CogBlog – A Cognitive Psychology Blog

For many students and faculty alike, testing is often considered a necessary evil in learning contexts. Tests – for students – are stressful, requiring hours of preparation, and may feel like a hurdle to be cleared. For many, testing is seen primarily as an assessment tool – it determines a grade, can have powerful and long-reaching implications on an individual’s future success and career options. Failing a test can have negative effects on academic performance and also on a student’s sense of worth. All in all, tests are typically seen as a rather negative event (unpublished data in our lab indicated an average rating of 3.04 on a 1-7 scale, where 1 was ‘extremely negative’ and 7 was ‘extremely positive’ for the word test).

For faculty, tests are often not held in much higher consideration. Tests can ‘take time away’ from more productive pursuits such as lectures or discussions. Preparing and grading a test is extremely time-consuming – thus pushing many faculty to opt for multiple choice questions that can be graded quickly and easily. Small wonder, then, that many college courses only include 2 or 3 exams over the course of a semester.

But what are some of the consequences of these choices about when and how to test? Among other things, presumably, high-stakes, infrequent testing places a greater burden on students because they cover more material; fewer tests probably results in each test being ‘worth more’ in terms of final assessments – thus, the cost associated with failure is much higher; infrequent testing does not allow students and faculty to regularly monitor learning, comprehension, and mastery of material; and, from a practical perspective, a first major test may occur too late in the semester for a struggling student to drop a course or seek effective support. In order to bring about positive change in this culture of “testing = BAD” we need to understand more about what tests can do to promote academic success and healthier, more engaged learning communities.

There has been a recent focus in the field of memory research on the potential benefits of testing memory (see Roediger & Karpicke, 2006; Roediger & Butler, 2011 for recent reviews). The benefits of testing identified thus far include direct and indirect effects. Let’s consider the indirect effects first. More frequent tests 1) encourage students to study more regularly; 2) help prevent falling behind with reading or other coursework; 3) allow students to learn from their performance and the feedback they receive; 4) allow instructors to assess their teaching effectiveness and where gaps in knowledge are occurring; 5) promote distributed learning and reviewing of material, which is far more effective than massed learning (AKA, cramming) for long-term retention.

The direct effects of testing operate on slightly different levels. Namely, these seem to exert their influence more specifically on how learning and remembering occur. Testing – such as a low-stakes quiz prepared by an instructor – administered shortly after learning results in better performance on a delayed test compared to repeated study or re-reading of the same material. Thus, testing seems to improve memory – even when factors such as time spent and the amount of material tested or reviewed are equivalent. To date, numerous studies have reported that taking tests over studied material improves performance compared to re-studying the same material. Taking a test has been shown to improve memory in middle school children (e.g., McDaniel, Agarwal, Huelser, McDermott, & roediger, 2011) and older adults (Tse, Balota, & Roediger, 2010) in addition to college students. The benefits occur with word pairs, face-name pairs, prose passages, complex visual stimuli like chinese characters or maps (Kang, 2010), functions (Kang, McDaniel, & Pashler, 2011), among other materials.

There are at least two potential factors involved in the effectiveness of testing. The “testing effect” proposes that taking a test directly improves memory for the items that have been successfully recalled on a test. Although this is likely an important factor, one can also observe testing benefits for non-tested information (Chan, 2010), indicating that successful recall is not necessary for testing to enhance memory. A second likely “culprit” in the testing effect is that taking a test improves learning after the test. This “test-potentiated learning” was recently investigated by Arnold and McDermott (in press, Psychonomic Bulletin and Review). In two experiments, they found that the benefits of studying the same material (in this case, pictures of simple objects) a second time only appeared when the second study phase was preceded by a test. Thus, taking a test allowed for more effective study the second time around. Szpunar, McDermott, and Roediger (2008) reported a similar finding: Taking a test after learning some material made learning new material easier, by reducing proactive interference (the phenomenon that refers to the fact that learning new information can be made more difficult by a prior learning event).

Given the building evidence for the benefits of tests for achieving long-lasting learning goals, it makes sense to test more, not less. Testing helps create and promote long-term learning.