The present study examined the nature of verbal memory deficits in individuals with traumatic brain injury (TBI) compared to healthy controls. The study was designed to control for methodological shortcomings of previous related research. Three groups of participants were used: (a) a head injured sample with moderate to severe traumatic brain injuries (N=55), (b) a control sample matched on age and initial performance on CVLT Trial 5 and Sum of Trials 1 to 5 (N=55), and (c) a control sample matched on age, education, and race, but not on initial CVLT learning performance (N=55). Current findings indicate that: (a) rate of learning was comparable across groups, consistent with no encoding differences, (b) TBI patients have a significantly more rapid rate of forgetting of new information than either acquisition-matched or demographic-matched controls, consistent with consolidation problems in TBI, (c) TBI patients have less proactive interference than demographic-matched control participants, consistent with a consolidation problem in the TBI group, (d) TBI patients and acquisition-matched controls have comparably low rates of proactive interference, consistent with impaired acquisition in both of these groups, and (e) TBI patients and controls do not differ in the benefit experienced from semantic or recognition retrieval cues, consistent with no differences in retrieval processes. These data support an impaired consolidation hypothesis, rather than encoding or retrieval deficits, as the primary deficit underlying memory impairment in TBI.