Frank Crary said:
The correlation with gun ownership is equally easy to explain. In
the neighborhoods studied, a large fraction of the gun owners are
either criminals (confirmed by the study's correlation previous
convictions)
The study found a correlation between criminal record and homicide,
not between criminal record and gun ownership. In any event there was
a higher risk of homicide associated with gun ownership even after
controlling for criminality.
or purchased a gun in reaction to threats of crime
and violence.
This is unproven.
Both criminals and people threatened by criminals
have a much greater chance of being murdered than the average
person.
The study also controlled for a history of violence in the home. It
is possible that some small number of people threatened with violence
where the threats had not spilled over into actual violence may have
acquired guns for self defence and the guns failed to prevent a
homicide. This number is probably small, or else the data would have
shown a correlation between gun ownership and non-gun homicide.