You are viewing a read-only archive of the Blogs.Harvard network. Learn more.

Does Teaching for Understanding Improve Student Achievement?

No wide-scale, experimental studies have been conducted of the Teaching for Understanding framework, but several wide-scale studies of constructivist teaching have made connections between instructional practices and student gains on standardized tests. Among the most recent and rigorous is that of Martin Abbott and Jeffrey Fouts of the Washington School Research Center. In their study, Constructivist Teaching and Student Achievement, Abbott and Fouts conducted observations of 669 classrooms in 2000-2001 and found, after controlling for income, that constructivist teaching was clearly correlated with higher student achievement on standardized tests in reading, writing, and math. While analyses such as this are seldom straightforward, these researchers felt that the benefit of constructivist teaching would have shown up as even greater if certain study variables had been aggregated in different ways. We’d add that more precise measurement of what constitutes constructivist teaching would have strengthened the correlations as well. In the National Research Council’s synthesis of research around student learning in history, math, and science, How Students Learn, the authors write the following:

. . . in some cases there is evidence that teaching for understanding can increase scores on standardized measures (e.g., Resnick et al., 1991); in other cases, scores on standardized tests are unaffected, but the students show sizable advantages on assessments that are sensitive to their comprehension and understanding rather than reflecting sheer memorization (e.g., Carpenter et al., 1996; Secules et al., 1997). (p. 177)

Relevant here is the statement by leading educational evaluator Jim Popham, of UCLA, that “the vast majority of [state standardized tests] are instructionally insensitive—that is, they’re unable to detect even striking instructional improvements when such improvements occur” [1]. Since arguably the most prominent trend in educational testing in the last five years has been greater demands for tests that tap into higher order thinking, we may soon start to see research results that show greater effects from instructional changes such as the adoption of the Teaching for Understanding approach. Other reports that have received attention in linking constructivist teaching to student outcomes include:


[1] Popham, James W. (2006) “Assessment for Learning: An Endangered Species?” Educational Leadership. (63)5. p. 82-83


1 Comment

  1. Roland Stark

    October 31, 2006 @ 5:56 pm


    I checked into the three sources cited by How Students Learn that deal with the effects of constructivist teaching on either standardized or alternative test scores. I’m dissatisfied with all three as erstwhile sources of evidence. Lauren Resnick, Victoria Bill, Sharon Lesgold and Mary Leer (“Thinking in Arithmetic Class;” pp. 27-53 of Barbara Means’ Teaching Advanced Skills to At-risk Students [Jossey-Bass, 1991]) offer intriguing findings on young children’s math gains, but their report is sparse: they don’t even tell the size of their sample, let alone say anything about the selection or comparability of the different student groups. In short, though I find their approach to teaching math appealing, their write-up doesn’t convince a skeptic like me about their results. T. Secules et al. (“Schools for Thought,” Educational Leadership 54: 6, 56-60, 1997) provide even fewer details and nothing about any controls that might have been applied in their Schools for Thought experiment. They refer to a Helen Bateman of Vanderbilt U who was to produce an independent report on their study, but on the Web I was unable to locate any mention of such a report. Finally, T. Carpenter et al. (“Cognitively Guided Instruction,” in Elementary School Journal 97:1, pp. 3-20, 1996) propose an educational model and do not offer empirical results. I think if we’re going to find more substantial evidence of constructivism’s benefits, we’ll need to wait on some of the current research efforts spawned by, or married to, efforts to test children more frequently and more authentically.

Log in