Minimum wage policy is a fraught issue. People involved in the public debate — writers, business leaders, politicians and think-tankers — often defend their positions with a passion and commitment reserved for religious disputes. That makes it challenging to bring data to bear on the conversation. Nevertheless, as the weight of evidence piles up, it’s getting harder and harder to ignore.

The minimum wage is one of the few issues in economics where really good evidence does exist. The wage floor is typically set at the city or state level, so it’s possible to compare places that raise their minimum wages to neighboring places that choose not to.

In 1993, economists David Card and Alan Krueger surprised the world when they studied a New Jersey minimum wage hike. Basic undergraduate economic theory would predict that a higher minimum wage would put people out of work, but Card and Krueger found the opposite — employment increased after the hike, relative to neighboring Pennsylvania. And stores that were forced to raise their wages more actually hired more people.

Card and Krueger’s results were initially greeted with extreme skepticism. But as similar studies trickled in, the finding turned out not to be much of an outlier — minimum wages just don’t look as if they have much of a negative impact on employment in the short run. Soon, economists were looking around for other theories to replace the simplistic one taught to Econ 101 students.

Meanwhile, anyone who wagers on small or zero effects from minimum wage hikes continues to win those bets. Studies of Seattle’s recent minimum wage increase find, once again, little change in employment levels. Results such as these are no longer surprising. But if you have a friend who’s a dedicated opponent of minimum wages, you might still be able to win some money betting on the effects of the next city’s hike.

Of course, it’s important to realize the limitations of these studies. First, they focus on the short term — maybe one or two years after the change. Higher minimum wages might damage the economy in a more subtle, long-term fashion by slowing down growth in low-wage industries. But since these long-term effects are hard to disentangle from conditions in the broader economy, it’s hard to know how much this happens. Second, high minimum wages might make a city or state more vulnerable to a big recession — that’s something that’s impossible to know until disaster strikes. And crucially, these studies don’t tell us how a high minimum wage could go before it starts doing damage; a minimum of $1,000 per hour, for example, would certainly be a bad idea.

Recently, a number of studies have been coming out that look at the minimum wage from different angles. These are useful and good, since they can help economists understand exactly how a minimum wage works. But sadly, the new studies are often seized on by minimum wage partisans as evidence that their positions are right. This is jumping the gun.

For example, a recent study by Dara Dee Luca of Mathematica Policy Research and Michael Luca of Harvard Business School is making the rounds. They found that minimum wage increases in the San Francisco Bay Area tended to make restaurants with bad Yelp reviews more likely to close down, while slowing the rate at which new restaurants started up.

Luca and Luca’s study is being hailed by minimum wage proponents as proof that wage increases improve quality by weeding out bad restaurants. But this doesn’t tell us whether the restaurants that remain are raising their prices — in fact, the authors find little correlation between menu prices and the likelihood of exit. So we don’t know if restaurant-goers are getting more bang for their buck.

Meanwhile, minimum wage opponents are using the study to say minimum wages are bad for business. But, as the authors note, there’s no way to tell whether existing restaurants or new restaurants expand to compensate. The number of restaurants per capita, they find, falls only very slightly.

A second study, by John Horton of New York University’s Stern School of Business, is also being cited by both sides of the debate. Horton did an experiment with an online labor market, assigning different employers different minimum wages. The companies that were forced to pay higher wages hired about the same number of workers but for fewer hours. But the workers they hired tended to be more productive ones, so overall they got about the same amount of work done.

Minimum wage opponents claim that the reduction in hours represents a negative effect, while proponents argue that the productivity increase means minimum wages either make employees work harder or lures more qualified people into the workforce. In fact, either or both of these could be true. Without observing what happens when all companies have to pay a higher minimum wage, it’s impossible to know the aggregate effect across the economy. Horton’s study offers valuable insights into how companies try to respond to higher minimum wages, but as the author notes it doesn’t tell us much about what happens when they all try to do the same thing at once.

So when you read about minimum wage studies and you see the two sides of the debate clash in the media, it’s important to remember the limits of empirical research. We know that moderate minimum wage hikes rarely have big immediate effects on employment. We know minimum wages tends to raise the average quality of restaurants and workers, though we don’t know exactly why. But just don’t expect academics’ uncertainty to make its way into the public debate on this contentious issue.

Noah Smith is a Bloomberg View columnist. He was an assistant professor of finance at Stony Brook University in New York, and he blogs at Noahpinion.