James Taranto, in Opinion Journal's Best of the Web draws attention to the media habit of referring to the defendants and suspects in criminal cases as being war veterans. This contrasts with the general media view (incorporated in Britain's PCC Code) that someone's ethnicity should not be referred to gratuitously.
In what may be an attempt to justify the stereotype, The New York Times tries to make an assessment of whether or not war veterans are disproportionately likely to be involved in crime. Taranto - quite properly - ridicules their methodology. They do not make an attempt to compare crimes by veterans with the total number of crimes in the US population; do not take account of the fact that most veterans of current conflicts are between 19 and 30 and male; and use media reports as the basis for their analysis. The NYT seems to be saying "other media are using this stereotype so it is okay for us to do so".
But there is another point here. The analysis the paper conducts is predicated on an unstated premise that if war veterans are more likley to be involved with crime than others then it would be okay to gratuitously refer to the status of a veteran in a crime report. This is the complete opposite of the assumption behind the movement to ban gratuitous references to race. Here the premise is that even though people from ethnic minorities are known to be disproportionately involved with crime, journalists should not refer to ethnicity unless it is relevant to the crime. This double standard is not even discussed, let alone justified in the NY Times article.