This study aimed to understand the interplay between correctness and confidence in a multiple-choice assessment, specifically to identify misconceptions and evaluate question quality. Existing literature lacks a unifying theory on confidence, with concepts like self-efficacy often overlapping. Shrauger and Schohn define confidence as "perceived (assuredness) in competence," while Bandura's self-efficacy incorporates broader factors. Confidence in task completion refers to certainty, whereas self-efficacy involves emotional and sensitive factors influencing this certainty. Prior research reveals a positive correlation between confidence and accuracy. Assessing student confidence helps identify various confidence-correctness interactions: (1) correct and confident (calibration), (2) correct and unconfident (potential guessing), (3) incorrect and confident (misconception), and (4) incorrect and unconfident (awareness of deficiency). This information enables educators to provide targeted feedback and improve teaching. The study builds on previous US-based research exploring dental student confidence and aims to investigate the generalizability of these findings to a Canadian context.

The literature on confidence and its relationship to accuracy is reviewed. Studies have shown a positive correlation between confidence and correctness. Confidence, along with self-concept, self-efficacy, and anxiety, are reliable indicators of academic achievement. Inquiring about student confidence in assessments helps identify different confidence-correctness interactions. Four situations arise: correct and confident, correct and unconfident, incorrect and confident (misconception), and incorrect and unconfident. Understanding these interactions helps educators identify uninformed and misinformed students and the topics causing misunderstandings. This allows for tailored feedback, encouraging self-reflection and critical thinking. Confidence scoring in multiple-choice exams can reward students for admitting uncertainty. Combining confidence questions with multiple-choice questions can validate and improve assessment question standards. Previous research with US dentistry students identified misconceptions and overconfidence. This study aimed to replicate and extend this research to a Canadian setting.

This cross-sectional study used a convenience sample of 29 second-year dental students in a preclinical endodontics course. Students answered 20 multiple-choice questions (10 'basic' and 10 'moderate' difficulty) on endodontics. Each question was followed by a confidence question (scale: very unsure, unsure, sure, very sure), dichotomized into 'confident' and 'unconfident'. Four situations were analyzed: correct and confident, correct and unconfident, incorrect and confident (misconception), and incorrect and unconfident. Statistical analysis (α = 5%) used Chi-square and Mann-Whitney tests to examine interactions between: (a) students' performance, misconceptions, and confidence; (b) question difficulty, correctness, and confidence; and (c) misconceptions and question type (negative vs. positive, clinical scenario vs. theoretical). OpenEpi and Jamovi software were used for data analysis.

Students demonstrated high correctness (92.5%) and confidence (84.6%). Nine students (31.03%) were responsible for 12 misconceptions. Students with more misconceptions had lower overall performance (P < 0.001). High-achieving students showed low confidence in incorrect responses (P = 0.047). 'Moderate' questions had more incorrect answers (P < 0.05) and less confidence (P = 0.02) than 'basic' questions. All questions were deemed valuable; however, those with negative wording or lacking clinical context led to more misconceptions. Questions with images or requiring clinical scenario visualization had fewer misconceptions (P = 0.007). There was no significant difference in misconceptions between negative and other questions (P = 0.96).

This study demonstrated that incorporating confidence ratings into multiple-choice assessments provides valuable insights into student understanding and question quality. The high correctness and confidence levels might be attributed to the straightforward nature of the assessment and personalized instruction in a small class. However, the identification of misconceptions and the correlation between misconceptions and lower performance highlight the assessment's effectiveness. The finding that high-achieving students were uncertain about incorrect answers suggests a valuable level of self-awareness. The higher incidence of incorrect answers and lower confidence in 'moderate' questions indicates areas for improvement in question design or teaching. Questions involving clinical scenarios and images proved more effective, suggesting the value of visual aids and application-based questions. Although negative questions did not significantly differ in misconceptions, the observed trend warrants further investigation and potential revision for future assessments.

Preclinical endodontic students showed high correctness and confidence. Students with more misconceptions performed poorly. All questions were valuable, but some require improvement. Combining multiple-choice questions with confidence questions helps identify misconceptions and assess question value, enriching feedback and improving future assessments. Future research could explore the generalizability of these findings to larger student populations and different learning environments.

The study used a convenience sample, limiting generalizability. Findings are based on a single assessment, potentially overlooking variations in student performance across multiple assessments. The confidence scale, while helpful, involves subjective self-assessment and may not perfectly capture true understanding. The relatively simple nature of the assessment might not fully assess higher-order cognitive skills.