Honest follow-through needed on this project Imprimer Envoyer
La recherche - Les données probantes
Écrit par Marian Kester Coombs   
Mardi, 30 Avril 2013 14:23

Honest follow-through needed on this project

Marian Kester Coombs
The Washington Time
24 mars 1998

The Washington Times


What if the federal government spent $1 billion over nearly three decades to study thoroughly the question of which teaching method best instills knowledge, cognitive skills and positive self-concept in students?

What if that study were able to conclude exactly which method best does all three?

Wouldn't the American people like to know about it?

Both the study and its conclusion do exist. Project Follow Through, begun in 1967 under President Lyndon Johnson's War on Poverty to “follow through” on Head Start, spent an estimated $1 billion through the Office (now Department) of Education, the Office of Economic Opportunity and dozens of private sponsors to test and evaluate an array of educational methods.

A total of 700,000 students in 170 socioeconomically disadvantaged communities around the country were involved. Parents were asked to decide which model would be adopted at their school, the government then funded the model through such sponsors as universities and private institutes.

The last funds for the study - the largest educational experiment ever undertaken – were disbursed in 1995. What happened then is another study – in the politics of bureaucracy, says Douglas Carnine, professor of education at the University of Oregon.

« The education profession has never been particularly interested in results, especially if they run counter to the prejudices of the profession, » says Mr. Carnine, who was involved with Project Follow Through when his university served as one of its sponsors.

Of the nine education methods evaluated under Project Follow Through, three each fell into one of three types:

  • Basic skills – a semi-behavioristic approach similar to the Suzuki method.
  • Cognitive – a learning to-learn approach that stresses the child's discovery or “construction” of knowledge on his own.
  • Affective – a whole-child approach that aims to boost student self-esteem on the theory that a higher sense of self-worth promotes achievement.

By the mid-'70s, data on me program had been gathered by Stanford Research Institute and analyzed by Abt Associates of Cambridge, Mass. These data were derived from a battery of five tests administered to “cohorts” (followed either from kindergarten through third grade or from first through third grade) of more than 9,000 Follow Through students matched with a control group of 6,500 students from non-Follow Through school sites.

The 11 “outcome measures” assessed by these tests consisted of basics such as spelling and computation, problem-solving ability or cognition and self-concept or self-esteem (“affective development”) – a division that happened to correspond to the three approaches taken by the test models.

Direct instruction, one of the basic skills approaches, showed the greatest positive impact on all three types of development (see graphs).


Follow Through


Ironically, those methods aim specifically at improving cognition or boosting self-esteem showed either negative average effects or no average effect on all three types of measure. This result anticipates later studies that have shown “the dark side” of efforts to found education upon self-esteem rather than viewing self-esteem as an outgrowth of acquiring proficiency at skills.

Direct Instruction (DI), devised by Siegfried Engelmann in the early 1960's as he taught his own children, is defined by the researcher James Baumann: « The teacher, in a face to face, reasonably formal manner, tells, shows, models, demonstrates and teaches the skill to be learned. The key word is teacher, for it is the teacher who is in command. »

Gary Adams - co-author of Research on Direct Instruction, a recent evaluation of DI- adds that « the difference is the curriculum, not so much the method. It's the sequence of concepts presented that matters. Not one other model has been field tested to the extent this one has... with very good teachers and very difficult kids. »

Skills such as reading, spelling and computation are presented step by step, and reinforcement ensures that each child has mastered one step before moving on to the next.

In 1977, the Ford Foundation hired four researchers to re-evaluate the Project Follow Through data, which were showing the superiority of Direct Instruction.

Published as “No Simple Answer” in the Harvard Educational Review, the re-evaluation by E.R. House, G.V. Glass, L.F. McLean and D.F. Walker argued that Dl's superiority was not that statistically significant if the analysis was refocused; was due more to environment than to the method used; and that while such a method might work for at-risk students, it was not relevant to the needs of “normal” students.

Bonnie Grossen, editor of the journal Effective School Practices, replies that the writers' « re-analysis not only continued to show the much greater strength of DI on all measures, but they even proved how academically valid the measures were. »

« What they did was change the question because they didn't like the answer. Their argument was that we should ignore the question of what leads to better academic learning. So they essentially gave people permission to ignore the whole point of the project. »

« It's hard to figure out why they would do such a thing, » she muses.

After the Harvard article appeared, all the test models were recommended equally for dissemination to the school districts, and by 1982, the least-effective models were receiving higher levels of funding than the most effective ones, in an apparent effort to equalize results.

Deputy Secretary of Education Marshall “Mike” Smith was involved with Project Follow Through at Harvard during the early 1970s.

The program « was a very ambitious effort to try and understand how children learn, » he recalls. « We found out how difficult it is to change the schools to make them more effective. »

Were the findings swept under the rug, as critics claim? « Oh, I think that's wrong. There wasn't just one finding that came out of Follow Through. There was a general finding that highly structured classes focused on basic skills produced better results on basics skills tests. »

« The world of the classroom is a pretty complicated place, » concludes Mr. Smith, a former dean of education at Stanford University. « This is difficult stuff. »

The president of the National Council of Teachers of Mathematics, Gail Burrill, was asked about Project Follow Through in a recent interview and responded, « I have never heard of it » (Notices of the American Mathematical Society, January 1998).

Mr. Adams shakes his head. « The most puzzling thing is how the very models like whole language and discovery learning that the data showed to be ineffective and even harmful are still being pushed. Parents should be asking, 'Where is the proof these programs work?' »

« Instead, the screaming only starts when the awful test results come in, » Mr. Adams says.

Une réalisation LSG Conseil.