This study sought to examine the nature of metacognitive judgments: feeling of knowing and subjective confidence, with an emphasis on examining the impact of question weight and explicitly added accurate and inaccurate information about consensuality. The aim of the study was to examine the accuracy of these metacognitive judgments, their correlation considering recall accuracy and the impact of question weight and additional information about consensuality on the metacognitive judgments. The study included 82 psychology students of Faculty of Humanities and Social Sciences, University of Rijeka. In Experiment 1 participated 30 students and in Experiment 2 participated 52 students. The age range was between 19 and 27. In Experiment 1 participants answered the 90 general knowledge questions and assessed feeling of knowing and confidence in answer accuracy, for each question presented. In Experiment 2, subjects answered the same questions as in the Experiment 1 and also assessed the feeling of knowing. After that, additional information about consensuality were presented: an information about the percentage of respondents who answered correct in Experiment 1. This information was true in half of question, and incorrect for the second half.The results indicated that metacognitive judgments are accurate in predicting performance. Feeling of knowing and confidence estimates for correct and incorrect answers are significantly positively related. Results of item analysis pointed that the confidence judgments are higher in a situation with no additional information in relation to the confidence judgments in situation with the additional information. Also, confidence estimates for easy questions are significantly higher than estimates for medium and difficult questions. Effect of the accuracy of consensuality information and interaction are not confirmed. In situation with the additional consensuality information, the results of participants analysis pointed that confidence estimates are higher than feeling of knowing estimates. Also, the estimates are higher with presenting accurate consensuality information compared to estimates with inaccurate information, and they are the highest following easy questions, then following medium, and lowest for difficult questions. Interaction effects suggest that participants estimate the highest confidence to easy questions with correct consensualityinformation, and the lowest feeling of knowing estimates for difficult questions, regardless of the accuracy of the information presented.