Many educators have had the experience of reviewing assessment results and wondering whether the scores tell the whole story. Assessments can be an excellent tool for measuring whether your students are competent.
However, the percentage of questions a student answered correctly on an exam can leave one important question unanswered:
"How certain was the learner?"
At first glance, certainty may seem secondary to accuracy. After all, if a student selects the right answer, haven't they demonstrated competence?
Not necessarily.
Research consistently shows that knowing the correct answer and knowing that you know it are not always the same thing. Studies have found that learners' confidence in their responses may not correspond well with whether those responses are actually correct.1
One learner can answer incorrectly while feeling highly confident, revealing a misconception that requires attention, while yet another might answer correctly but with significant uncertainty. Both suggest fragile knowledge that may not yet be ready for clinical application.
In both cases, the answer alone tells only part of the story.
In health care, knowing the correct answer is important. Knowing when to trust your judgment is equally important.
Clinical practice requires constant self-monitoring. Physicians, nurses, physician assistants, and other health professionals must continually assess what they know, what they do not know, and when they need additional information or support. Overconfidence can contribute to errors, while underconfidence can lead to delays, hesitation, or missed opportunities to act.6
This ability to accurately judge one's own knowledge is often referred to as calibration or self-monitoring. It is a critical component of clinical reasoning and lifelong learning. Yet many assessments measure only whether an answer is right or wrong.
By incorporating confidence into assessment, educators gain insight into a learner's ability to evaluate their own understanding, a skill that extends far beyond a single test question.
Consider two students who answer the same question correctly:
Student A is highly confident.
Student B is unsure and selects the correct answer with hesitation.
Traditional assessment treats these learners identically. Both receive credit and move on.
From an educational perspective, though, they may have very different learning needs.
The first learner may have achieved durable understanding; that is, the relevant information is more likely to be locked in long-term. The second student could have arrived at the correct answer through guessing, educated or otherwise. Research suggests that these low-confidence correct responses often represent opportunities where reinforcement and feedback can meaningfully improve retention.1,2
The same principle applies when students answer incorrectly. An incorrect answer paired with low confidence often reflects incomplete knowledge. An incorrect answer paired with high confidence can signal a deeply held misconception that might be more difficult to correct.2,3
For learners, these distinctions matter because they create actionable feedback. Instead of seeing only what they got wrong, students gain insight into where they may be overconfident, underconfident, or still developing mastery.
|
In other words, confidence helps every learner answer the question: What should I focus on next? |
Confidence-informed assessment also provides educators with richer information than correctness alone.
When faculty can distinguish between learners who are uninformed, misinformed, or simply uncertain, they can provide more targeted coaching and remediation. Research suggests that confidence data also can help faculty identify misconceptions, target remediation efforts, and better understand patterns of learning across a cohort.2
For example, a group of students may perform adequately on an assessment while demonstrating widespread uncertainty in a particular content area. That finding may suggest opportunities to strengthen instruction, provide additional practice, or revisit key concepts before learners encounter them in clinical settings.
The result is a more nuanced understanding of learner development and a stronger foundation for educational decision-making.
These ideas are one reason Aquifer designed Calibrate: to incorporate both knowledge and certainty into the learning process.
Rather than evaluating learners solely on whether an answer is correct, Calibrate helps students reflect on how confident they are in their understanding. The resulting reports provide a more complete picture of learning by highlighting areas of strength, uncertainty, overconfidence, and potential misconceptions.
Studies suggest that combining accuracy and confidence provides a more precise picture of learner understanding and can help identify knowledge gaps that might otherwise remain hidden.4,5
For students, this creates opportunities for more focused study and self-reflection. For educators, it provides additional context for coaching, feedback, remediation, and curricular improvement.
Perhaps most importantly, it reinforces a skill that extends well beyond assessment itself: the ability to evaluate one's own knowledge accurately.
Interested in learning more about Calibrate? Request a meeting with our team to explore Aquifer's formative assessments and more.
As health professions educators, our goal is not simply to help learners accumulate facts. We are preparing future clinicians to make decisions in complex, uncertain environments.
Correct answers will always matter. But clinical readiness also depends on learners recognizing when their understanding is solid, when it is incomplete, and when they need additional information or support.
By helping learners develop that self-awareness alongside clinical knowledge, assessments can become more than measures of performance. They can become tools for developing sound clinical judgment.
Sometimes the most important insight isn't whether a learner got the answer right; it's whether they knew they were right.
References:
1 Butler AC, Karpicke JD, Roediger HL III. Correcting a metacognitive error: feedback increases retention of low-confidence correct responses. J Exp Psychol Learn Mem Cogn. 2008;34(4):918-928. https://learninglab.psych.purdue.edu/downloads/2008/2008_Butler_Karpicke_Roediger_JEPLMC.pdf. Accessed June 15, 2026.
2 Curtis DA, Lind SL, Boscardin CK, Dellinges M. Does student confidence on multiple-choice question assessments provide useful information? Med Educ. 2013;47(6):578-584. https://escholarship.org/content/qt4849x0m3/qt4849x0m3.pdf. Accessed June 15, 2026.
3 Hasan S, Bagayoko D, Kelley EL. Misconceptions and the Certainty of Response Index (CRI). Phys Educ. 1999;34(5):294-299. https://www.researchgate.net/profile/Diola-Bagayoko/publication/241530804_Misconceptions_and_the_Certainty_of_Response_Index_CRI/links/53d2e74d0cf220632f3cc30a/Misconceptions-and-the-Certainty-of-Response-Index-CRI.pdf. Accessed June 15, 2026.
4 Smrkolj Š, Bančov E, Smrkolj V. The reliability and medical students’ appreciation of certainty-based marking. Int J Environ Res Public Health. 2022;19(3):1706. https://www.mdpi.com/1660-4601/19/3/1706/pdf. Accessed June 15, 2026.
5 Tabibzadeh N, Mullaert J, Zafrani L, et al. Knowledge self-monitoring, efficiency, and determinants of self-confidence statement in multiple choice questions in medical students. BMC Med Educ. 2020;20:445. https://link.springer.com/content/pdf/10.1186/s12909-020-02352-6.pdf. Accessed June 15, 2026.
6 Tweed MJ, Purdie G, Wilkinson TJ. Defining and tracking medical student self-monitoring using multiple-choice question item certainty. BMC Med Educ. 2020;20:344. https://link.springer.com/content/pdf/10.1186/s12909-020-02250-x.pdf. Accessed June 15, 2026.