Washington: Overconfidence can be perilous, says new research that claims past studies on the subject were flawed as they relied on the participants' own assessment of their confidence levels.

For instance, the investment banker who ended up losing millions on a "can't-miss" start-up or the drunken driver who insisted on driving back home and crashed his car are unlikely to rate themselves as over-confident.

Now, Pascal Mamassian, a French researcher, believes he has found a way to circumvent this problem. He has demonstrated that over-confidence can be assessed by a natural and objective visual-motor task.

Participants in Mamassian's study sat at a computer and were asked to press a key in synchrony with a visual "blob" that would appear on the screen.

They would be awarded points if they succeeded and docked points if they pressed the key prematurely or too late.

Mamassian then used a mathematical model to examine how participants would need to adjust their key tapping strategy in order to maximise gain and minimise loss.

He found that participants routinely failed to aim toward the optimal time, instead displaying over-confidence in their action. Specifically, "they underestimated the magnitude of their uncertainty and the cost of their error", he said.

Because of the objective nature of the task, Mamassian said: "Overconfidence was not limited to the realm of subjective beliefs and cognitive judgments but appeared instead to reflect a general characteristic of human decision-making."

The findings of the study have been published in the latest issue of Psychological Science, a journal of the Association for Psychological Science. IANS