Cognition & Education

 

Check out the new Center for Integrative Research on Cognition, Learning, and Education (CIRCLE).


Our work on cognition and education has focused on four basic determinants of student learning: study activities, individual differences, text structure/properties, and testing.  Our experimental work is designed to reveal how student learning is affected and can be improved by variation in these determinants.


STUDY STRATEGIES


Rereading:  Popular but Ineffective.  Past progress reports have described the results of three experiments with textbook chapters showing that reading twice produced no better learning than reading once.  This past year these experiments were presented at the American Educational Research Association annual meeting (April, 2007) and a manuscript based on this research was submitted for publication to Journal of Educational Psychology.


Generative Study Strategies.  This line of work, conducted in collaboration with Dan Howard (a graduate student in McDaniel’s lab), is investigating the effectiveness of using various generative study strategies for retention of educational information.  In a new experiment completed this past project year, participants studied an educational text conveying the technical operation of a particular mechanical device (e.g., a bicycle pump; car brakes).  Three study conditions were implemented: a reread control, reading with note taking, and reading with the 3R method (read the article, recall as much of the article as possible, and then reread the article). This 3R method has fallen out of favor in educational practice, although theoretically it should be a potent study method.  Performance on free recall, multiple choice, and inference tests were compared between the conditions, both immediately after study and after a one-week delay.  Use of the 3R strategy resulted in significantly improved performance relative to the reread control on all three tests at both the immediate and delayed retention intervals.  The 3R strategy also  improved performance on free recall relative to note-taking. 


INDIVIDUAL DIFFERENCES


Structure Building, Quizzing, and Concept Mapping. We continued to investigate the effect of individual differences in students’ ability to build coherent cognitive structures of presented information (structure building) on learning from textbooks and to develop techniques to improve learning and retention for low structure builders.  With Aimee Callender and Kathy Wildman (graduate students in McDaniel’s lab), we completed a study in which some experimental groups received a text and either had to answer fill-in-the-blank questions while reading the text or fill in concept maps of the text.  Fill-in-the blank questions showed little benefit on multiple choice and recall tests relative to a read-only control.  Low structure builders’ recall (but not multiple choice performance) was aided by concept maps. Unexpectedly, concept maps produced benefits to multiple choice and recall performances for high-structure builders. 

Consequently, we are now attempting to produce more robust benefits for lower-ability readers. We have designed  different mapping procedures, which we will apply to different kinds of texts.  In addition, a new study in development will investigate concept mapping not only as a study method, but also as a criterial test.  Because concept mapping presumably reveals the organizational structure of a representation, we will investigate if readers of differing levels of structure building ability show dramatically different performances on a concept mapping test.


Structure Building and Working MemoryTo further investigate the factors that contribute to structure building ability, we have designed a new experiment that will investigate the relationship between working memory ability and structure building ability, as measured by the MultiMedia Comprehension Battery (MMCB).  One question we are addressing is what cognitive components are involved in structure building.  One candidate is working memory.  The idea here is that larger working memory capacity may be associated with inhibiting irrelevant meaning and establishing richer linkages among related information, thereby creating coherent representations of the text.


Reader Appropriate DifficultyThe research outline above has focused on one measure of reading ability, structure building ability.  Other factors are also important, including decoding, or reading fluency.  In this study (conducted by Aimee Callender), participants took two reading measures, one assessing structure building and one assessing decoding ability.    compared study methods that should help a low structure builder better organize the text (outlining) against those that should affect one’s fluency while reading (filling in deleted letters). Participants also completed various processing tasks with texts, tasks designed to either help readers better organize the text (outlining) or more elaboratively process individual propositions of the text (filling in missing letters).  Initial results suggest that those who score either very high or very low on both reading measures are not affected by the study tasks, with the highly competent readers remembering significantly more information than the low-ability readers.  However, those who are good structure builders but have some difficulty with decoding benefit from a task which focuses their attention on individual propositions in the text (letter deletion task).  It appears that improving learning with specified study tasks requires consideration of specific combinations of reading abilities (see McDaniel & Callender, in press, for development of a framework that helps prescribe optimal matches among study strategies, reader characteristics, and text characteristics).  This work will be presented at a special session at APA, August, 2007. 


TEXT FORMAT


Interactive—Visual “Texts”. Mark McDaniel, Keith Lyle (a post-doc in McDaniel’s and Roediger’s lab), Andrea Young (an undergraduate researcher in McDaniel’s lab) have collaborated with Robin Heyden, an author of high school and college biology textbooks, to investigate how different text formats affect learning of a real-world biology lesson.  Robin Heyden is authoring a computer-based biology textbook called iBiology, which augments standard text with embedded activities that depict concepts visually and require active generation of information by learners.  This past project period we completed two experiments contrasting the effectiveness of iBiology with a parallel a text-only version that includes neither visual displays nor generation and a parallel text with visual version that excludes the interactive exercises (activities are completed by the computer rather than the learner).  In Experiment 1, using non-biology majors, iBiology  produced significantly worse performance on learning definitions, multiple choice questions, and short-answer questions than the text-only version (with processing time controlled for).  However, interestingly, learners preferred the iBiology version.   In contrast, with biology majors, in Experiment 2 all conditions produced equivalent definition learning and multiple choice performance. Both conditions with visual presentations (iBiology and the passive visual version)  improved short-answer performance relative to the text-only conditions.  A tentative and perhaps surprising conclusion is that the effectiveness of demanding visuals in textbooks may be limited to those learners with more background knowledge.



TEST ENHANCED LEARNING


See The Test-Enhanced Learning in the Classroom Project here


Level of Test Questions.  Much of our past work (see also Roediger’s, McDermott’s, Bjork’s and Marsh’s projects) has established the potent benefits of tests as a learning activity.  One theoretical distinction reflected our present work  is the level of the criterial information.  Summative tests can focus on “factual” or on “conceptual” information. Kathy Wildman and I are investigating the testing effect for retention of different kinds of textual content—factual and conceptual—and the influence of different testing formats (multiple choice and short answer) on the testing effect for factual versus conceptual information.  Experiment 1 has just been completed, and preliminary analyses show that testing effects are more robust for conceptual information than for factual information.  This work will be presented at a special session at APA, August, 2007.


Test Enhanced Learning in the ClassroomJeanine Sun (an undergraduate student), Aimee Callender, and I have investigated the effects of quizzing in an introductory psychology class.  Students in this course took a quiz on 10 of the chapters covered in the course.  Questions on the quizzes also appeared on the final exam either in the same format and wording, or using different wording.  Additionally, some information on each exam was not included on the quizzes at all.  Data collection for this experiment is complete and analyses are underway.  Jeanine Sun will continue this line of research in additional large-enrollment classrooms in the fall semester.


Want to Learn More?


If you are interested in learning more see the publications page. If you would like to get involved in this lab’s education research contact principal investigator Dr. Mark McDaniel (mmcdanie@artsci.wustl.edu).