How do tests help students learn?

In a recent study, researchers demonstrated that giving students short, topic-specific quizzes after an exam improves their chemistry learning performance.

We often use testing to quantify how much knowledge students have gained. However, tests can also contribute to learning. This testing effect is defined in early psychology literature as a boost in learning due to actively retrieving information for some sort of test. Results consistently show that any form of testing is better than rereading or restudying.

Flavours of testing

There are two flavours of the testing effect. Testing previously studied information improves retention of that information. This is known as the ‘backward testing effect’. The ‘forward testing effect’ refers to retention of new information through testing.

The forward testing effect also means that testing can make students better at retrieving information, whether or not that specific information was on the test. This isn’t fully understood, but seems to hold out in practice.

Testing in chemistry

Researchers at the University of Florida in the US examined how the forward testing effect enhances students’ learning of chemistry. Few previous investigations have studied the forward testing effect in chemistry.

Significantly, the researchers carried out their investigation in the classroom, rather than in laboratories where much of the research into testing effects is done. They presented 146 university students on a general chemistry course with four term-time exams. Three retrieval quizzes followed each of the exams. The quizzes were based on the topics of the previous exam and participation in the quizzes was voluntary. There was also a final course exam.

Students who completed more than 50% of the retrieval quizzes performed significantly better on the final exam than those who were at below 50% participation. There was no difference in performance between students who completed slightly more than 50% or much more than 50% of the quizzes. Importantly, students’ prior achievement levels had no impact on whether they chose to participate in the quizzes. This suggests the method would not introduce further biases between students with different abilities.