Tom Wright demonstrates perfectly the misleading nature of Lott's postings on the coding errors:
He also claims that the 100's of errors claimed for the study could make a difference without mentioning that these errors were corrected and the study still showed the same results to within thousandths of a percent of the original result. This is like claiming an elephant is a mouse because the claimed multi-ton weight was off by a few grams.
You see, while Lott admits that the "estimates do change somewhat", he does not tell you how much they change. Instead he talks about how only a few thousands of one percent of the values were wrong, creating the impression that the estimates changed by that much. They did not. Lott does not tell you that they actually changed by more than a factor of two in some cases. It is not that easy to see this because the "corrected" tables are hard to find and in a password protected section of his web site. I have made a copy of one of them, so it can be compared with the ones in the original paper.
Wright also says:
Unfortunately, in the debate on guns, refusal to release data by pro-gun-control advocates is so common that I sometimes jump to conclusions. We still do not have the data or methodology for the commonly quoted canard that '10/11/15/16 children a day are killed by guns'. The number varies day to day.
Actually what is common is false claims by pro-gun folks that their opponents don't release their data. The number of children killed with guns is easily obtained from the CDC here. The number given varies because the number of such deaths changes from year to year.