Is formative assessment more effective than summative assessment for learning?

Summative assessment boosts scores, but formative assessment predicts who needs help

If your goal is to get students to score higher on a specific test, summative assessment (graded, high-stakes exams) is clearly more effective than formative assessment (low-stakes, feedback-only quizzes). A large 2025 study of 1,894 medical students across four Dutch universities found that when the same national pharmacotherapy test was used as a summative assessment, the average score was 84.3% and the pass rate was 60.4%. When it was used as a formative assessment, the average score dropped to 67.5% and the pass rate plummeted to just 5.9% [1]. That's a 17-percentage-point gap in scores and a tenfold difference in pass rates—a massive effect. The reason is simple: students study harder when the stakes are high.

However, if your goal is to identify which students are at risk of failing before the final exam, formative assessments are more valuable. A 2024 study of 82 first-year medical students found that scores on formative assessments (along with later summative assessments) could predict final university exam scores with reasonable accuracy (R = 0.76, meaning 58% of the variation in final scores could be explained by earlier assessments). But crucially, the formative assessments themselves were not the strongest predictors—it was the later summative assessments (the second and third in the series) that were statistically significant [2]. This means formative assessments are useful early warning signals, but they need to be combined with later, more formal assessments to reliably flag struggling students.

Another 2023 study of 250 first-year medical students confirmed this pattern: students who scored above 50% on formative assessments went on to score significantly higher on summative exams than those who skipped or failed the formative assessments. The difference was statistically significant (P < 0.05), meaning it's not due to chance [5]. So formative assessments act like a canary in the coal mine—they don't guarantee high scores, but they reliably identify students who need extra support.

Blending formative and summative assessment works better than using either alone

The most effective approach is not to choose one over the other, but to design an assessment system that uses both in a coordinated way. A 2022 study from dental education tested a model where students completed a formative assessment at the end of one course (with detailed feedback on their level of understanding using the SOLO taxonomy—a framework that categorizes answers as incorrect, descriptive, or relational/deep understanding), and then a summative assessment in a subsequent course. The result: most students' responses developed to a higher level of understanding between the two assessments [4]. The formative feedback wasn't just a grade—it told each student exactly where their thinking was shallow, and the summative assessment then checked whether they had deepened that understanding. This interdependence is key: the formative assessment guides learning, and the summative assessment certifies it.

A 2023 study from veterinary education tested a clever hybrid called the "On The Spot Presentation-based Assessment" (OTSPA)—a low-weight summative assessment that felt formative because it was conducted in a supportive environment. Students prepared for all topics, but only a subset was tested on the day. 79.6% of students said the preparation drove their understanding of the material, and 80.4% said it improved their communication skills. Importantly, performance on this low-stakes summative assessment correlated positively with performance on the final written summative assessment (a small but significant correlation across all three modules tested) [3]. This suggests that even a low-stakes summative assessment can serve a formative purpose—giving students feedback without the crushing pressure of a high-stakes exam—while still predicting later success.

The takeaway: don't pit formative against summative. Use formative assessments early and often to give feedback and identify struggling students. Use summative assessments later to confirm learning and motivate effort. When designed together, they reinforce each other.

The effectiveness depends on the assessment system—programmatic vs. traditional

The same assessment can perform very differently depending on the overall assessment system it's embedded in. The 2025 Dutch study compared three systems: traditional with summative assessment (high-stakes, graded), traditional with formative assessment (low-stakes, ungraded), and programmatic assessment (where many low-stakes assessments are combined to make a holistic judgment, with no single high-stakes test). The results were striking: when schools switched from formative to summative within a traditional system, scores jumped by 14.4 percentage points and pass rates soared by 42.3 percentage points [1]. But when schools switched from a traditional summative system to a programmatic system (where the same test became non-high-stakes), scores dropped by 3.3 percentage points and pass rates fell by 14.2 percentage points [1].

This means that the "effectiveness" of formative vs. summative assessment is not a fixed property—it depends on the culture and stakes of the entire program. In a traditional system where students are used to high-stakes exams, making an assessment formative (ungraded) can dramatically reduce effort and performance. But in a programmatic system where multiple low-stakes assessments are the norm, the same test can still motivate learning without the same drop-off. The key is alignment: the stakes and feedback must match the system's overall design.

For educators: if you're in a traditional exam-heavy culture, summative assessments will likely produce higher scores. If you're building a programmatic assessment system, you can use formative assessments effectively, but you need to design the whole system carefully to maintain student motivation.

Sources used in this answer

Related research questions