College admissions, a process that thousands of students undergo each year, involves great administrative and human resources. This process has only been investigated in terms of affirmative action, discrimination, correlates for admissions, and the creation of linear systems for prediction. This research applies the decision making under uncertainty framework to this high-involvement, expert decision. Through simulating an actual college admissions process, studies uncover evidence of a variety seeking bias caused by the arbitrary bracketing based on geographic origin. Due to this need for variety, low quality candidates experience a boost in their rankings, relative to an unbracketed condition. A new dependent measure is used, namely the relative value placed upon variety as a characteristic of a candidate. Policy implications are discussed and areas of future research outlined.