The fact is that most 'objective' measures for academic merit are at best only moderately predictive of success in college. That includes the ACTs and SATs, which don't do a great job of predicting anything.
http://www.washingtonpost.com/blogs/answer-sheet/wp/2014/02/21/a-telling-study-about-act-sat-scores/
(There are many other links demonstrating this I can find if anyone doubts it)
Therefore, there is no clear-cut 'objective' preference between a huge mess of applicants at the margins of being accepted. The question, then, is how you select among these many applicants who have indistinguishable chance of succeeding once admitted.
You're an admissions officers and have 100 applicants for the final admittance spot, and all of them have an indistinguishable but good chance of succeeding at the university, even if some of them have lower GPA, others have lower SATs, others have fewer extracurriculars, others have worse essays, etc. Do you not think the university gains more from
having different perspectives in the classrooms than it does from having the person with a 2% higher SAT score but the same background as the vast majority of other acceptances?