In 2013, two users of the popular online message board Reddit began tracking mass shootings in the United States. The effort began at the Guns Are Cool subreddit, an ironically named group of gun-control proponents who share articles about shootings in the United States.

Brock Weller, a cook in Oregon who was involved in the early evolution of the tracker, told the online news organization the Trace earlier this year that he became interested in the subject of gun violence after he was held at gunpoint by a man wielding a shotgun in 2008.

Since then, a crowd-sourced mass shooting tracker has become a near-daily source of information about shootings in the United States. It is a list of all shooting incidents in which four or more people are killed or injured by gunfire, culled mainly from media reports.

Reddit users submit mass shooting incidents on the subreddit, which then feeds the standalone tracker website.

As the nation has endured several years of prominent shootings, most recently in San Bernardino, California, on Wednesday, the tracker has become a source of data for stories in media across the country. But it also raised a question about what we define as a mass shooting in the United States - particularly how many people must be harmed or killed - and the differences between regular shootings, mass shootings, mass killings and massacres.

Given that the numbers on the mass shooting tracker come from a mostly anonymous group of pro-gun-control Reddit users, some have asked whether their numbers are biased. The tracker's contributors say that they have been careful when documenting their sources. Each incident in the database is accompanied by at least one link to a local media story reporting on the details of the shooting. They invite users of the data to check their work and report any problems they find.

But there is no question that it uses a broader definition than what has been previously understood. The government has never defined "mass shooting" as a standalone category.

The FBI used to consider someone who killed four or more people during one event a "mass murderer," regardless of weapons used. But starting in 2013, federal statutes defined "mass killing" as three or more people killed, regardless of weapons.

But, according to the tracker's users, that definition has a bit of a problem. It includes non-gun killings, for instance, and it excludes cases in which a lot of people are shot but few of them die.

Earlier this year, a gunman killed two people and wounded nine others during a shooting at a theater in Lafayette, Louisiana. Because only two people died, not including the gunman, that incident would not meet the federal definition of a "mass killing" - even though it garnered widespread media attention.

The Redditors who started the tracker deliberately wanted to use a broader definition that would capture instances similar to the Lafayette shooting as well as other multiple-casualty events. So they settled on the definition of four or more people killed or injured.

Some observers, such as independent researcher and gun-rights proponent John Lott, object to the tracker's methodology because it counts "a lot of shootings where people aren't killed, criminal gangs are involved, and cases take place out of public view."

He argues that gang violence and domestic violence spring from different causes than instances in the Colorado Springs or San Bernardino shootings and that they should be considered separately.

"The causes and solutions to drug gang violence are dramatically different than for the vast majority of mass public shootings, where attacks are designed to kill or wound as many people as possible," Lott said in a post on the Crime Prevention Research Center Web site earlier this year. "Padding the numbers by lumping the two together doesn't make much sense."

Others argue that the use of the tracker's data sends a misleading signal to the public. "Fear is running rampant," criminologist James Alan Fox of Northeastern University said in an e-mail. "The source of the panic rests squarely with the way in which media outlets cover these tragedies, and in this case, the mass shooting tracker that is attracting so much ink."

"I'm not saying four-plus injured isn't important," he added. "But it is certainly not as serious as four-plus killed, or three-plus killed, or even two-plus killed (as in Roanoke). The problem is that we then are dealing with two variables: number of victims and type of victimization."

But the people running the tracker consider their approach a feature, not a bug. If we are looking to wrap our heads around the totality of gratuitous gun violence in this country, the argument goes, it makes sense to consider the many instances of this type of violence that happen behind closed doors or within regions plagued by gun violence. A life lost is a life lost, whether that person lived in the South Side of Chicago or suburban Colorado Springs.

The tracker uses the same definition of mass shooting as the Gun Violence Archive, a nonprofit organization that tracks firearm injuries and deaths across all categories in the United States, but with one crucial distinction: The tracker includes the attackers in its death tallies if the incident ends in the attacker's death.

"Just because someone commits suicide by cop doesn't mean that the bystanders weren't subjected to an act of violence," one of the tracker's founders told the Trace last month, "or that the cop who pulled the trigger to end the incident won't have to deal with the psychological ramifications of killing another human being."

Other researchers have also been including domestic violence and gang-related shootings as mass shootings for years. Harvard University's David Hemenway said he considers public shootings, domestic violence and gang violence as the three main categories of mass shootings. He said that in many cases, it makes sense to consider them all together.

Regardless of the cause, "four people are still dead," he said in an interview.

Still, it can be useful to separate the three subcategories. "Last year, overall mass shootings weren't increasing that much," he said. But, "the more public ones were increasing. Those look like they're 'contagious' much more than the intimate partner violence ones."

Jim Bueermann, a former police chief in Redlands, Calif. - the city where the San Bernardino suspects lived - said it does not make much sense to distinguish for this purpose between an injury and a death.

"I would submit that sometimes the only difference between a shooting and a murder could be a centimeter, an inch, an unlikely ricochet, whatever," Bueermann, who is now president of the Police Foundation, a kind of think tank for research into effective law enforcement practices, said in an interview. "If we're trying to capture true gun violence in our country, a broader definition [of mass shooting] is probably more useful than a narrow one."