State prison admissions ran 3% higher in years when district attorneys stood for election

District attorneys are elected officials who win their jobs in county contests ^[17], and that election calendar gives a study like this its comparison. If prosecutors behave differently in the year their jobs are on the ballot, the difference should show up in the cases they send to prison. Okafor took the names of more than 4,200 district attorneys from the National Directory of Prosecuting Attorneys ^[6] and matched them to offender-level admissions records from the National Corrections Reporting Program, which cover nearly 40 states ^[7]. Election results came from state and county sources. Prosecutor characteristics came partly from the DIME database ^[9].

Prosecutors are a sensible place to look. They decide whether to charge, what to charge, whether to offer a plea, what sentence to recommend and whether to drop a case ^[10].

The size of the effect depends on the baseline. Against all non-election years, admissions ran about 3% higher. Against the year just before the election, they ran 4.3% higher ^[4]. Months sentenced followed the same pattern, about 3% and 4.0% ^[5]. All four estimates were statistically significant. Switching the admissions baseline moves the estimate by 1.3 percentage points ^[3].

Spread over the 21 years from 1986 through 2006, the 63,000 extra admissions come to roughly 3,000 a year ^[1]. Taken at face value, 63,000 as a 3% increase implies a base of about 2.1 million admissions in the election years the estimate covers ^[2]. Each of those admissions is a person sent to state prison. The bump shows up only in election years, and it sits inside what the phys.org account calls the steepest rise in incarceration in U.S. history ^[16].

That size limits what the finding can carry. The same account says prisoner numbers soared in the 1980s and 1990s regardless of the crime rate ^[11], but that is the article's background. It is not a result of this study. A 3% effect confined to election years cannot by itself account for growth the article describes as soaring. The thing this doesn't tell you, at least in the phys.org account, is how the study handled crime trends inside election years. The scope is also narrower than "incarceration": Okafor counted admissions to state prisons, not jails, and admissions, not total prison populations ^[8].

Okafor attributes the pattern to voters. From the War on Drugs through the 1994 Crime Bill, he said, public sentiment on crime was more punitive, and DAs may have felt incentives to be more aggressive ^[15]. "The study provides evidence consistent with prosecutors having responded to perceived voter preferences," he said ^[12]. The standard he holds offices to is simple: "an office should charge and make plea offers the same way in an election year as in any other year" ^[13].

I think the design supports the narrow claim. In the offices and states the data covers, election years came with measurably more admissions and longer sentences. The evidence does not show that elections drove the era's prison growth. Okafor's own conclusion looks beyond the courthouse. "That is why efforts to change public attitudes toward punishment deserve as much attention as efforts to change policies or officeholders," he said ^[14].