Recently, a new study was published that claimed three types of gun control laws could reduce firearm-related fatalities by up to 10 percent. Now, for the anti-gunners, this is big news. This is something they’re likely to jump up and down about and publicize until the cows come home. After all, many of these regulations are things they’ve been pushing for lawmakers to adopt for some time now.

However, there are problems with the study. Big problems.

In fact, they’re big enough that no one should act on anything just yet.

U.S. gun laws vary considerably by state. Some, like Kansas, allow citizens to carry firearms in public and make it legal for gun owners to shoot an assailant in self-defense in some situations (known as the stand your ground doctrine). Others, like California, are more restrictive, limiting not only who is allowed to carry guns in public, but also access to firearms in the home by requiring safety devices such as trigger locks or gun safes.

This patchwork—combined with limited funding for reasearch—has made it hard for scientists to predict the effects of gun laws on gun deaths, says Terry Schell, a senior behavioral scientist at the nonprofit RAND Corporation, which aims to improve public policy through research and analysis.

To limit these problems, Schell and colleagues focused on just three kinds of laws and one outcome: gun deaths per capita. To understand how laws affect death rates, they screened hundreds of existing and novel statistical approaches, finally zeroing in on a model that reduces statistical noise by paying special attention to how different variables affect deaths year by year, rather than averaged over long periods of time.

The researchers counted the number of gun deaths from all 50 states for each year from 1980 to 2016. They then examined each instance of a new law limiting or allowing right to carry, stand your ground, or child access, state by state, through 2013. Finally, they compared that with mortality data from the Centers for Disease Control and Prevention (CDC) for the next 6 years.

So what’s the big problem? They’re not accounting for any other changes that may or may not have led to these differences. There’s absolutely no way to control for these factors.

Anyone who has lived in multiple states can tell you that while there are some things that are fairly universal across all American cities, there are also profound differences. That includes a whole litany of laws, including gun control laws.

As it stands, it’s impossible for any study to control for these other factors when looking at whether a bit of gun control legislation works. For example, a reduction in joblessness can reduce violent crime all on its own, all without the first bit of gun control being passed. Yet if gun control is passed around the same time, studies like this will chalk up that reduction to the gun control law.

That’s true even if there were evidence to suggest that the reduction might have been greater had the gun control law not gone into place.

At the end of the day, you just can’t compare states. As the above-linked story notes, while six of the top ten safest states have gun control policies like the ones espoused by the study, four of those ten don’t. Further, some of the states with the strictest gun control also have some of the most violent communities, such as Illinois and Maryland.

The factors that contribute to firearm fatalities are complicated and no study actually tries to look at other potential causes for any decrease in violent crime. They, instead, focus on guns.

And they wonder why we think these studies are biased and don’t want taxpayer money going toward such things.