This article presents the results of a study of the institutional accreditation processes that took place in Chile during the period of 2003 to 2006. The purpose of the study is to estimate the capacity of peer evaluation to predict the final results of institutional accreditation. Thus, a regression model was developed, in which the score given by the academic peers to the institutions -in the accreditation areas of 'Institutional Management' and 'Undergraduate Teaching Process' - were entered as the independent variable, and the accreditation results, expressed in the accreditation decision and the period of accreditation, were included in the model as the dependent variable. The results show that there is consistency between the peer evaluation and the accreditation decision; that the score given by the academic peer to the area of 'Undergraduate Teaching Process' has more capacity to predict the accreditation results than the score of the area 'Institutional Management'; and, that there are arguments to support that the focus on 'evaluation'(in comparison to the focus on audit) was predominant in the model of institutional accreditation in Chile during the studied period.