Academic journal article Journal of Economics and Economic Education Research

Are Worst Students Really More Overconfident? a Preliminary Test of Different Measures

Academic journal article Journal of Economics and Economic Education Research

Are Worst Students Really More Overconfident? a Preliminary Test of Different Measures

Article excerpt


The idea that overconfidence is a common problem among students-particularly the less able-is widely spread among those who work on educational economics and is supported by a series of articles. In Grimes (2002) the score forecasts of the most able students are more pessimistic (under confident) and anyway nearer-in absolute value-to the effective score value with respect to the forecasts of the less able students, who tend to overestimate their performance. Similar results are found for students in Economics by Grimes et al. (2004),Nowell & Aston (2007) and for psychology students by Ochse (2001), Johnson et al. (2012) as well as by the cognitive psychologist Chew (2011) in the data shown in his video guide for teaching and in different contributions. In a previous study Novarese (2009) finds a correlation between ability and overconfidence also among the students of the course in Economics of the Law Faculty in the University of Piemonte Orientale. Grimes (2002), Ambrose et al. (2010) and Kennedy et al. (2002) and many others connect overconfidence to meta-cognitive abilities. The best students are better able to think about what they know: this makes them more aware of their own knowledge and therefore more able to evaluate themselves. However, this mainly explains the higher predictive ability of the better students, but not necessarily overconfidence. On the contrary, one could think that students who have the poorest metacognition have no clue how weak the understanding of a concept is (see for instance Chew, 2012). If on one side this sounds reasonable, on the other it suggests that less able students do not know they are so and this seems a less credible statement: many students know that they do not understand, have lower aspiration levels (Castellani et al., 2010), they do not feel much like studying and so on.

The aim of this work is to give a contribution to this discussion through an empirical analysis about the link that may exist between metacognitive abilities and overconfidence. In particular we want to empirically discuss the existence and the sign of its direction.

Many contributions in the literature (see for instance Clayson, 2005) suggest that the relation between metacognitive abilities and overconfidence needs to be investigated further. For example, in Willingham's (2009) model on students' ability based on cognitive science, knowledge depends on integrated units of knowledge (chunks) or, more simply, in a series of building blocks, without which short time memory has to record and organize too much information5. For example, when a child who is learning to read must understand a sentence, the short time memory effort is enormous: he must recognize letters, keep them in mind to compose words, keep these in mind and so on, while an adult reads the words directly. A student who hardly reads English will strive enormously to study for instance philosophy in English-his mind has to organize too much information simultaneously (understand words, sentences and then interpret them).

In the framework of this reasoning the difficulties and effort that learning implies might limit the student's interest in the subject and cause uncertainty in his abilities to such an extent to make it unlikely for him to be overconfident.

The literature on overconfidence in general is vast, following the article by Russo & Schoemaker (1992). A key point of this literature already proposed by Russo and Schoemaker themselves underlines the connection between the possibility of having a feedback and potential overconfidence. Those who receive clear and continuous feedback should be more able to evaluate themselves. Students have by definition a continuous feedback and in this perspective it seems odd that they cannot adapt to the situation.

In addition the existence of an inverse relation between score and overconfidence may be the result of some measurement artifice. Overconfidence is defined by comparing the predicted score (generally asked immediately before or after the exam) with the actual one. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.