaching a conclusion, an analyst must ask whether the stops differ in ways that legitimately affect the outcome, such as how far over the limit the drivers were going. The code below compares ticket rates overall and within speed bands.

# Ticket rates after speeding stops: overall and within speed bands (illustrative data)
stops = {   # band: (young stops, young tickets, older stops, older tickets)
    "under 20 over": (300, 90, 2000, 600),
    "20 to 39 over": (500, 300, 1500, 900),
    "40 or more over": (200, 190, 200, 190),
}
ys = sum(v[0] for v in stops.values()); yt = sum(v[1] for v in stops.values())
os_ = sum(v[2] for v in stops.values()); ot = sum(v[3] for v in stops.values())
print(f"overall: young {yt / ys:.1%} ticketed, older {ot / os_:.1%}, ratio {(yt / ys) / (ot / os_):.2f}")
for band, (a, b, c, d) in stops.items():
    print(f"{band:16s} young {b / a:5.1%}  older {d / c:5.1%}")
print(f"share of young drivers' stops at 20 or more over: {(500 + 200) / ys:.0%}; older: {(1500 + 200) / os_:.0%}")

overall: young 58.0% ticketed, older 45.7%, ratio 1.27
under 20 over    young 30.0%  older 30.0%
20 to 39 over    young 60.0%  older 60.0%
40 or more over  young 95.0%  older 95.0%
share of young drivers' stops at 20 or more over: 70%; older: 46%

Overall, young drivers are ticketed at 1.27 times the rate of older drivers, but within each speed band the rates are identical. The overall difference arises because a larger share of young drivers were stopped for higher speeds. This pattern, known as Simpson's paradox, shows why raw disparities do not prove unfair treatment. The reverse is also true: equal overall rates can hide unequal treatment within groups. And the analysis leaves another question open: are young drivers stopped more often in the first place for the same driving? Good analysis of discretion asks about every decision point, which Linh, the aspiring analyst, finds fascinating.