experience combat or participate in abusive violence had lower conviction rates than era controls. However, if the veteran experienced combat or abusive violence, the results showed impressive effects of heavy combat exposure on arrests and convictions. The arrest rate increased by an average of 2.32 percentage points and the conviction rate by 1.23 percentage points for every point increase on the Combat Exposure Scale. When preservice background factors were statistically controlled for, combat exposure continued to show an association with arrests and convictions (generally for nonviolent offenses). The authors concluded that arrests in the years after service, including arrests that led to conviction, increased sharply with exposure to combat. Of the Vietnam veterans with heavy combat exposure who were arrested after discharge, 19.6% had been arrested for nonviolent offenses and 4.9% for violent offenses, compared with 17.7% and 2.1%, respectively, of veterans with moderate combat exposure and 5.7% and 0.6% of those with light combat exposure.
In the 1986-1987 NVVRS, Kulka et al. (1990) found that the 406 Vietnam theater veterans with high levels of war-zone stress were more likely to have been arrested or jailed than the 783 theater veterans with moderate-low war-zone stress (39.1% vs 27.7%) and were almost 3 times as likely (8.8% vs 2.8%) to have been convicted of a felony. Involvement with the criminal-justice system was based on self-reports of number of times arrested since the age of 18 years, nights spent in jail or prison (since the age of 18 years), and number of lifetime convictions for a felony offense and on whether the veteran was in jail or prison at the time of the interview. Of the 319 Vietnam theater veterans with current PTSD, 45.7% had been arrested or jailed more than once in their lives compared with 11.6% of the 871 theater veterans without PTSD; 11.5% of the veterans with PTSD had been convicted of a felony. PTSD symptom level at the time of the interview was assessed with the Mississippi Scale for Combat-Related PTSD; a study cutoff score of 94 was used as a threshold for an assessment of current PTSD. A diagnosis of PTSD was validated in a subsample of the veterans with the SCID. Black theater veterans were more likely to be involved in the criminal-justice system than whites or Hispanics, but black and Hispanic theater veterans with low war-zone stress were less likely to be involved in the criminal-justice system than black or Hispanic era veterans.
In contrast, the other three secondary studies were less likely to attribute incarceration to deployment. In a small study of 61 Vietnam veterans (31 incarcerated and 30 not) in Iowa state prisons, Shaw et al. (1987) concluded that previous personal characteristics, such as a history of antisocial behavior, and not wartime experiences best explained incarceration. Combat stress was similar in both groups although high stress was associated with the development of PTSD. Incarcerated veterans were found to have a higher prevalence of antisocial personality disorder than the community sample (36% vs 7%) and a higher prevalence of alcohol and drug abuse. In the study by Hiley-Young et al. (1995) of 177 Vietnam combat veterans who were receiving treatment for PTSD as inpatients at a VA hospital, high levels of combat exposure were not found to predict criminal activities. A strength of this study is that the authors verified combat status with military records.
Boivin (1987), studying incarcerated male Vietnam veterans in a maximum-security section of a Michigan state prison, found that incarcerated veterans (n = 46) were more likely to be black, to have come from a less supportive family background, to have been assigned to an Army infantry combat unit, and to have been injured in combat than incarcerated nonveterans at the same prison (n = 19), nonincarcerated Vietnam veterans from a Vietnam Veterans of America chapter (n = 28), and nonincarcerated non-Vietnam veterans from a National Guard group at a local college (n = 28). The incarcerated veterans were also more likely to have been