Why Making Mistakes Might Be The Smartest Learning Strategy

Most people treat a wrong answer as a failure. New research suggests the brain sees it differently. A study from the National University of Singapore found that guessing at foreign vocabulary words before being shown the correct answer leads to better retention than simply reading the right answer from the start, even when the guesser has no idea what they’re doing.

Researchers Tabitha J. E. Chua and Steven C. Pan put this to the test across four controlled experiments, asking adults to learn Spanish vocabulary through one of two methods: guess first, then see the answer, or just read the word paired with its correct image. In every experiment, the guessers came out ahead on memory tests afterward.

What makes the finding worth paying attention to is who was doing the guessing. Participants were screened to exclude prior Spanish knowledge, though the authors note some words may still have felt partly familiar. For the most part, these were people taking shots in the dark, and the wrong answers still helped them learn.

Why Getting It Wrong Helps the Brain Learn

Memory researchers have long known that attempting to retrieve something, even unsuccessfully, can strengthen learning. One explanation holds that making a guess activates related knowledge in the brain, building a mental bridge that makes the correct answer easier to attach and remember. Prior research demonstrated this with semantically related word pairs like “doctor” and “nurse,” where a wrong guess still creates a useful mental connection.

But Spanish vocabulary words have no such connection to their English equivalents for a first-time learner. Nothing about the word “manzana” hints at “apple.” According to the researchers, this is where a second explanation gains traction: the mismatch between a wrong guess and the correct answer may trigger a kind of heightened attention. When the brain registers that it got something wrong, it pays closer attention to the feedback, encoding the right answer more firmly as a result.

Guessing accuracy across the experiments hovered around 35 to 38 percent, modestly above the 25 percent expected from pure chance. Critically, the authors found that learning gains also appeared for items participants guessed incorrectly, meaning the benefit wasn’t limited to words that may have felt half-familiar going in.

Guessing vs. Reading: How Adults Learned Spanish Vocabulary

Chua and Pan recruited adult participants through an online platform called Prolific Academic. All were fluent English speakers between the ages of 21 and 45, living in English-speaking countries, and screened to exclude prior Spanish exposure. Individual experiment sample sizes ranged from 47 to 123 after excluding those who failed screening checks or didn’t complete the study, published in Cognitive Research: Principles and Implications.

Each participant learned 36 Spanish nouns, concrete everyday words covering categories like animals, clothing, tools, and food. In the guessing condition, participants saw a Spanish word and chose the correct meaning from four picture options, with eight seconds to guess and five seconds to view the correct answer. In the reading condition, participants simply studied the word paired with its correct image for five seconds. Both groups spent the same amount of time viewing the correct word-and-picture pairing, though the guessing group also had additional exposure to the Spanish word alone before feedback arrived.

After completing unrelated filler tasks, everyone took a memory test. Some questions required participants to recall the English meaning of a Spanish word from scratch, while others offered multiple choices. Researchers varied the structure across all four experiments, testing different combinations of question formats and learning directions to see whether the guessing advantage would hold.

Wrong Answers Outperformed Reading Across the Board

On free recall tests, where participants had to pull the correct answer entirely from memory, guessing beforehand consistently beat reading. On multiple-choice tests, the advantage was often even larger. Only one experiment out of four failed to show a statistically clear benefit on multiple-choice questions, and the researchers attributed that to performance in both groups being so high that there was little room to show a difference.

Whether test questions were scrambled together or grouped by type, the guessing advantage persisted. That consistency across different experimental setups strengthens the case that the effect reflects something real about how memory works, not a quirk of one particular design.

Most participants also felt the guessing method was more useful. In several experiments, around 60 percent said they found it more helpful and would prefer to use it going forward.

Apps like Duolingo and Rosetta Stone already build their vocabulary lessons around this exact format, presenting a word, asking users to guess, then revealing the correct answer. According to Duolingo’s own reporting, as cited in the paper, the app has surpassed 100 million monthly active users. This research adds scientific weight to the idea that the format isn’t just engaging. It may work precisely because it asks learners to be wrong first.

That said, the study tested a single learning session with 36 concrete vocabulary items. How well these results hold up across longer periods of study, with younger or older learners, or with more abstract material remains to be seen.

---