Predictors of subjective well-being in patients with paranoid symptoms: is insight necessarily advantageous?