Do Studies Show Gun Control Works? No.

Gun violence in America has fallen to historic levels since the middle of 2010. IncreasedThey have remained constant in recent years higherThere are more laws in the United States than there is in other countries. While there are many laws at both the state and federal level restricting Americans’ guns access, advocates claim that social science research has shown that there may be a few more.simple, Commonsense“Laws can be Significantly reducedThe number of deaths and injuries that firearms are responsible for.

It has seen a huge Research effortWe can go back to decades and determine if gun control measures are effective. 2020 AnalyseRAND Corporation (a nonprofit research group) analyzed 27,900 scientific publications about the effectiveness gun control laws. RAND’s authors identified only 123 (or 0.4%) studies which tested the effects thoroughly from the vast amount of data. Although some 27,777 other studies might have been helpful for discussions that were not empirical, many of them were severely flawed. 

This article examines the importance of the 123 empirically rigorous studies, and the actual conclusions they provide about gun control laws’ efficacy. 

It is not true. RAND considered the 123 most reliable of all the 27,900 studies. However, they had significant statistical flaws such as inadequate controls, excessive parameters and hypotheses, insufficient data disclosures, misspecified models and data. 

These methodological flaws, which aren’t specific to gun control research, are also typical of academic publishing’s response to requests from political partisans seeking scientific evidence.

The social science literature about gun control is largely useless. It also provides endless ammunition for those who claim that studies prove that certain policies would be beneficial. It matters because even though gun laws may not achieve their intended goals, they are still useful. High costs. You can transform otherwise law-abiding citizens in to criminals. 

RAND evaluated 722 different hypotheses regarding the effect of gun control policies on “statistical significance” in the 123 papers. The majority of statistically significant results accepted by peer-reviewed journals are those that have a 1-in-20 chance or lower to be due to random chance. So if researchers run 100 tests on the relationship between two things that obviously have no connection to each other at all—say, milk consumption and car crashes—by pure chance, they can be expected to get five statistically significant results that are entirely coincidental, such as that milk drinkers get into more accidents. 

In terms of the gun control studies deemed rigorous by RAND, this means that even if there were no relationship between gun laws and violence—much like the relationship between drinking milk and getting into car accidents—we’d nevertheless expect about five percent of the studies’ 722 tests, or 36 results, to show that gun regulations had a significant impact. However, the papers actually found only positive results for 18 combinations of gun controls measures and outcomes (such a waiting period and gun suicides). It’s still not the same as the 36 expected false negatives. However, some combinations received support from multiple studies. However, it is in line with what one would expect to see if there were no gun control measures.

Concerning is the finding that the only negative outcome that the researchers found was that gun control led to an increase of violence and bad outcomes. Due to the sheer number of researches on the subject, it is highly probable that there would have been more. These results could suggest that guns control measures may not be working as they should. 

Many inconclusive researches are never published and many of the inconclusive findings are left out of published studies. This is why the rare pro-gun control results and near total absence of antigun-control result strongly suggest that we don’t know much about gun control.

We don’t have any good evidence of the effectiveness or harm caused by gun control. These are important reasons and will not be addressed in the near future. Researchers cannot discern the signals from the noise in data about gun violence. 

It is common to look at homicides in different states over the past years. It can take months, or even years for such legislation to be implemented fully and enforced. Modern gun control measures affect only a few gun sales and new guns are just a fraction of the total firearms. Few of those who are prevented from buying guns under the law will actually kill anybody. Many of those who do intend to murder someone would still kill them with another weapon or get a gun elsewhere.

Accordingly, the best estimate of the impact of laws on gun violence in their first year is a decrease of just a fraction percent. However, gun homicides in a given state increase by six percent per year if there are no legislative changes. This is based upon FBI Uniform Crime Reporting (UCR) dataAll the way back to 1990. This type of study, which is a before-and-after one can only detect effects around 1990 according to statistics. It is three times largerThese studies cannot be used to determine the effect of gun laws on the annual gun crime rate. This is at most an order of magnitude greater than the likely effects of legislation.

You can overcome these limitations by using what statisticians refer to as “controls”. These are mathematical tools that enable them to compare different things and to isolate the effect that they are looking for. Gun control studies compare violence rates between two states with stricter or less restrictive gun laws. They try to eliminate any differences among the states, except those that are more liberal. 

Researchers also have another option: comparing the violence rate in one state with that of other states. It is believed that factors that affect homicide rate other than legislative changes will have the same effect on both the state and national figures. The comparison of changes in state and national homicide rates supposedly allows for the control of other factors that can affect violence such as an increase in crime or a decline in shootings.

This is because national violence levels don’t always correlate with state violence rates. According to the FBI’s UCR data there is only a 0.4% correlation between annual state changes and national rates, if legislation has not changed. This means that the differences between state rates and national rates are more volatile than changes in state rates. This control is not a filter, but adds noise to your study. If you want to compare the state with similar states or nearby ones, this is also a problem. Unfortunately, we don’t have any controls on state homicide rates.

Sometimes, researchers in gun control are able to find enough evidence of an effect for measurement. Get togetherConsider dozens, if not hundreds, of state legislations and look at the effect on homicide. However, states with strict gun control regulations differ from those with less. These states are generally more economically oriented, liberaler and urbane, with lower rates of murder and suicide. There are cultural differences too great and too many uncertainties in data for us to know what would happen if Greenwich, Connecticut laws were applied in Festus Missouri.

Researchers try to avoid the pitfalls of before-and-after studies or inter-state comparisons by using longer periods of time—say, by studying the change in gun homicide rates in the 10 years after legislation was enacted compared to the 10 years before. You might be able to discern a difference from just one year of noise but not 20 years.

A second limitation of gun control studies’ utility is the incompletion and lackluster reliability of their underlying data. Estimates of the number of working firearms in the U.S. differ by a factor of two—from approximately 250 millionsYou can find more information here 600,000,000—and most uses of firearms go unreported unless someone is killed or injured. Although we have information about gun sales from licensed dealers and gun crime, not all crimes are reported by the FBI. Many non-crimes also get reported.

Even if there was statistical proof that gun violence has declined since the passing of the gun control legislation, other factors that took place around that time could still be a plausible explanation.

More research is needed on violence and crime, not more on legislation governing gun control. A better understanding of crime and violence can enable precise measurement and experimentation to identify changes that are too small for statistical analysis. 

Consider the impact of smoking on cancer. The claims that smoking causes cancer were based on aggregate statistics for years. However, the research was flawed. Eventually, however, medical researchers—not statisticians or policy analysts—figured out how cigarette smoking affected cells in the lungs, and how that developed into cancer. There were certain types of lung cancers and other problems that could only be found in smokers. With this better understanding it became possible to provide overwhelming statistical support for every link in the chain.

As far as studies on gun violence go, let’s suppose that psychological researchers will be able to demonstrate the impact of being abused in childhood on the likelihood of someone committing a gun crime. This would permit more focused and precise studies to examine the effects of gun control legislation. Instead of trying to compare large numbers of people, it is possible for researchers to focus only on those with high likelihood of gun violence.

The only way to understand the effects of gun control is to learn more about the reasons people commit suicide and the impact of guns on rape, assault, and other criminal acts.

This isn’t just about gun control. Nearly every similar study into the impacts of legislation on specific individuals suffers from this problem. There is too many variables to be able to analyze. Scientists are unable to provide the statistical support that political partisans need. Researchers flood into favored areas because they have lots of funding and are interested in the results. They peer-review each other’s papers and don’t apply the kind of rigorous analysis required to make policy decisions.

It doesn’t mean gun control legislation cannot be used. However, without credible scientific evidence, the belief in additional gun control laws’ efficacy is a matter of faith and not reason.

RAND approved 123 studies, but the ones that received the greatest media attention or legislative attention don’t belong among them. These were the most difficult studies, with claims too vague, weak and insufficient to please partisans or sensationalist media. The worst and most bizarre claims were made by these studies. Check out the headlines.

One prominent Study, which was touted from the debate stage by Sen. Cory Booker (D–N.J.) Sen. Cory Booker (D-N.J.) made the astonishing claim, while running for the presidency in 2016, that Connecticut had a 40% reduction in gun deaths due to a handgun purchase permit. Although the murder rate in Connecticut fell quickly after the law was adopted, it’s true that gun deaths dropped across the nation. The study’s 40 percent claim is the actual murder rate in Connecticut compared to something the researchers call “synthetic Connecticut,” which they constructed for the purpose of their study—a combination of mostly Rhode Island, but also Maryland, California, Nevada, and New Hampshire. 

It turns out that the entire effect claimed by the authors (the 40% reduction) was caused by Rhode Island’s temporary spike of 20 additional murders in 1999-2003. Synthetic Connecticut, however, was greater than 72 percent Rhode Island.

The decline that the authors observed didn’t continue, even when compared with synthetic Connecticut. The law was still in force, but by 2006 the real Connecticut gun murder rate had exceeded that of synthetic Connecticut. It then continued rising to 46 percent. Despite publication in 2015, the authors decided to ignore 2006 data. 

The field is known for making strong claims that are based on complicated models and unreliable data. This is a typical example of this study. To make matters worse, many researchers cherry-pick outcome measures, locations, or time periods to achieve their preferred results. 

Take, for example, the study that looked at whether large-capacity magazines and assault weapons were banned. These studies are frequently paired together. Researchers DefinitionBasic terms such as “assault weapons”, “mass shootings” are defined differently. The data they collect is restricted by the time and place of events, as well as other factors such as whether an event was classified as a terrorist act or gang violence, thus not being considered a mass shooting.

This research has more data problems than any other on gun violence research. Mass shootings can be Very rareComparable to most other types of gun violence Don’t includeassault weapons. Although estimates of mass killings using assault weapons vary depending on their definitions, they are generally high. ConsiderOnly 1% of all gun-related homicides occur in this country. 

The U.S. Federal banOn assault weapons and large-capacity magazines. This was the topic of many studiesThe law was in effect for ten consecutive years (from 1994 to 2004). Its effectiveness was questioned by many and varied opinions. Many social factors, including the length of the law’s implementation, caused criminal rates to fluctuate widely. It was impossible to predict what would happen. Make useful and insightful conclusionsConcerning the effects of the ban, especially on something rare Mass shootings. However, researchers can easily find weak evidence to support their conclusions if there is too much noise.

Furthermore, laws defining assault weapons and key components of laws are different in countries that have bans. This is in addition to the problems that data can present when comparing populations over different time periods. ComparationsThe distinctions between nations and states are nearly meaningless.

RAND Corporation Another meta-analysisUpdated 2020. Inconclusive evidence found that assault weapons and large capacity magazines bans had no effect on mass shootings, or on violent crime.

What about the simple question: Is owning guns more or less secure? The following is an example: StudySince its publication in the newspaper, it has been a constant headline. New England Journal of MedicineIn 1993, it was concluded that guns in the home were not meant to confer protection. 

You can find them here Problems of major importanceWith this research. The researchers found that having a gun in your home can increase the risk of someone being shot, however, nearly all of the murders included in the analysis weren’t committed using a firearm. The authors did not establish whether the gun used in killing gun owners was their own or that of another individual. 

Research on this topic can be so inconsistent because of individual differences that social science research cannot easily control. While a gun expert living in a high-crime area may be safer having a gun in their possession, an alcoholic who lives in low-crime neighborhoods and keeps guns hidden in the kids’ clothes is likely to feel less secure. The people want to know if guns make them safer. But social science can’t provide that answer. 

Population averages can be useful when one rule has to be applied to everyone—for example, estimating how many lives would be saved by a pollution control regulation, or how many dental cavities are prevented by fluoridating the water supply. Personal safety and guns are not about making the average gun owner more secure, it is about which guns make people safer.

Building on the sand is not building a gun control position based on scientific evidence. No empirical data is available on this subject and there are no studies that show a difference in the results. It is also false to claim there are simple, commonsense laws that could significantly reduce gun violence if we have the political will.

Complex issues like these require a rigorous scientific investigation in order to reach any useful conclusions. They are more dependent on individuals and wide social and cultural factors that any regulations. It is wrong to look for laws that will sweep away innocent victims and do more harm than good. All this with the supposed backing of science that cannot possibly know what it means.

Justin Monticello produced and edited the book. Monticello and Aaron Brown wrote the book. Isaac Reese designed the graphic. Ian Keyser audio production.

Music: Aerial Cliff by Michele Nobler, Land of the Lion by C.K. Martin, The Working of the Plan by Cooper Cannell, Thoughts by ANBR, The Inner Bird: Flight by Sivan Talmor and Yehezkel Raz, and Running by Tristan Barton.

Photos: Hollandse-Hoogte/ZUMA Press/Newscom; Robin Rayne/ZUMAPRESS/Newscom; Ted Soqui/Sipa USA/Newscom; YES Market Media/Yaroslav Sabitov/YES Market Medi/Newscom; Chuck Liddy/TNS/Newscom; YES Market Media/Yaroslav Sabitov/YES Market Medi/Newscom; Brett Coomer/Rapport Press/Newscom; Martha Asencio-Rhine/ZUMAPRESS/Newscom; Jebb Harris/ZUMA Press/Newscom; John Gastaldo/ZUMA Press/Newscom; Greg Smith/ZUMA Press/Newscom; Richard Ellis/ZUMA Press/Newscom; Matthew McDermott/Polaris/Newscom; KEVIN DIETSCH/UPI/Newscom; Bill Clark/CQ Roll Call/Newscom; Michael Brochstein/ZUMAPRESS/Newscom; Sandy Macys / UPI Photo Service/Newscom; E. Jason Wambsgans/TNS/Newscom; Eye Ubiquitous/Newscom; Matthew McDermott/Polaris/Newscom