This is especially true when statistics are used to reinforce entrenched positions during highly charged political debates.
One particular false statistic -- even under the most charitable reading of the data, it's highly misleading -- is trotted out after virtually every major terrorist attack in order to claim that Islamic terrorism is a big nothingburger.
I've had to deal with various incarnations of this here before:
94% of all terror attacks are committed by non-Muslims
Only 6% of terror attacks in the U.S. are by Muslims
And yet another (one that is flatly false, because the data cited is never broken out by sex or race) says:
94% of terror attacks are committed by white men
The data behind this statistic comes from the FBI's Terrorism 2002-2005 report. You can find it on the FBI's website here.
First thing to consider when you see this statistic bandied about: realize the data is more than a decade old.
Including the data from 2006-present will give very different results (results that purveyors of these claims may not like), but since this is the data set they choose to use, I will use their preferred source.
Second, note that this statistic is counting "incidents." The data ends up being portrayed like this:
The problem with using this standard of measure is that it is basically useless. An ecoterrorist mailbox bomb that doesn't injure or kill anyone is given the exact same weight as the Oklahoma City bombing or the 9/11 attacks. So how does measuring terrorism in this way tell you anything important?