We assess whether the difference between teacher and test-based assessment at age 14 and 10 -as a measure of bias or uncertainty in assessment specifically– is linked systematically to observable pupil characteristics. To the best of our knowledge this is the first paper, focused on the Spanish educational system, which has compared teacher and test-based assessments. Secondly we consider the important, but largely unknown question of whether these measures of uncertainty and bias in assessment have any relationship with pupil’s subsequent educational choices at age 16. The results suggest that the gap between test scores and teacher assessments is particularly large for female students in mathematics, who obtain significantly higher teacher assessments as compared to their test scores. This is consistent with the notion that teachers under estimate the achievement of boys in mathematics, over estimating the achievement of girls as measured by the test score. Additionally we find that students from semi-private schools (concertadas), who are more likely to be wealthier and attend schools with wealthier peers, get higher test scores, as compared to teacher assessments. This is particularly noticeable in reading- which is consistent with the idea that teachers tend to assess pupils in a relative way. Finally there is some evidence of higher probability of attendig Sciences or Social Sciences postcompulsory education for those who present higher difference between teacher and test-based assessment, although do not appear to tell a convincing story about this divergence having any real impact on post-school participation decisions.