The small group work and presentations on the Capital Region data base have provided us a demonstration of the interpersonal dynamics and political coalition dynamics that frame the core question of conceptual credibility about horizontal keiretsu. 1) How can a forum deliberately designed to foster cooperation and collaboration among highly competitive units persist as a networking arrangement? For a network of highly performing school districts engaged in Regents reform for secondary improvement the question might be framed as "if we can convince the public and educators of my district that public education is high quality for us what reason would I have to extend that concern for quality to neighboring districts, our area of the state or for public education throughout New York?" For a group of post masters graduate students considering the application of information about districts in the Capital Region the question might be, " if I can convince myself that I have achieved a greater familiarization with actual information used in considering the Business Review ranking, information derived from New York State of Learning and the Office of the Comptroller what reason would I have to extend those understandings in any collaborative way?" Two lessons learned from 745 efforts to date that have nothing to do with Japan or cultural applications of keiretsu.. 1) The interpersonal tensions of our horizontal keiretsu efforts have a data driven component and a political coalition component. The data driven issue is the extent to which the utilization of information facilitates a "common language" that promotes a total group development in policy analysis more than emphasizes the individual differences in expertise in using data to stratify the group. The Networking text raises such questions on pages 68-69 (and footnote 17). The 745 experience was expressed in how each small group decided how to prepare for a report (e.g. , did every class member actually try to move information around on a spread sheet, did the analyst explain the technical aspects of analyzing well enough that the non analyst would feel comfortable talking to someone about details) and the reaction of the total class to each report effort. The horizontal keiretsu persistence question rests with possible extensions of using Capital Region information to talk about educational policy and politics throughout the rest of the semester. The political coalition issue is the Etzioni counsel about civility and mutuality in any social situation of latent adversarial feelings or open competition. To what extent does the common sense of the 745 politics seminar provide a "glue" of association in relation to the tugs of being individual graduate students and persons with real lives. Civility rests with codes like "agree to disagree" and "turn the other cheek" that get stretched to the limits in highly competitive and deliberately artificial social arrangements. A postmasters seminar with the traditional format of instructor and grade structure is inherently unstable in the civil sense. Mutuality rests with the code of "guard each other's back" and maintaining one another's "respect" in reciprocating exchanges. Paper chase tensions are less to do with what strains mutuality than the simple facts of different interests and gaps of time between connections. The horizontal keiretsu persistence question rests with interpersonal understandings about the meaning of "ending" the 745 experience when grades are given in mid May. 2) Herbert Kaufman was right about the balance of inherent contradictions being able to be played on either end of the pendulum. Further, no matter which end of the pendulum is chosen and emphasized, there will remain an inherent tension of countervailing expectation toward the other alternative. For horizontal keiretsu in the EAPS 745 seminar the Kaufman argument means the rest of the class experience can either emphasize development of your personal "research in progress" efforts or can be total group follow ups and extensions on where the small group reporting left us. Of course, it will be a matter of emphasis rather than exclusive preoccupation but this is the issue about the fundamental form of persistence that we face. If the choice is emphasize a more confederated, individual meaning of class, you need time to work on your "research in progress" efforts. There are some great projects undertaken which I will be outlining in another link. We will establish a schedule to start reporting to one another after the Easter break. If the choice is to continue as a total group exploration then here are some possible "next step" extensions from the small group reports. a) Preparing for the Business Review l996 report and adding l994-l995 State of Learning information to the existing Capital Region data base. I do not think we should under estimate what we did find out about the specific area housing 79 districts and the credibility of the Business Review rating of school districts. From the multiple regression analysis of all 17 descriptive measures to the l995 Business Review ratings (as the dependent variable) we can say some policy related things. First, regardless of the comments and impressions of people who speak without data, we cannot "account" for much in the way of patterns of systematic variation. True, the correlation squared (coefficient of determinism) was .63 and Analysis of Variance did generate a statistically significant F Test score (6.04 at .00001) so we can say that Business Review ratings did have some "association" with the pool of descriptors. Yet, the beta coefficients of each individual descriptive measure only found two measures, Regents Diploma percents (+.33) and percent of students eligible for free and reduced lunch (-.37), with any hint of accounting for some direct one-on-one association. As guides for policy investment, we can say something about the wealth capability of districts and what the Business Review rates (-.37 within.63) but the ice is pretty thin for Kozol-like pronouncements. It may be more interesting to probe the differences of what free and reduced lunch means as the actions of energenic administrators as they enhance eligibility count for federal allocations. The same for Regents Diploma percent as we consider (+33. within .63) as an expression of the "overlap" between part of our dependent variable and one of seventeen descriptors. Perhaps we could push this problem by reducing the 79 districtsto the 55 that received some University Press points and reanalyzing for the overlaps of Regent driven measures. Acknowledging what systematic consideration of data does NOT find may be the first reason we should continue to plug away with the spreadsheets. Remember we are "backward mapping" from given ratings and state collected information. Maybe the class should focus upon what we would do to "measures" if we were designing the collection, analysis and interpretation sequence that could move toward "indicators" of what varies systematically with academic performance. b) looking closer at Comptroller definitions and State Education Department definitions of "direct teaching" and "administration of instructional and media" spending. Who would we ask to determine the actual placement of details of particular functions (e.g., special education) and personnel? What would happen if we added supply and equipment costs to operations or debt service to overall spending in Comptroller data? Who rationalizes the difference between SED estimates of revenues and spending and Comptroller final audit? c) looking directly at race and ethnicity in the equation of relationship to achievement and performance in the Capital region. Starting with the Bracey report from NAEP data presented in Phi Delta Kappan. What does the Capital region mean when we look closer at race, poverty and social meanings? We do have access to l990 Census information and we have the Center for Economic Growth for information about business and economics of the area. d) looking at the impact of the continued impasse and inactivity on state and federal funding arrangements on the 79 capital region districts. We already know which districts are most dependent on these avenues, could we find out if those districts are anticipating and responding to next year's budget differently than less dependent districts? e) the evolution of my thinking about state wide implementation since the Network text was written. You have the summary sheets from the Suffolk presentation. I have also written a draft of the bullseye chapter shifting the focus from 33 districts to 67 and the target chapter to emphasize the 137 k-12 districts in the top 20 % state wide. Further, the evaluation sheet you are filling out (in addition to the written critique of the text) was sent to the sixty top 10% superintendents. I am particularly interested in how this new information fits with the recent state efforts to promote an unitary Regents focus for secondary education and the possibilities for area or regional report cards. You get the idea. Horizontal keiretsu for a seminar in educational politics will never run out of "topics" and "issue foci." The real issue is investment of energy and expectation by people who are extremely busy and have already made great concessions to participate in the forum. We are at the threshold of deciding about the persistence question and no matter what form of activities chosen the keiretsu concept will be satisfied.