Mild traumatic brain injury (MTBI) is common, yet few studies have examined neuropsychological outcomes more than 1 year postinjury. Studies of nonreferred individuals with MTBI or studies with appropriate control groups are lacking, but necessary to draw conclusions regarding natural recovery from MTBI. We examined the long-term neuropsychological outcomes of a self-reported MTBI an average of 8 years postinjury in a nonreferred community-dwelling sample of male veterans. This was a cross-sectional cohort study derived from the Vietnam Experience Study. Three groups matched on premorbid cognitive ability were examined, those who (1) had not been injured in a MVA nor had a head injury (Normal Control; n = 3214), (2) had been injured in a motor vehicle accident (MVA) but did not have a head injury (MVA Control; n = 539), and (3) had a head injury with altered consciousness (MTBI; n = 254). A MANOVA found no group differences on a standard neuropsychological test battery of 15 measures. Across 15 measures, the average neuropsychological effect size of MTBI compared with either control group was -.03. Subtle aspects of attention and working memory also were examined by comparing groups on Paced Auditory Serial Addition Test (PASAT) continuation rate and California Verbal Learning Test (CVLT) proactive interference (PI). Compared with normal controls, the MTBI group evidenced attention problems in their lower rate of continuation to completion on the PASAT (odds ratio = 1.32, CI = 1.0-1.73) and in excessive PI (odds ratio = 1.66, CI = 1.11-2.47). Unique to the MTBI group, PASAT continuation problems were associated with left-sided visual imperceptions and excessive PI was associated with impaired tandem gait. These results show that MTBI can have adverse long-term neuropsychological outcomes on subtle aspects of complex attention and working memory.