Several months ago one of National Review's blogs ran posts by contributor W. Thomas Smith, Jr. from Beirut. Smith described several things"a sprawling Hezbollah tent city" with "some 200-plus heavily armed Hezbollah militiamen," and "4,000 and 5,000 HezB gunmen deployed to the Christian areas"that appear not to have existed. Thomas Edsall has the details here, with more here from David Kenner.
This in itself is no judgment on National Review. While some publications do better fact-checking than others (Mother Jones is one of the best, in my experience), no one can produce 100% error-free journalism.
What does matter is how editors react when problems are brought to their attention. Glenn Greenwald explains what's happened to date here. National Review's online editor Kathryn Jean Lopez was alerted to the implausibility of Smith's claims six weeks ago by two journalists with experience in Lebanon. Yet she appears to have acted only when she realized Edsall would be writing about it. (ADDED CLARIFICATION: Mitchell Prothero, a Beirut photographer, writes here that he contacted National Review "more than six weeks ago." Harper's publishes here an email sent by journalist Chris Allbritton on October 6th calling Smith "a liar." However, it's unclear who at National Review would have received it. Lopez posts here her version of events, which does not mention Prothero or Allbritton.)
This reminds me of my own experience with Lopez. In October of last year, she wrote a post about the Johns Hopkins study estimating the number of deaths in Iraq since the invasion as 655,000. Lopez printed email from "A Hill Guy" reading: