How much should we worry about Islamic terrorism? How much should we worry about other kinds? There's no exact right answer to this question. Who is out there in dark places plotting murder most foul? We can only guess, using imperfect information. Of course, there's "imperfect" and then there's downright distorted.

The New York Times highlighted one data set recently, in an article headlined "Homegrown Extremists Tied to Deadlier Toll Than Jihadists in U.S. Since 9/11." "Since Sept. 11, 2001," the article says, "nearly twice as many people have been killed by white supremacists, antigovernment fanatics and other non-Muslim extremists than by radical Muslims: 48 have been killed by extremists who are not Muslim, including the recent mass killing in Charleston, S.C., compared with 26 by self-proclaimed jihadists, according to a count by New America, a Washington research center." The article goes on to cite a nationwide survey of police and sheriffs departments, noting that "74 percent listed antigovernment violence, while 39 percent listed 'Al Qaeda-inspired' violence, according to the researchers." Well, I guess that settles that, then.

Ah, no. Statistics are useful, but fragile. How you handle them makes a big difference.

CartoonMatt Davies' latest cartoon: HourglassCommentSubmit your letterReader essaysGet published in Newsday

The most obvious thing to note is the choice of start date: Sept. 12, 2001. That neatly excludes an attack that would dwarf all those homegrown terror attacks by several orders of magnitude. Ah, you will say, but that was a one-time event. Sort of.

It is no longer possible to destroy the World Trade Center, but we can't be certain to never again have a large-scale terror attack that kills many people. If you have high-magnitude but low-frequency events, then during most intervals you choose to study, other threats will seem larger -- but if you zoom out, the big, rare events will still kill more people. We don't say that California should stop worrying about earthquake-proofing its buildings, just because in most years bathtub drownings are a much larger threat to its citizens.

The other thing to ask is how we're defining a terror event and classifying the motivation. I took a little stroll through the underlying data, and on the "jihadist violence" side, the definition is pretty clear: with the exception of one case in which a Muslim who seemed fond of jihadist propaganda beheaded a coworker for reasons that are not entirely clear, the rest of the attacks involved someone with an ideological commitment to radical Islam trying to kill a bunch of people in a way that made it clear that this was about U.S. foreign policy in the Middle East.

advertisement | advertise on newsday

Counting the other types of extremist terrorism is a little murkier. Some of them are fairly obvious: When a white supremacist starts shooting people at a Sikh temple, I don't think we need to wonder too hard what his motives were. On the other hand, the data set The Times relies on also includes Andrew Joseph Stack, who you may remember piloted a small plane into an IRS building in Austin. Stack left a manifesto behind, and it doesn't exactly read like an anarcho-capitalist treatise. Oh, he's mad at the government, all right, but he's mad about ... the 1986 revision to Section 1706 of the tax code, which governs the treatment of technical contractors.

Here are some other things Stack was angry about: -- The bailouts of GM and Wall Street -- Drug companies and health insurers (the health-care bill was then stalled in Congress) -- The Catholic Church and the "monsters of organized religion" -- The Pennsylvania steel bankruptcies that gutted steelworker pensions -- Now-defunct accounting firm Arthur Andersen -- Sen. Daniel Patrick Moynihan (because of Section 1706) -- The California base closings of the early 1990s -- The 1980s S&L crisis -- Government aid to airlines after 9/11 -- His accountant -- George W. Bush Its closing lines are "The communist creed: From each according to his ability, to each according to his need. The capitalist creed: From each according to his gullibility, to each according to his greed." Labeling this as a "deadly right- wing attack" is beyond a stretch; it's not even arguably correct.

Nor is this the only questionable inclusion. Consider Raymond Peake, who was convicted of shooting someone at a firing range, apparently in the course of stealing his gun (it was not the first time Peake had stolen a gun, but it was the first time he'd shot anyone); he appears to be on the list on the basis of a single vague statement from law-enforcement that Peake had been stealing guns for an unidentified organization aimed at overthrowing the United States government. His "co-conspirator," whose lawyer denied that he had any knowledge of Peake's alleged crimes, ultimately plead guilty not to conspiracy to overthrow the government, but to receiving stolen property. Maybe there was a shadowy plot to overthrow the U.S. government with the four guns they found in the co-conspirator's home. On the other hand, maybe a suspect just started rambling when he was arrested for murder.

Sign up for The Point

Go inside New York politics.

Then there was Joshua Cartwright of Ft. Walton Beach, Florida, who shot two deputies when his wife called the cops to stop him from hitting her. This was elevated to a "deadly right- wing attack" because, according to New America, "Cartwright had a history of non-compliance with the police and Cartwright's wife told police that he held anti-government views and was 'severely disturbed' by President Obama's election." All this may be true. But it's dangerous to profile so that every person with vaguely stated right-wing views, or even not-so-right-wing views, becomes an avatar of that group, rather than an individual who happens to be a member of that group, and also happens to have done something bad.

The case of Robert Poplawski is similarly questionable. He ambushed three officers who responded when his mother called the police on him. He also frequented white supremacist websites and espoused anti-government racist views, according to the database. He also wrote his grandmother's name in his own blood on a bedroom wall on the day of the shooting, and told the police negotiator "You know, I'm a good kid, officer. ... This is really an unfortunate occurrence, sir." Which does not exactly sound like a crazed right-wing terrorist determined to take down the government By Any Means Necessary.

Add to the list of "not clear what he was thinking, but probably not domestic terrorism" Curtis Wade Holley, who set fire to his own home and then shot at the first responders. The timeline suggests he was upset because his ex-girlfriend finally had his utilities shut off and he was worried about being evicted or losing his car, something he'd vowed not to endure without a fight. The evidence for him as a "right-wing attacker," rather than just a paranoid and broke marijuana grower, seems to be that someone, possibly the ex-girlfriend, had called police to say that he had anti-government views and would shoot cops if they came to his place. Would a similar situation with someone known to be an Irish nationalist be an example that The Troubles had crossed the Atlantic to the United States? I find it very hard to understand why these cases were included, except to pad out the count of "deadly right-wing attacks." Presumably we are looking for political terror for a political purpose, not every violent crime by a Muslim or a right-winger. This means the acts must include some amount of premeditation, some intent to pursue an ideology, not a flash shootout precipitated by a completely unrelated event, like beating your wife or getting your utilities shut off. Restricting the count to attacks that seem to have had a political purpose, and an ideology that could be convincingly described as "right wing," drops the tally of right-wing terror to 41 or less.

I'm also somewhat dubious about Albert Gaxiola, Shawna Forde and Joshua Bush, who killed a man and his 9-year-old daughter while robbing their house. The database says "The three conducted the robbery to help fund their anti-immigrant organization." But prosecutors told jurors that "it was Gaxiola who suggested Forde and Bush ought to rob and kill Flores. Gaxiola wanted Flores dead because he was a rival drug smuggler." Forde and Bush were, according to prosecutors, seeking money to fund their Minutemen organization, but once you start to bring black-market assassinations into this, things start looking a little murkier than a case of "deadly right-wing attacks." To be generous and round up the numbers for right-wing terror, I could argue for including the Gaxiola trio and Peake. However, once you start throwing in the gray cases on the right- wing side, shouldn't we be similarly permissive on the Islamic terror side? In prison, one of the Beltway snipers penned rambling anti-American screeds in which the Baltimore Sun said that "the most recurring theme is that of jihad - or holy war - against America."

The Beltway snipers killed 10 people, which all by itself would bring the number of jihadist killings up to 36. Then the story becomes less "right-wing terror is much more dangerous than jihad" and more "Muslim terrorists have killed some people in the United States, and other kinds of ideological murderers have too." What's the takeaway? Never think that because you have a nice, hard-sounding number, that number tells you what you want to know. Numbers don't just grow in the wild; they are chosen, by parameters that the researchers decide.

advertisement | advertise on newsday

The parameters these particular researchers chose might not be the criteria you would use; they are certainly not the ones that I would have chosen. And even if you agree that these are absolutely the right and proper numbers, that still doesn't tell us that right-wing terror is more dangerous to us, the living, than to the people during the time period they studied.

To know that, you would need to know who remains out there, plotting dark things.

Megan McArdle is a Bloomberg View columnist who writes on economics, business and public policy.