We investigate the theoretically proposed link between judgmental overconfidence and trading activity. In addition to applying classical measures of miscalibration, we introduce a measure to capture misperception of signal reliability, which is the relevant bias in the theoretical overconfidence literature. We relate the obtained overconfidence measures to trading activity in call and continuous experimental asset markets. Our results confirm prior findings that classical miscalibration measures are not related to trading activity. However, misperception of signal reliability is significantly linked to trading volume, particularly in the continuous market. In addition, we find that men trade more than women at high levels of risk aversion, but the gender trading gap vanishes as risk aversion lessens. The reason is that the trading activity of women seems to be more sensitive to risk attitudes than that of men.