The purpose of the study was to investigate the establishment of an early warning system and subsequent intervention with college freshmen, which addressed both the academic viability and retention of first-year students. The population of interest was first-semester students who were predicted to succeed by standard admissions criteria, but fail to achieve minimum academic standards at the university level. The two main goals of the study were to: a) validate the use of a prediction model to establish a system of early identification of first-semester college students who are at-risk of academic difficulty, and b) examine the efficacy of an intervention designed to support these students. Using first-semester GPA as the outcome variable, a prediction equation was developed using multiple regression with historical data (both cognitive and non-cognitive variables) from a prior cohort of freshmen. The equation was applied to a new cohort of freshmen, who were assigned levels of academic risk. An advising-based intervention was utilized with the most academically at-risk students in the participating academic unit. The results of the study showed that the risk prediction equation was modestly correlated with first-semester GPA (r=.48, p<.001), while a step-wise multiple regression revealed that individual predictors were differentially effective among the risk levels. Comparisons of first-year academic performance indicated that the intervention was not effective in remediating risk variables for the experimental group. While the risk prediction and categorization system showed promise, modifications could be made to further enhance effectiveness across the full continuum of risk levels, including the development of the equation with subgroups of interest as well as use of more sophisticated instrumentation to measure key non-cognitive variables. In order to establish an effective intervention, future efforts should be guided by knowledge of individualized risk factors and relevant theoretical approaches.