How good is a model based on bibliometric indicators in predicting the final decisions made by peers?

Several studies have demonstrated that a relationship exists between results from peer-review evaluations and results obtained with certain bibliometric indicators. However, very little has been done to analyse the predictive power of models based on bibliometric indicators. Indicators with high predictive power will be seen as good instruments to support peer evaluations. The goal of this study is to assess the predictive power of a model based on bibliometric indicators for the results of academic openings at the level of Associado and Catedrático at Portuguese universities. Our results suggest that the model can predict the results of peer-review at this level with a reasonable degree of accuracy. This predictive power is better when only the scientific performance is assessed by peers.