How to ace your next exam

It generally assumed that the way to improve your performance on a test of memory is to study as much as possible. But research has shown that there's something even better than studying: testing. The testing effect is the idea that being tested on material actually improves retention of that material. The testing effect is well-known in the cognitive psychology literature, but often seems under-appreciated in the field of education. Perhaps this is because although a short-term testing effect is well established, longer-term results are more dubious.

Larsen, Butler and Roediger (2009) wrote an article entitled "Repeated testing improves long-term retention relative to repeated study: a randomised controlled trial" in which they assess whether the testing effect holds over an extended period of time. They use a more realistic setup than some of the previous studies have utilized by training and testing medical residents on knowledge of two neurological situations.

In this study, the authors had 40 participants learn about the treatment of status epilepticus, and the diagnosis and treatment of myasthenia gravis. All of the participants took part in a highly-interactive one-hour teaching session that covered both topics. The residents were then assigned to two different groups.

In the first group, residents were repeatedly tested on status epilepticus and they repeatedly studied myasthenia gravis (testing SE, studying MG group). After the teaching session, this group took a test on the treatment of status epilepticus and studied a review sheet on myasthenia gravis. The tests were composed of short-answer questions and the review sheets consisted of identical information to that on the test answer sheets.

The second group was repeatedly tested over myasthenia gravis and they repeatedly studied status epilepticus (testing MG, studying SE group). After the teaching session, this group took a test on the treatment and diagnosis of myasthenia gravis and studied a review sheet on status epilepticus. After each test, the participants were allowed to check their answers.

Two weeks after the teaching session and the initial testing/studying, the participants repeated these tasks. The testing SE, studying MG group took a test on the treatment of status epilepticus and studied a review sheet on myasthenia gravis. The testing MG, studying SE group took a test on the treatment and diagnosis of myasthenia gravis and studied a review sheet on status epilepticus. This process was repeated again after another two weeks.

Finally, approximately six months after the testing session, all of the participants in both groups took a final test on both topics. The results are displayed below.






As you can see, for myasthenia gravis, the group that was repeatedly tested on MG outperformed the group that repeatedly studied MG on the final test by 17%. For status epilepticus, the group that was repeatedly tested on SE outperformed the group that repeatedly studied SE by 11%.

The results of this study show that testing not only serves as a tool for assessment, but additionally promotes learning.

Larsen, D.P., Butler, A.C., & Roediger, H.L. (2009). Repeated testing improves long-term retention relative to repeated study: a randomised controlled trial. Medical Education, 43, 1174-1181.

Additional studies on this topic:

Butler, A.C., & Roediger, H.L. (2007). Testing improves long-term retention in a simulated classroom setting. European Journal of Cognitive Psychology, 19(4/5), 514-527.

Karpicke, J.D., & Roediger, H.L. (2008). The critical importance of retrieval for learning. Science, 319, 966-968.

Roediger, H.L., & Karpicke, J.D. (2006). The power of testing memory: Basic research and implications for educational practice. Perspectives on Psychological Science, 1(3), 181-210.

The advantage to being bilingual and how it arises

People often ask whether growing up bilingual can put a child at a disadvantage. Will the child develop slower, and ultimately, know each language less well? Indeed, Ellen Bialystok, a leading researcher on the cognitive effects of bilingualism, points out that bilingual children have been shown to command smaller vocabularies in each of their two languages. In adults, the disadvantage persists, but in the form of a deficit in lexical retrieval - or how quickly you can access vocabulary words.

However, in a recent article entitled The source of enhanced cognitive control in bilinguals (2008), Ellen Bialystok and three other researchers discuss the cognitive advantage to being bilingual: enhanced executive control (as compared to monolinguals). According to Miyake et al. (2000), executive control consists of three frequently-cited functions: shifting mental states, monitoring short term memory representations, and inhibiting practiced responses. In the present article (hereto referred to as Emmorey et al., 2008), the authors specifically focus on inhibiting practiced responses. This simply means that you've been engaging in Task A, and in order to engage in Task B, you must first prevent yourself from engaging in Task A again. Emmorey et al. argue that bilinguals spend much of the time inhibiting one language in order to speak or comprehend the other language, and this is why they show enhanced executive control.

It should be noted that this conclusion is limited to unimodal bilinguals - bilinguals whose two languages use the same modality (namely, speech). Bimodal bilinguals, on the other hand, know both a spoken language and sign language. The interesting part about this distinction is that bimodal bilinguals do not need to inhibit one language in order to communicate with the other. They can technically "speak" both at the same time. Thus, if they (like unimodal bilinguals) display an advantage with regard to inhibiting practiced responses, then it shows that this advantage in bilinguals does not stem from the extra time spent inhibiting one language while using the other.

To restate: unimodal bilingual individuals show enhanced cognitive control compared to monolinguals. However, the cause remains unknown. If it is due to the practiced inhibition of one language in order to communicate with the other, then only unimodal bilinguals will show an advantage. If, however, it is due to something else, both unimodal and bimodal bilinguals will show an advantage. Note that not showing an advantage would mean performing on par with monolinguals on the executive control task.

In order to get at this distinction, Emmorey et al. tested unimodal bilinguals, bimodal bilinguals, and monolinguals on a task known as the "flanker task". In this task, participants must say which direction a particular arrow is pointing within an array of arrows. The diagram below shows an example.


Sometimes, as in the above example, the task is relatively easy. However, as in the diagram below, sometimes it's more difficult. This is because all of the other arrows (the flankers) are facing in the opposite direction, making you want to say that direction (in this case, right).

The former example is called the "congruent" condition and the latter is called the "incongruent" condition. Emmorey et al. added two more conditions to this, known as the "go" and "no-go" conditions. In the go condition (on the left, below), participants were asked to perform the same task as above, but the central red arrow was in the middle of two red diamonds. In the no-go condition (on the right, below), the central red arrow was in the middle of two red Xs, and participants were told that they must inhibit responses to this type of trial. Finally, there was a control condition, which consisted of just one arrow with no flankers.

The graph below summarizes Emmorey et al.'s findings.

The y-axis is the reaction times of the participants. A higher number means that participants were slower to react. A lower number means that they were faster, or more successful, at the task. As you can see, the unimodal bilingual individuals were fastest to react on the go, congruent, and incongruent trial types. (Note that the no-go trials are not displayed in the graph because participants were asked to inhibit a response.) In contrast, the bimodal bilinguals performed at the same speed as the monolinguals. Thus only the unimodal bilinguals showed an advantage. These results provide strong evidence for the claim that the bilingual executive control advantage (or really, the unimodal bilingual advantage) arises from extra practice at inhibiting one language while using the other.

Bialystok, E. (2009). Bilingualism: The good, the bad, and the indifferent. Bilingualism: Language and Cognition, 12(1), 3-11.

Emmorey, K., Luk, G., Pyers, J.E., & Bialystok, E. (2008). The source of enhanced cognitive control in bilinguals. Psychological Science, 19(12), 1201-1206.

Miyake, A., Friedman, N.P., Emerson, M.J., Witzki, A.H., Howerter, A., & Wager, T.D. (2000). The unity and diversity of executive functions and their contributions to complex "frontal lobe" tasks: a latent variable analysis. Cognitive Psychology, 41, 49-100.

English and Mandarin Chinese speakers conceive of time differently

One of the greatest questions of psycholinguistics is whether the language that one speaks determines the thoughts one can have -- this idea is known as linguistic determinism. Perhaps the most famous proponent of linguistic determinism was Benjamin Lee Whorf. In 1940, Whorf published an article entitled "Science and Linguistics." In this article, he discussed many examples of language determining thought, stating that "we dissect nature along lines laid down by our native languages." Many of these examples are still taught today in introductory linguistics classes. For example, Whorf argued that Eskimos have seven words for snow, while English only has one, and therefore, Eskimos could perceive more types of snow. This has since shown to be wrong -- if you're interested in reading more on this topic, check out Geoffrey Pullum's book "The Great Eskimo Vocabulary Hoax" (see full citation below).

Linguistic determinism has largely been discredited. The few examples that were proposed were found to be invalid. For example, it was suggested that individuals from cultures with fewer color terms than English could not distinguish between other colors. (If you only have words for black, white, and red, then you can't distinguish between blue and green). This was shown to not be the case -- color labels don't determine the colors you can perceive. What remains today of Whorf's linguistic determinism is a watered-down version called linguistic relativity. According to linguistic relativity, language doesn't determine thought, but it does affect it in more subtle ways. This brings us to the actual topic of today's post: Boroditsky, Fuhrman, and McCormick's article, "Do English and Mandarin speakers think about time differently?" In this article, the authors provide evidence for linguistic relativity by showing that manipulating linguistic context can actually affect an individual's behavior.

In English, metaphors about time involve front/back spatial terms. For example, we might say "the future is in front of us" or "the past is behind us." In Mandarin Chinese, time metaphors are both horizontal and vertical. For example, Mandarin uses the word "qian" (in front of) to describe time. One might say "qian yi tian" to mean before today (literally "in front one day"). Mandarin also uses the words "xia" (down) and "shang" (up) to describe time. For example, one might say "xia ge yue" to mean next month (literally, "down month") or "shang ge yue" to mean last month (literally, "up month").

The question that the authors wondered is whether these different linguistic metaphors for time might affect how English-speakers and Mandarin-speakers think about time. To study this, Boroditsky, Fuhrman, and McCormick manipulated participants' linguistic context. They gave English and Mandarin speakers keyboards with labels for earlier and later and had them judge whether a picture occurred earlier or later than another. For example, they showed a picture of a mug filled with coffee and a picture of coffee being poured into a mug.

The participants were always given a keyboard with two keys that, they were told, signified earlier and later. What was different is that some subjects used a keyboard where the earlier and later keys were next to each other ("horizontal setup"), and some used a keyboard where the keys were above and below each other ("vertical setup"). In the vertical setup, the keyboard was perpendicular to the desktop, so that the keys were literally vertically related to one another. Furthermore, in the horizontal setup, sometimes earlier was to the left of later, and sometimes to the right. In the vertical setup, sometimes earlier was above later, and sometimes it was below.

If it is the case that language affects thought, then for English speakers, the horizontal setup in which earlier is to the right of later would be more difficult than the condition with earlier to the left of later. This is because it contradicts the typical English time metaphors. However, for the vertical setup, there would be no effect of condition because English doesn't use vertical time metaphors.

For Mandarin speakers, both the horizontal and vertical setups should show different levels of difficulty. Like the English speakers, for the horizontal setup, the condition in which earlier was to the right of later would be more difficult than the condition with earlier to the left of later. For the vertical setup, Mandarin speakers would find the condition with earlier above later easier than the opposite condition.


The graph above shows that the authors' predictions regarding the English speakers were confirmed. In the study, the authors used reaction times as measurements of how easy or difficult a task was. The faster one reacts, the easier the task is. You can see from the graph that the reaction times for the vertical setup (red) were about the same in each condition. However, participants were much slower to respond when earlier was to the right than when earlier was to the left (blue), indicating that the former was more difficult than the latter.


The above graph shows that the authors' predictions regarding the Mandarin speakers were also confirmed. You can see from the graph that, like the English speakers, Mandarin-speaking participants were much slower to respond when earlier was to the right than when earlier was to the left (blue), indicating that the former was more difficult than the latter. But in addition, in the vertical setup, participants were also much slower to respond when earlier was down, than when it was up (red).

Together, these results show that time metaphors actually influence the way people conceive of time. They provide support for linguistic relativism, though not full-fledged linguistic determinism.

Boroditsky, L., Fuhrman, O., & McCormick, K. (2010). Do English and Mandarin speakers think about time differently? Cognition, doi:10.1016/j.cognition.2010.09.010.

Pullum, G.K. (1991). The great Eskimo vocabulary hoax and other irreverent essays on the study of language. University of Chicago Press.

Whorf, B. L. (1940). Science and linguistics. Technology Review, 42(6), 229-31, 247-8.