|
|
|
|
|
|
|
Teach Talk
Over the last several years, researchers in what is now called learning science, which includes work in such fields as cognitive science, developmental psychology, neuroscience, and cross-cultural anthropology, have begun to make significant progress in understanding how learning takes place at the postsecondary level. They are creating an increasingly sophisticated picture of what happens in the college classroom from the perspective of both student and instructor. And they are describing with impressive detail how the system created by the interaction of learners, teachers, technology, and discipline operates.
In all this, one finding, in particular, is so consistent that it should be singled out for special attention. Put simply, it is this: Researchers have seen that when students themselves are actively involved in the learning process, their learning improves. This finding, far from being an abstraction of interest only to a small group of academicians in the field, has direct consequences for what we do in the classroom.
In other "Teach Talks," I have alluded to the positive effects that actively engaging students in their learning can have (see, for example, "Newsletter Introduces New Regular Feature," February/March 1995; "What the Students Say," November/December 1997; and "Teaching Teamwork Skills," January/February and, March/April 1998). So important is the finding that active learning contributes to students’ improved performance, that the next three "Teach Talks" will be devoted to exploring this topic. In this column, I will present the data that underlies the claim for the efficacy of active learning techniques. In the next "Teach Talk," I will outline some specific ways science and engineering instructors can bring active learning into the classroom. And the third column in the series will describe specific ways active learning techniques are being used in MIT classrooms. This full picture of active learning will hopefully prompt some of you to experiment with it in your classrooms in ways you see fit.
But before I begin, a definition of terms and a disclaimer.
The phrase "active learning" – like many terms that have the misfortune of being too popular and too widely used – has taken on a number of meanings depending on the context in which it is being used and who is using it. It is sometimes used interchangeably with terms like "collaborative learning" or "cooperative learning," although I would argue that both are subsets of active learning. (More on this in the next "Teach Talk.") It can encompass a range of activities, from having students discuss a problem or a concept with one another during class to having them work on semester-long (or several semesters-long) design projects in teams. In this series, I am going to use "active learning" to describe both an approach to learning and a variety of techniques that are used if one employs that approach.
Active learning means, basically, that students are involved in some kind of guided activity in class, so that they are doing something in the classroom besides sitting and listening to the instructor give a lecture or watching him/her work problems on the board. This definition has two implications:
Karl Smith*, a professor of civil engineering at the University of Minnesota and an authority on collaborative learning, uses these two illustrations to show the differences in underlying philosophy between "traditional" learning and active learning. In the classroom as we know it, instructors try to "pour" knowledge into the heads of their students. Teachers who use active learning believe that knowledge can best be gained through the interaction of students, not only with them, but also with one another and with the material being taught.
The richest definition of active learning I have found comes from Richard Hake, a professor of physics at Indiana University. Hake, who uses the term "interactive engagement," writes that IE methods are "designed in part to promote conceptual understanding through interactive engagement of students in heads-on (always) and hands-on (usually) activities which yield immediate feedback through discussion with peers and/or instructors." (p. 65). ("Interactive-engagement versus traditional methods: A six-thousand-student survey of mechanics test data for introductory physics courses," American Journal of Physics, 66, 64-74, 1998.)
This does not mean, by the way, that students and teacher have equal roles in the classroom. A basketball coach and his/her players are not at the same level within the hierarchy of the team. But no one could argue that basketball players can become proficient at playing the game only by listening to the coach talk about it. They have to get out on the court, they have to practice with their teammates, and they have to play in front of the coach, so they can get feedback on their performance. Like athletes, students have to do physics or mathematics, chemistry or biology in order to master it. Doing it in front of both "teammates" and the "coach" in order to practice and receive feedback improves the student’s ability to understand and apply the crucial concepts of the discipline.
Now to the disclaimer. I am not advocating – although some people would – that active learning should replace more traditional approaches, like lectures, across the board. Lectures have their advantages, and have their place in a methodological toolbox. Nonetheless, the research makes a fairly compelling case that some kind of involvement with course content during class time leads to gains in learning not seen using more conventional approaches. Let me, then, turn to those studies themselves.
Meta-Analysis Finds Gains in Achievement, Persistence, and Attitude
One of the most persuasive studies comes from the National Institute for Science Education (NISE). Based at the University of Wisconsin - Madison and funded by the NSF, NISE is a group of faculty, researchers, and faculty developers who work to improve science, mathematics, and engineering education through research, publications, workshops, and assessment projects. In 1997, NISE published a research monograph, "Effects of Small-Group Learning on Undergraduates in Science, Mathematics, Engineering, and Technology: A Meta Analysis," by Leonard Springer, Mary Elizabeth Stanne, and Samuel S. Donovan.
Springer et al.’s meta-analysis is of 39 high-quality studies on the use of small-group learning. The original pool of studies was 383. In order to be included in the meta-analysis, a study had to: (1) examine undergraduates from an accredited postsecondary institution in North America; (2) look at small-group work among two to ten students; (3) have been conducted in an actual classroom; (4) have been published or reported in 1980 or later; (5) report enough statistical information to estimate effects size.
The metric used was the standardized mean difference (d-index) effect size. For two-sample analyses, effect size was calculated by subtracting the control group’s average score from the experimental group’s average score and dividing the difference by the average of two standard deviations. For single-sample analyses, the average score on a pre-test was subtracted from the average score on a post-test, and the difference was divided by the average of two standard deviations.
The authors write, "The main effect of small-group learning on achievement, persistence, and attitudes among undergraduates in SMET [science, mathematics, engineering, and technology] was significant and positive" (p. 7). Students who learned in small groups demonstrated greater achievement (d=0.51), persisted to a greater extent through SMET courses (d=0.46), and expressed more favorable attitudes towards their courses (d=0.55) than students who did not work collaboratively or cooperatively (see graphs).
To put these findings into perspective: The average effect size for all classroom-based interventions on student achievement is 0.40 as compared to 0.51 for small-group learning. The 0.51 effect size in achievement would move a student from the 50th percentile to the 70th on a standardized test. The 0.46 effect size for persistence is enough to reduce attrition in a SMET course or program by 22%. And the 0.55 effect size for attitudes far exceeds the average effect of 0.28 for all classroom-based interventions.
Physics Studies Also Report Positive Results
A significant body of research into learning has been done in physics, including, notable work by Lillian McDermott at the University of Washington, Edward Redish* at the University of Maryland, Richard Hake at Indiana University, and Eric Mazur at Harvard. (Each, by the way, is a physicist who has chosen to do research in physics education as well as research within the field itself.) The findings of the studies echo those of the studies in the Springer meta-analysis: Gains in learning are correlated with the use of active engagement methods.
Edward Redish, Jeffrey M. Saul, and Richard N. Steinberg examined 11 different introductory calculus-based mechanics classes for engineering students at the University of Maryland and reported their findings in "On the Effectiveness of Active-Engagement Microcomputer-Based Laboratories," which appeared in the American Journal of Physics (65, 45-54, 1997). Taught by six different instructors, each class was composed of three hours of lecture and one hour of a small section each week. However, in six classes the one-hour section was a traditional problem-solving recitation; in five classes the one-hour section was an interactive tutorial using microcomputer-based laboratory equipment.
Students were evaluated using questions from the Force Concept Inventory (FCI), a standardized, well-validated instrument used to measure how well students understand concepts from classical mechanics, and a long-answer examination question. The metric used was the figure of merit (h), which was defined as:
h = class post-test average – class pre-test average/100 – class pre-test average
The tutorial classes systematically produced better overall FCI gains than the non-tutorial classes. The average fractional gains of the classes were:
- <h> = 0.18 (classes with recitations)
- <h> = 0.35 (classes with tutorials)
As the graph shows, every non-tutorial class had a larger h than every non-tutorial class. (Light colored bars are the tutorial classes; dark-colored bars represent the recitation classes. Only 8 classes are depicted because not every class in the 11-class study was given the FCI.)
Results were somewhat disappointing on the long-answer examination question, which asked students to generate a velocity vs. time graph for a complicated situation. However, the tutorial students did better than recitation students. While only 12% of the recitation students were able to draw a correct graph, 22% of the tutorial students were able to do so.
Richard Hake, in the article cited above, examined the effect of using what he calls "interactive engagement" by surveying 62 introductory physics courses, including courses in high schools, colleges, and universities, with a total enrollment of 6542 students. Students were evaluated using the Force Concept Inventory (FCI). The measure of the average effectiveness of a course in promoting conceptualized understanding was taken to be the average normalized gain <g>, which was defined as:
<g> = % <post> - % <pre>/100 - %<pre>
Fourteen traditional courses (N = 2084) that made little or no use of interactive engagement (IE) methods achieved an average gain of 0.23 plus or minus 0.04 std dev. Forty-eight courses (N = 4458) that made substantial use of IE methods achieved an average gain of 0.48 plus or minus 0.14 std dev., almost two standard deviations above that of traditional courses. On the graph (which appears in the Redish study cited above) gains made in traditional courses are represented by the line closest to the x-axis; gains made in courses using IE methods are shown by the middle line; and gains made by IE courses using a research-based text are depicted by the steepest line.
Why Is Active Learning So Effective?
I can’t provide a definitive answer to the question of why we see the results we do when active learning techniques are used, but research in neuroscience and cognitive psychology, which confirms earlier theories put forth by developmental psychology, is beginning to provide us with a glimmer of an explanation.
The picture of the brain emerging from this research is that of an organ continually molding itself by making synaptic connections. Of more importance to the argument here is that this process is influenced through the interaction of the individual with his/her environment. In other words, as John D. Bransford, Ann L. Brown, and Rodney R. Cocking write in their well-received new book, How People Learn: Brain Mind, Experience, and School, "Learning changes this physical structure of the brain" (p. 103). They then go on to write, "These structural changes alter the functional organization of the brain; in other words, learning organizes and reorganizes the brain" (p. 103). While once it may have been thought that brain development occurred only in children before a certain age, we now know that the process of synaptic addition occurs throughout an individual’s life.
And this is a process that unites quantity with quality: that is, the more the brain is stimulated because the individual is in a rich environment, the more brain activity. For example, Bransford and his co-authors, citing a study by J.E. Black and others, report animals raised in complex environments have a greater volume of capillaries per nerve cell than animals raised in sterile cages. ("Complex Experience Promotes Capillary Formation in Young Rate Visual Cortex," Neuroscience Letters, 83, 351-355, 1987). Other studies Bransford, Brown, and Cocking include in their book support this finding that "experience increases the overall quality of functioning of the brain" (p. 107). Here, of course, I am simply scratching the surface of a complex body of research. I would urge you to read How People Learn yourself, not only for its description of research on the brain, but for interesting findings on science teaching as well.
While it may be a leap from animals in stimulus-rich cages to a college classroom, the research that confirms that the brain continues to change throughout the human life span allows me to ask this question: Isn’t it possible that by marrying what we know about the physiology of the brain with the research findings on active learning, we can at least posit a hypothesis for why active learning is such a powerful teaching technique? Active learning allows for more intense engagement on the part of the students which, if the brain research is right, should produce more substantial learning.
Possible Weaknesses in the Argument
When I have presented these findings on active learning in workshops with MIT faculty, several questions and criticisms arise without fail. MIT faculty are a dubious lot, and changes in conventional methods of teaching are not to be taken lightly. Therefore, I think it is important to surface several of these concerns:
There are methodological flaws in the studies cited. Hardly a study in social science goes by without some concern over methodology, and these are no exception. Both Hake and Springer address possible methodological difficulties. These range from the problems associated with meta-analysis due to the fact that journals tend to publish studies that report statistically significant results (Springer, p. 18), to the problem of "teaching to the test" in the studies where gains in the FCI are used to measure effects (Hake, p. 69).
It is impossible for me to address all these methodological concerns here. (Please contact me if you would like copies of the articles, and/or would like to discuss specific problems in methodology.) What I can say is that the authors are aware of these limitations, and have taken care to deal with them within the constraints imposed by social science research. In addition, the studies cited have been well received in the field; the Hake and Redish studies appear in peer reviewed journals.
Aren’t what we are seeing simply another version of the "Hawthorne Effect"? The "Hawthorne Effect" is a term coined from studies undertaken at the Hawthorne Works of the Bell System’s Western Electric plant between 1924 and 1933 on, among other things, worker productivity. The common interpretation of the studies’ results is that when changes are made to an environment in which a group operates, the group’s members will respond positively not because of the intrinsic worth of the intervention, but simply because they are being given special attention.
Alhough there is great debate over exactly what the Hawthorne experiments do show, even if the above interpretation is true, the NSIE study reports a significantly larger positive effect for active learning than other kinds of interventions (p. 17). In other words, even if most every intervention produced positive results (a claim which, in actuality, can’t be made), active learning remains a technique that results in more gains than other techniques. In addition, Hake reports positive gains in courses in which active learning has been employed for a number of years (p. 70).
These findings may be applicable to other college-level students, but MIT students are different. Hake’s is the only study I have read that discusses a correlation between students’ academic ability and their responsiveness to active learning techniques, and it reports higher gains for honors high school students than for those in regular courses. It does make sense to me that highly intelligent students would derive great benefit from the ability to engage with ideas and concepts directly under the watchful eye of a gifted teacher, but that’s only gut-level instinct. It would be an interesting – and important – contribution to the research literature if we could undertake a study to test that hypothesis.
As promised, this "Teach Talk" has examined the more conceptual, research-oriented part of active learning. I hope I have made the case that this is a tool to be used, explored, experimented with. In the next "Teach Talk," I’ll take a more concrete approach, describing the range of techniques that can actually be put into practice in science and engineering classrooms.
|
|
|
|
|
|
|