John Lott is a right-wing author who has made claims about the benefits of guns using fabricated evidence. To support his points on the Internet, he adopts various pseudonyms (known as sock puppets) who write in supporting John Lott and giving his books good reviews.
Lott is Resident Scholar at the American Enterprise Institute, a conservative think tank, where he "studies crime, antitrust, education, gun control, campaign finance, and voting and legislative behavior". 
More Guns, More Crime
Lott published a book called More Guns, Less Crime presenting evidence that concealed-carry laws correlated with lower crime rates. However, a more extensive 2002 study found that, if anything, more guns led to more crime. Lott responded with an even more extensive study which confirmed his original claim. But Lott's study was plagued by coding errors which, when fixed, eliminated his conclusion. (This was the second time people had found coding errors in Lott's work.) Lott changed the model to add more bias to his results and restore his conclusions, but did not say so. When confronted with the change, he denied having made it and attempted to cover his tracks by tampering with file-modification dates. Even with the new biased model, "placebo laws" found that Lott's results are not statistically significant.
Tim Lambert, Coding errors in More Guns, Less Crime data.
Lott claimed that 98% of defensive gun uses required only brandishing the gun, originally citing national polls. Lott later cited Gary Kleck as his source. In 2000, Lott later claimed that the number was based on his own survey, not Kleck. Lott claimed the survey was done over three months in 1997, but he first cites it in February 1997. Lott also failed to mention his study when asked by a reporter about it in 1999.
Lott cannot provide details of the survey, saying he lost all his data in a hard drive crash. When asked for funding records or other evidence, Lott said he paid for the project himself, used volunteers students, had them make calls from their own phones, and did not discuss the survey design with anyone else. Lott says he can't remember the names of the volunteers. Meanwhile, nine published surveys showed between 21% and 67% of gun uses required shooting (far more than the 2% Lott suggests).
In fact, the only evidence Lott can provide is that a couple colleagues vaguely remember him mentioning such a survey.
All this has led even extreme conservatives like Michelle Malkin to suggest Lott never conducted the survey. But even if Lott did conduct such a survey, it was unethical of him to cite other sources for the number and continue to publicly cite its results after all of the original evidence had been lost.
Tim Lambert, John Lott's Mysterious Survey.
James Lindgren, Comments on Questions About John R. Lott's Claims Regarding a 1997 Survey.
Michelle Malkin, The Other Lott Controversy, WorldNetDaily, February 5, 2003.
Lott has given himself seventeen five-star reviews using various techniques to conceal that he was the author. He's also written concealed negative reviews of books whose authors he didn't like. (For example, after Michelle Malkin's column denouncing him (see above), he gave her book a one-star review.) He has posed as another person and posted defenses of himself on discussion groups, under false names like "Mary Rosh". ("I have to say that he was the best professor I ever had," Rosh gushed in one Internet posting. "There were a group of us students who would try to take any class that he taught. Lott finally had to tell us that it was best for us to try and take classes from other professors.")
Unmasked by bloggers, he conceded "I shouldn?'t have used [the psuedonym". He later told the Washington Post, "I probably shouldn't have done it -- I know I shouldn't have done it -- but it's hard to think of any big advantage I got except to be able to comment fictitiously."
But the very next day Lott started up again under a new name. He later went so far as to have one puppet disagree with things he posted as another puppet!
Science editor-in-chief Donald Kennedy concluded "What [Lott] did was to construct a false identity for a scholar, whom he then deployed in repeated support of his positions and in repeated attacks on his opponents. In most circles, this goes down as fraud." 
Tim Lambert, John Lott's online book reviews, July 12, 2003.
Richard Morin, Scholar Invents Fan To Answer His Critics, Washington Post, February 1, 2003.
Tim Lambert, John Lott's Fabricated Identities.
Lott claimed that "For every two additional black Republicans in the average precinct, there was one additional non-voted ballot."  In other words, half of all black Republican ballots were invalid. While this is obviously absurd, following it out even contradicts a later Lott claim. Using Lott's numbers, black Republicans alone meant 2.5% of all black ballots were invalid, but Lott later says that less than 1% were invalid.
Tim Lambert, Lott's bizarre claim about Florida 2000.