Old Faithful but not El Capitan: multiple-choice pretesting helps learning about cronartium ribicola
Thursday, August 11, 2016
Old Faithful…. Hayden Valley…. Mammoth Hot springs. Clearly I am talking about the natural wonders of Yellowstone National Park. Now try and answer the following question: What state does the majority of the park reside in? The choices are: (a) Idaho; (b)South Dakota; (c) Wyoming; (d) Montana.
If you correctly picked Wyoming, you clearly know too much. Try another question: As of 2007, which tree in Yellowstone, found only in scattered groves, is threatened by a specific fungus? Gotcha.
What are the effects of answering such questions on subsequent learning of the information?
Perhaps surprisingly, there is considerable evidence that pretesting of material facilitates subsequent learning, even if you cannot initially retrieve (or even choose) the correct answer. You will learn more about cronartium ribicola, the white pine blister rust that has been affecting whitebark pine in Yellowstone since 2007, if you struggle to answer a question about it first.
This finding has clear implications for teaching, and it is perhaps not surprising that many university instructors now pretest their students at the beginning of a class, perhaps usingclickers, before the material is presented.
Although this pretesting effect appears to be robust and replicable, it is less clear whether pretesting also facilitates the learning of information that was not itself pretested. Does pretesting help this information, too, or might it interfere with its acquisition?
The idea of interference is not entirely far-fetched, given that students need to absorb large amounts of related and thus potentially confusable information: Being pretested on the characteristics of short-term memory in a psychology class might therefore make learning of the principles of long-term memory more difficult.
A recent article in the Psychonomic Society’s journal Memory & Cognition examined this question. Researchers Little and Bjork were particularly interested in comparing the benefits of multiple-choice pretests (such as the first item above about the state that Yellowstone resides in) to those resulting from cued-recall pretests (such as the above question about the fungus). The relatedness of the critical information was defined in terms of its competitiveness to the pretested material.
The design of their main experiment is shown in the figure below:
In the figure, MC stands for multiple choice and CR for cued recall. Participants were randomly assigned to the sequence on the left or on the right. The experimental design permitted Little and Bjork to compare the effects of pretesting within participants, by comparing performance on passage A (after pretest) with performance on passage B (no pretest but study of facts containing comparable information) on the final cued recall test. A third passage, C, served as control and involved extended reading only.
In addition, between-participant comparisons could be performed between the type of pretest (multiple choice vs. cued recall), and whether or not the fact study for passage B involved competitors from passage A. Of course, across participants, the identity of the passages and the order of presentation was counterbalanced.
To illustrate the principal manipulation, each passage—for example about Yellowstone—contained two paired facts on the same topic (e.g., geysers) that could be tested using the same four multiple-choice alternatives (e.g., Old Faithful, Steamboat Geyser, Castle Geyser, and Daisy Geyser). If that passage was used with a pretest (i.e., passage A in the figure above), one of the pair members would be used on the pretest (e.g., What is the tallest geyser in Yellowstone National Park? Answer: Steamboat Geyser), and the other, related pair member would be used on the final test (e.g., What geyser is thought to be the oldest in the world? Answer: Castle Geyser).
Conversely, when the Yellowstone passage was preceded not by a pretest but by the study of facts (i.e., it served as passage B above), the factual statements containing competitors included information that would have been presented as incorrect alternatives for the corresponding multiple-choice question (e.g., Castle Geyser is thought to be the oldest geyser in the world, not Old Faithful, Steamboat Geyser, or Daisy Geyser). If the passage served in the condition without competitors, the factual statements were the same except for the absence of the alternatives (e.g., Castle Geyser is thought to be the oldest geyser in the world).
The figure below provides a first snapshot of the results. Final recall performance was better after pretesting than after study of the same facts. (The difference between the two olive bars is slight but was statistically significant.) In addition, the type of pretest did not appear to matter. Performance was virtually identical after multiple-choice pretesting and cued-recall pretest.
The next figure shows the effects of pretesting on related material—that is, material that was not itself pretested but was related to the pretest.
The figure shows that information related to the pretest (the dark olive bars) was recalled better when it followed multiple-choice pretesting than when it followed cued-recall pretesting, and this was confirmed by statistical test.
Indeed, a cued-recall pretest had no greater benefit for related information than study of the same information as facts, and both led to worse performance than the control condition in which participants merely spent their time reading the passage. Little and Bjork conclude that “being exposed to incorrect alternatives associated with related information in the context of a multiple-choice question induces processing that improves later learning of that related information beyond that provided by a cued-recall question.”
Crucially, this result does not arise from mere exposure to the related information at pretest: Studying the same facts, including the potentially competing multiple-choice alternatives, does not provide the same benefits. Clearly, the active differentiation between the correct response to a multiple-choice item and the other alternatives that occurs during an MC pretest sets the stage for subsequent learning particularly well.
So, if you are one of those instructors who likes to pretest a class with a clicker, you can now be assured that you are boosting your students’ learning outcome.
Article focused on in this post:
Little, J. L., & Bjork, E. L. (2016). Multiple-choice pretesting potentiates learning of related information. Memory & Cognition. DOI: 10.3758/s13421-016-0621-z.