It's your ConCon. What do you want to do with it? Learn - Discuss - Decide
Ouchi on School Decentralization
In school reform, the work of organizational-structure scholar Dr. William Ouchi of UCLA has often been cited here in Hawaii and in other states as a blueprint for
improving results through decentralization. All of us who are motivated to improve the public schools, and especially those who advocate decentralization, really need to do their homework by reading Dr. Ouchi’s work, as well as other scholarly research.
In the past two weeks I have studied Ouchi’s report of his seminal research about school decentralization. I would like to give my reaction to his research report.
It was not primarily a correlational study but a pilot study to acquire some initial empirical data about the organization of school systems, with only a secondary purpose of testing the hypothesis that decentralization would be correlated with achievement. Ouchi achieved his first purpose of acquiring data, but was not able to conclude that decentralization and achievement went together.
His data, though scant and unclear, suggest the opposite: in a student cohort comparison he found that Los Angles, a centralized district, showed the clearest improvement in test scores. The other analyses that he performed indicated that Seattle, one of the decentralized districts, declined dismally in both cohort scores and the achievement gap between ethnic groups. There was no other significant finding about the relation between decentralization and achievement.
But, curiously, in his discussion Ouchi asserted that Houston, a decentralized district, outperformed L.A., thus confirming his assumption that decentralized districts would outperform centralized districts. The reason this was a curious assertion is that both L.A. and Houston increased their scores by the same amount during the two-year period examined, although Houston was already performing better at the start of the period; and Houston did not implement decentralization until the end of the period! There is no logical way that decentralization could account for Houston’s initial advantage over L.A.
Ouchi’s report abounds in this kind of verbal sleight-of-hand. No wonder the Hawaii DOE liked him.
His research method was not designed adequately to discover any objective correlation between decentralization and achievement. The principal technique which his study group used was to interview school personnel about their perceptions of local control. This was to establish whether the districts could be classified as decentralized or not. But the interview technique is usually considered the weakest way to do research.
In addition, they did intensive interviews, with recordings and transcriptions, of only one (1!) school in each of the six cities which formed the main comparison groups. Ouchi’s promotional literature says they studied 223 (or 232) schools, but that is hyperbole. Originally each superintendent gave Ouchi a list of 5% of their schools which the study group would be permitted to visit for one day. That, plus the Catholic and independent schools which they visited, makes up the 223. Ouchi admits that he has to assume the superintendents picked their best schools. But then, on the basis of these initial visits, which consisted of informal interviews of the principals and tours of the schools, Ouchi and his team picked one school in each city to focus on with a two-day follow-up visit. I think we have to assume that Ouchi picked these with a view to promoting decentralization. Indeed, Ouchi expresses disappointment that the hard statistics of achievement contradicted his glowing personal descriptions of schools which he liked. The end result is that six schools were double-subjectively selected out of over 3,000 schools in those six cities.
The selection of comparison cities was also questionable. He said the reason for selecting Los Angeles, Chicago, and New York (2635 schools) was that they were the largest centralized districts in the country, and therefore the effects of centralization would be most noticeable in them. But these districts are several times larger than the decentralized cities in the study, Edmonton, Seattle, and Houston (591 schools). If Ouchi were interested in an objective comparison, he would have tried harder to match the cities on the basis of size, wealth, ethnic makeup, etc. The choice of cities may have been appropriate for his main purpose of gathering data, but not for comparisons.
In Ouchi’s discussions in his report, he does not seem completely objective. Much of the report is given to florid quotes from the interviews praising local control. Instead of the most germane facts, he selects information that can serve as selling points for decentralization. For example, despite admitting that his cohort comparisons are more meaningful, he says little about them and instead emphasizes district-wide test results, perhaps because the cohort comparisons make decentralization look worse than the district-wide comparisons do. For some reason he prioritizes retention and graduation rates, in which decentralized schools had the upper hand, as his main indicators of achievement, even though retention and graduation are relatively ambiguous compared to achievement scores, in which centralization prevailed. He tries to explain the dismal decentralization results in Seattle by saying that, perhaps, it is because Seattle has had an influx of immigrants; while this is an appropriate comment in a research study, he does not offer any similar alternative explanation for the facts that seem to support his bias for decentralization.
I would like to add one of his results which may be useful. He found that accountability is not related to results. Neither closing schools nor removing principals had anything to do with achievement, nor were they correlated with degree of decentralization. But on second thought, his study was not really thorough enough to reach these conclusions, either.
All in all, it seems to me that Professor Ouchi has a product to sell, and his publications are not entirely objective. I hope we can find better research in preparation for a con-con. Any suggestions?
improving results through decentralization. All of us who are motivated to improve the public schools, and especially those who advocate decentralization, really need to do their homework by reading Dr. Ouchi’s work, as well as other scholarly research.
In the past two weeks I have studied Ouchi’s report of his seminal research about school decentralization. I would like to give my reaction to his research report.
It was not primarily a correlational study but a pilot study to acquire some initial empirical data about the organization of school systems, with only a secondary purpose of testing the hypothesis that decentralization would be correlated with achievement. Ouchi achieved his first purpose of acquiring data, but was not able to conclude that decentralization and achievement went together.
His data, though scant and unclear, suggest the opposite: in a student cohort comparison he found that Los Angles, a centralized district, showed the clearest improvement in test scores. The other analyses that he performed indicated that Seattle, one of the decentralized districts, declined dismally in both cohort scores and the achievement gap between ethnic groups. There was no other significant finding about the relation between decentralization and achievement.
But, curiously, in his discussion Ouchi asserted that Houston, a decentralized district, outperformed L.A., thus confirming his assumption that decentralized districts would outperform centralized districts. The reason this was a curious assertion is that both L.A. and Houston increased their scores by the same amount during the two-year period examined, although Houston was already performing better at the start of the period; and Houston did not implement decentralization until the end of the period! There is no logical way that decentralization could account for Houston’s initial advantage over L.A.
Ouchi’s report abounds in this kind of verbal sleight-of-hand. No wonder the Hawaii DOE liked him.
His research method was not designed adequately to discover any objective correlation between decentralization and achievement. The principal technique which his study group used was to interview school personnel about their perceptions of local control. This was to establish whether the districts could be classified as decentralized or not. But the interview technique is usually considered the weakest way to do research.
In addition, they did intensive interviews, with recordings and transcriptions, of only one (1!) school in each of the six cities which formed the main comparison groups. Ouchi’s promotional literature says they studied 223 (or 232) schools, but that is hyperbole. Originally each superintendent gave Ouchi a list of 5% of their schools which the study group would be permitted to visit for one day. That, plus the Catholic and independent schools which they visited, makes up the 223. Ouchi admits that he has to assume the superintendents picked their best schools. But then, on the basis of these initial visits, which consisted of informal interviews of the principals and tours of the schools, Ouchi and his team picked one school in each city to focus on with a two-day follow-up visit. I think we have to assume that Ouchi picked these with a view to promoting decentralization. Indeed, Ouchi expresses disappointment that the hard statistics of achievement contradicted his glowing personal descriptions of schools which he liked. The end result is that six schools were double-subjectively selected out of over 3,000 schools in those six cities.
The selection of comparison cities was also questionable. He said the reason for selecting Los Angeles, Chicago, and New York (2635 schools) was that they were the largest centralized districts in the country, and therefore the effects of centralization would be most noticeable in them. But these districts are several times larger than the decentralized cities in the study, Edmonton, Seattle, and Houston (591 schools). If Ouchi were interested in an objective comparison, he would have tried harder to match the cities on the basis of size, wealth, ethnic makeup, etc. The choice of cities may have been appropriate for his main purpose of gathering data, but not for comparisons.
In Ouchi’s discussions in his report, he does not seem completely objective. Much of the report is given to florid quotes from the interviews praising local control. Instead of the most germane facts, he selects information that can serve as selling points for decentralization. For example, despite admitting that his cohort comparisons are more meaningful, he says little about them and instead emphasizes district-wide test results, perhaps because the cohort comparisons make decentralization look worse than the district-wide comparisons do. For some reason he prioritizes retention and graduation rates, in which decentralized schools had the upper hand, as his main indicators of achievement, even though retention and graduation are relatively ambiguous compared to achievement scores, in which centralization prevailed. He tries to explain the dismal decentralization results in Seattle by saying that, perhaps, it is because Seattle has had an influx of immigrants; while this is an appropriate comment in a research study, he does not offer any similar alternative explanation for the facts that seem to support his bias for decentralization.
I would like to add one of his results which may be useful. He found that accountability is not related to results. Neither closing schools nor removing principals had anything to do with achievement, nor were they correlated with degree of decentralization. But on second thought, his study was not really thorough enough to reach these conclusions, either.
All in all, it seems to me that Professor Ouchi has a product to sell, and his publications are not entirely objective. I hope we can find better research in preparation for a con-con. Any suggestions?
Replies to This Discussion
-
Permalink Reply by John Mussack on -
Aloha Jimmy!
With all due respect, Jimmy, I would like to ask you to be more careful about quoting Dr. Ouchi, because you may be misleading readers who are not familiar with his research. No offense intended, but you spoke inaccurately when you said his study showed unions had no impact on performance, unions became an asset in making faster progress, and structure and type of school management made a huge difference in performance.
The only thing he said about unions was that they prevented the firing of principals and teachers. In the context of his study, that was a negative factor, because it prevented accountability. By the way, Jimmy, how can unions have no impact, and cause faster progress, at the same time? Or did you mean they had no negative impact? Either way, Ouchi did not draw any such conclusion.
If you read the research report, then you know that it is inconclusive about the relation between decentralization and performance. Furthermore, Ouchi acknowledged in his report that the sample was too small for meaningful comparisons.
Dr. Ouchi may believe in decentralization, and he may argue eloquently in favor of it, but please don't claim that his research supports it.
Aloha!
-
Permalink Reply by John Mussack on
-
Aloha Jimmy!
Thank you for offering me Ouchi’s 2003 paper, "The Impact of Organization on the Performance of Nine School Systems: Lessons for California". No need to send it to me. I read it a few weeks ago. After your recent offer of sending it to me, I read it again to see if it contained the claims you implied it did.
It was not a second study, and it did not confirm your claims about unions or about decentralization. It merely rehashed the same 2002 research report which we have been discussing. It says absolutely nothing about unions.
As far as “huge difference in performance” goes, I have to admit that in this 2003 paper Ouchi says, “These school districts have produced dramatic results in student achievement, cost containment, and control of fraud and corruption”.
But you should always look at the actual data. In the 2003 paper, Ouchi goes on to describe these so-called “dramatic results” as consisting of the superior performance of one district, Houston, over one other district, L.A., before decentralization was implemented. Perhaps Houston borrowed a time machine from NASA. Ouchi does not explain how Houston’s decentralization in 2001 caused superior test scores in 1999 and 2001.
As I understand it, this 2003 paper was Ouchi’s sales pitch to the governor of California. I think we should stop talking about Ouchi, and instead look for objective research by serious scholars who are not just jumping on a profitable bandwagon.
Aloha!
About
Badge
Photos
© 2009 Created by Peter Kay on Ning. Create a Ning Network!
