USF Libraries
USF Digital Collections

Educational policy analysis archives


Material Information

Educational policy analysis archives
Physical Description:
Arizona State University
University of South Florida
Arizona State University
University of South Florida.
Place of Publication:
Tempe, Ariz
Tampa, Fla
Publication Date:


Subjects / Keywords:
Education -- Research -- Periodicals   ( lcsh )
non-fiction   ( marcgt )
serial   ( sobekcm )

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
usfldc doi - E11-00022
usfldc handle - e11.22
System ID:

This item is only available as the following downloads:

Full Text


1 of 14 Education Policy Analysis Archives Volume 2 Number 7April 15, 1994ISSN 1068-2341A peer-reviewed scholarly electronic journal. Editor: Gene V Glass,Glass@ASU.EDU. College of Educ ation, Arizona State University,Tempe AZ 85287-2411 Copyright 1996, the EDUCATION POLICY ANALYSIS ARCHIVES.Permission is hereby granted to copy any a rticle provided that EDU POLICY ANALYSIS ARCHIVES is credi ted and copies are not sold.Policy Questions: A Conceptual Study Thomas F. Green Syracuse University Abstract: A policy question is a request for a fairly stable but modifiable authoritative line of action aimed at securing an optimal balance between different goods, all of which must be pursued, but cannot be jointly maximized. To such q uestions there are no purely technical solutions, a point that is revealed by the etiology of policy questions. They appear to arise from conflicts among humans over the distribution of goo ds, i.e., conflicts of interest. However, the deeper roots of such questions lie not in a conflic t of human interests, but in the incompatibility of the actual goods that human beings seek. Policy questions ask how to allocate such goods. But this allocation is the business of politics. No pol icy without politics nor politics without polity. Policy questions must be distinguished from constit utional questions, moral questions, and questions aimed at forming the set of alternatives from which policies might be selected. Finally, different facets of the policy process are distingu ished -policy analysis, formation, decision and the political analysis of policy. These never occur discretely or in fixed sequence. But they are different activities, having different theories and requiring different skills, and capacities. Together, they model what is required for that most comprehensive of all offices within our political system the office of citizen.I. Policy Questions No single definition of "policy" is likely to mirr or the full range of ordinary usage. Usage is too diffuse for that. Such a definition would ha ve to capture the likenesses and differences between managerial decisions, guides to practice, a nd rules of thumb as well as rules of conduct embodied in legislation. It would have to capture t he difference between basic and procedural


2 of 14policy, between prescriptive and permissive policie s and policies simply expressing the bare application of standard requirements in administrat ion like, "No one gets unemployment assistance for more than twenty-one weeks." It woul d have to subsume matters that fall under "Standard Operating Procedures" (File expense accou nts within ten days with receipts), and matters of personal practice (I don't answer the ph one at home because it is never for me. Avoid arguments in the office). Although each item in the full range of these examples can be called "policy," the term has, in each case, a slightly di fferent meaning. But despite the difficulty in reflecting the range of ordinary usage, progress, o f a sort, is possible by searching for the features of policy questions, especially if we focus upon pu blic policy in contrast to what might be thought of as personal policy. A policy question is a request for a fairly stable, but modifiable, line of action aimed at securing an optimal adjustment of the conflict betw een different goods, all of which must be pursued, but which, taken together, cannot all be m aximized. We do not have a wellformed policy question or a fully formulated statement of a policy problem until we can state the set of values or goods from which the question arises, and do so, moreover, so as to reveal their mutual inconsistency. The problems of education finance provide about as clear a model of policy questions generally as it is possible to shape. The policy is sues are always "nested" within a set of mutually incompatible values or goods: equal educational opportunity for children an equitable distribution of the tax burden local control of education, and responsible management of the State budget. Maximizing any one of these goods, that is, getting as much of it as we can, will inhibit the advancement of the others. The policy problem is ge nerated by the fact that we accept all four of these aims and yet they cannot all be maximized. We cannot have all the local control possible, for example, because doing so, at least in a system of local and State partnership, would mean allowing the local district to write the bill for e ducation and oblige the State to pay it. The State would have lost control over its expenditures. Maxi mizing local control would probably also mean getting less equity for children and taxpayers and less equity among districts than would be good. On the other hand, if we maximize equity for children, then we are likely to get more inequity in the spread of the tax burden and less l ocal control. The problems of educational finance, in short, do not arise merely from the nee d to establish a more equitable system for taxpayers and children. They arise rather from the need to do so within a system of public goods that seeks greater equity, preserves local control and encourages responsible public management. Virtually all issues of public (or even personal) policy have this feature. They are "nested" in a set of social goods all of which must be consi dered, but which, taken together, are at some point, mutually incompatible. Consider the issues s urrounding the imposition of exit standards at the secondary school. Here we seek the mutual benef its of: universal levels of attainment racially or ethnical ly balanced, high academic achievement, and culturally pluralistic communities The first two present the familiar conflict between "quantity" and "quality," that is, the old puzzle as to whether education can be universal and at the same time, excellent. Problems involved in attempting to balance the entire set ar e illustrated by experience in Florida and Virginia. Not many years ago, in an effort to raise the standards of high school achievement, these states introduced uniform examinations as an exit requirement from high school. Where


3 of 14should we place the cut-off score? The question is certain to be asked once a decision is made to raise standards in that way. In Virginia, the cut-o ff was set by the legislature without much attention to relevant data As a result, the attai nment rate declined. That some students would fail the exit examination was acceptable in prospec t to those who thought such results would signal a rise in academic standards. That would be progress toward one goal. To many, however, including, presumably, many who had approved of the new policy, the result was less pleasing because it produced what was, for them, an unaccept able decline in the level of attainment. That meant retreat from reaching another goal. Thus, the cut-off was adjusted in order to find a balance between the goods of high standards and hig h attainment rates. The new policy turned out to be very different from the old in appearance but not in results. Such advancements and retreats, stories of steps t aken and then taken back, are not unusual in movements of school reform or in efforts at change in public policy generally. Such adjustments are likely to result more from the pres ence of incompatible values or goods than from administrative bungling or blindness, or from inefficiency, or from political chicanery or legislative incompetence. That is to say, neither t he most efficient action nor the most technically proficient analysis is likely to resolve the centra l discord among the social aims in which policy questions are rooted. In general, there is no techn ical solution to a policy question, no technical means of evading the fact that if exit standards ar e raised, for example, then at least in the short run, the cost will be reduced attainment levels. Tw o goods will contend. Policy Questions, Practical Reason and the Technicist Delusion This conclusion, that there is no technical soluti on to policy questions, deserves further exploration. Note first, that policy questions are always practical questions, never theoretical. What counts as an answer to a policy question is al ways a statement stating what we should do, never a statement setting forth simply what we know In any public forum where policy is shaped, only practical questions are admissible, ne ver theoretical questions. And this is fortunate because it means that in the domain of policy we ar e able to arrive at agreement on what to do without having to agree on the reasons for doing it We are able to agree on a line of action and stick to it, even when we do not agree on what is g ood and even when we have different goals. (Note 2) The answer to a policy question is always a line of action, a specification of what to do. The answer to a theoretical question, on the other hand, is always a truth claim. Policy deliberation is aimed at action, not at the acquisi tion of knowledge; whereas theoretical questions are aimed at the acquisition of knowledge, not at a ction. I do not mean by this that we can or ever should make public decisions without knowledge. Soc ial action should be informed. Nor do I mean that we can ever gain greater knowledge withou t some action. Research, the search for answers to theoretical questions after all, is itse lf a kind of action. I mean rather that in any poli cy debate the thing to be decided is what we should do not what we must believe. Wise policy is never made with enough knowledge to determine a dec ision, and policy questions are never asked out of a primary interest in adding to our kn owledge. These days -when a return to paradise seems often to be viewed as a mere technical difficulty shortly to be overcome -it is easy to imagine someone saying, "If we just had methodology sophisticated enough and relevant data sufficiently refined, then we could answer whatever policy questions might come along." Such a person has fallen victim to an illusion, the illusion that policy questions are theoretical. Whe never we suppose that a policy question can be resolved by some addition to our knowledge, we supp ose that what presents itself as a problem of policy is in fact a problem of management or admini stration. My point is not that we should abandon all attempts to improve our methods of evaluation


4 of 14or policy analysis. It is rather that since indecis ion in matters of policy does not arise entirely from the lack of such methods, therefore, it is unl ikely to be laid to rest by their development. In matters of policy, we are confronted with indecisio n not because our knowledge or technical facility is faulty but simply because we are confro nted with a kind of question that, in principle, cannot be answered simply by any increment or impro vement of knowledge. Answers to policy questions may be improved by better information and better analyses only in the sense that such gains will make our answers rationally more persuas ive. But policy questions can, will, and usually are, answered even without such information Furthermore, it is not always obvious that the answers given in the absence of such analyses a re worse than or even often very different from answers that would be given in their presence. Answers might be better grounded rationally without being better or even different in any other sense. (Note 3) We can, no doubt, do something rationally more persuasive than consultin g chicken entrails, but we are unlikely to get anything that produces results quite as decisive. A nd this is so because of the properties of policy questions, not because of deficiencies in policy ev aluation.II. The Etiology of Policy Questions Scarcity : This last observation opens a second approach tow ard understanding why there are no technical solutions to policy problems. We c an wonder where policy questions come from. Are there peculiar features of human life that ines capably produce such questions, aspects of the human condition, as it were, from which of necessit y they arise? In paradise, it might be observed, there are no policies--except, perhaps, a dmissions policies. Policy questions do not occur there. Why not? Ideas of paradise, of course; have appeared in great variety, but I am inclined to the view that, on the whole, human bein gs conceive of it (and they will inevitably conceive of it in some way) as a perfected state of affairs, one in which desires presently denied in this imperfect world, somewhere, sometime, will be satisfied. (Note 4) What else could paradise be except a condition in which all benefic ent human yearnings are satisfied and the deepest ones most deeply? This suggests, of course, that the problem of optimality and therefore the need for policy would be banished from any worl d in which human desires are perfectly matched by satisfactions near at hand. Ask and it s hall be given. Yearn and you shall be gratified. Two general strategies exist for arriving at such a n elysian state. The first lies on the side of increasing the satisfactions available to human beings (increased productivity); and the second on the side of doing something about their desires (improved moral education and discipline). The first view holds that desires and satisfactions can be balanced by abundance Wherever there is enough of everything, including enough justice, enough virtue, and enough bread and wine, there is no allocational problem, no problem of sca rcity, and hence, no questions of policy. The other strategy is the converse. There is no sc arcity of what nobody wants. So desire and satisfaction can be made to meet not by the sat isfaction of wants, but by their proper and harmonious composition. This can be viewed as simpl y another side of the same path toward abundance. Thus, for Ghandi, diamonds and mink were plentiful, not because they were any the less scarce, but because they were not wanted. And not being wanted, they were abundant. If heaven is that condition in which the wants and des ires of an imperfect world are satisfied in a world like this one except perfected, then there ma y be abundance in heaven not because goods are maximized, but because wants are composed, not because the world is different, but because we are. In neither case do problems of policy arise, and th ey do not arise simply because such a bounty, secured by either strategy, makes choice re dundant. In paradise, there is neither discord nor delay between desire and satisfaction and there fore no place for policy. There is no need for it. I offer this fact in proof that one source of p olicy questions is the mismatch between human desires and satisfactions.


5 of 14 Conflict of Goods : But there is more. Beyond this discord of desire and satisfaction, there may be strife among persons, and moreover, as alrea dy suggested in the cases of school finance and exit standards, among the goods themselves that human beings seek. Why can we not imagine paradise to contain interpersonal conflicts of wants? The answer is that doing so would involve either the judgment that wants are improper ly controlled or that goods are too meagerly supplied to satisfy them. Along with such a conditi on would come a problem Paradise would no longer be like this world only perfected. It would be like this world still unperfected, no longer paradise It would contain the problem of composin g the most satisfactory combination of what goods do exist, who should get them, and in what de gree. In short, such a condition would introduce into paradise precisely those circumstanc es that create the need for policy and that dictate the features of a well-formed policy questi on, namely, What are the goods in conflict? What is their best possible adjustment? How can we reach it? What are the trade-offs? The main force of this incursion into the conceptu al territory of heaven is to show that problems of policy are an immediate and direct refl ection of some immensely fundamental characteristics of the world and of human existence within it. Policy questions arise because the goods--not simply the interests--that human beings seek to secure in the world are interdependent and often jointly discordant, not all of the time i n every respect, but all of the time in some respects. Only in paradise can we imagine all human goods simultaneously in sufficient supply so that there is no conflict in their allocation. It i s important to note that since knowledge is a certa in kind of good, then the idea of paradise includes th e assumption that knowledge is ample. Paradise does not arise, however, because there is an overflow of knowledge. Policy questions are not banished from paradise because our capacity to know is perfected. They are banished rather because all goods are plentiful, including v irtue and justice, or because there is a suitable composition of desires. Policy, Politics, and Utopia : This formulation, however, is still insufficient to expose the presuppositions of policy questions. It deals with the discord of satisfaction and desire and with the scarcity of goods, but not with their interdepe ndence. It is sufficient to show that when scarcity is absent--as in heaven--then no policy pr oblems arise. But the example presupposes coherence in the view that all human goods can exis t in abundance simultaneously. Our imagination poses the simplifying assumption that a plenitude of some goods sought by human beings can exist without significantly reducing the supply of others. And this assumption is false. Human goods do conflict. They conflict, moreover, in such a way that they cannot all be present sufficiently to satisfy human desires. (Note 5) The point is central. Human discontent is endemic. (Note 6) The price of securing the abundance of some human goods is always failure to secure as much as we would like of some other. Thus For example, if we succeed in providing as much equality as is wanted, we are unlikely to have as much liberty as is wanted. Or if persons develop as much tolerance for ambiguity as is wante d, we are unlikely to have as much courage as is wanted, and the price will sooner or later be come apparent. The view then is that human goods conflict (read values" if you wish) not simply because they are in short supply, nor simply because interp ersonal preferences conflict, but simply because they are not structurally consistent. They cannot all be maximized, even in paradise. It's a familiar idea that human wants or human interests c onflict. But the view here is that human goods conflict. The conflict of interests produces political problems, but inconsistency among human goods produces policy problems. Human goods c onflict even when human interests do not. Even when, rarely, all are agreed on a single predominating interest--victory in an all-out war, for example--policy problems remain. In short, there are no free goods, none that can be provided in plenitude without making certain others in short supply. There is no ultimat e solution to any problem of policy, except in paradise, and there problems of policy are not so m uch solved as they are simply non-existent. That the ultimate solution of policy problems is to be found only in paradise, may be exactly


6 of 14what gives rise to the consistent and apparently in eradicable human impulse to think of social solutions to policy problems in utopian terms, rath er than in the proximate terms of both politics and policy. Politics and Policy in Utopia : But utopian thinking is faulty not only because i t attends too little to what is feasible, but because it atte nds not at all to politics. It is an odd fact, and one of large significance, I think, that no well develo ped literary exposition of Utopia ever includes an account of politics. The central assumption is a lways that in Utopia the needed balance between conflicting human goods is resolved. The po rtrayal of Utopia consists precisely in offering an account of that resolution. Reconsidera tion in Utopia, therefore, is neither needed nor desired. To introduce politics into Utopia would be a threat to, not a part of, the good life. There, problems of policy remain only in the guise of moni toring the society and managing its affairs. Politics is always replaced by administration in Ut opia. Policy questions do not arise in paradise and serio us ones do not arise in Utopia. But the reasons are different. The reason that serious ones do not arise in Utopia is not that goods are abundant or desires composed, but that the inherent conflict between them is taken as resolved. All that remains is management. Not even the presum ed Utopia of putting evaluators in charge would alter that result. There is no technical solu tion, but neither is there any permanent solution to a policy problem.III Other Presuppositions Policy questions should be distinguished also from moral and constitutional questions as well as research questions. Failure to do so is a s ource of much confusion. Aspects of these contrasts can be discovered in the different ways i n which these different questions are related to duration and to the human desire for assurance. Constraints of time : Like a reporter filing a story for the evening ed ition, whoever will answer a policy question, in the real world, must d o so within strict constraints of time. The reporter meets a deadline, but always in the knowle dge that there will be a next edition. Today's story is amended by the next as events unfold and f urther facts emerge. Note two things about this observation: first, that policy questions gene rally are answered in anticipation that the answer will be revised. and secondly, that they are questi ons that have to be answered on time, even though the information needed for the answer is not on time. Both points arise from temporal constraints, but they have different implications. The first implies simply that policies are imperma nent. We expect them to change. Often, they are not even very durable. They are not suppos ed to be. In that respect, they are like the reporter's story and unlike constitutional and mora l questions. If we imagine some institution or government in which policies change with speed and frequency, say, every three months, or worse, every week, then we imagine circumstances in which there is no policy at all. Policies are guides to action subject to amendment, but not that much. They should be durable, but not permanent. Consider, for example, the American fore ign policy of "containment in "confronting world communism" as proposed by George Kennan in th e forties. Here was basic policy that remained in force for nearly fifty years. Policy as durable as that, like policy that changes every week, tends not to be viewed as policy at all, but simply as "a given" in deliberation, a condition of the world, like the seasons, something that life must take into account, but which is not subject to change. Permanence is as alien to the idea of po licy as transience. Policies are less durable than constitutional rules which, in turn are less durable than moral standards or principles. We do not expect persons t o change their moral principles (note the offense involved in describing such "principles" as "policies) as rapidly as they may change policy. Anyone for whom honesty is merely a good po licy is not to be trusted. We do, however, expect policy to change with experience and fairly often. The dimension of durability is among


7 of 14the defining features of context for policy questio ns. Just as it is better for the reporter to file a sto ry on time without all the facts than to get all the facts and file the story too late, so also in t he case of policy it is better to make a decision o n time, but without all the facts, than to get all th e facts and make the decision too late. In the case of policy, decisions have to be made within large l imits of uncertainty. Some reduction in the degree of uncertainty will be helpful, but the degr ee of reduction normally required for academic research is both improbable for policy decision and would often be undesirable even if it were not improbable. In other words, crude data arriving on time are always to be preferred over refined data arriving too late. So it is acceptable even fortunate, that research methods required for policy decision are crude, even though the stan dard methods of research are highly refined. We are cautioned in social science research to refr ain from conclusion until the data are in. But the data relevant to a policy decision are neve r in. The argument is never complete. To answer a question of research we are cautioned to g et as much data as is needed to arrive at a reasoned conclusion and to formulate no conclusion until that information is at hand and has been examined. Caution is among the virtues bred in to the researcher by his or her methodology. To answer a policy question, however, the constrain ts are different. In that case, we need as much information as we can get, but "as much as we can g et" usually turns out to be less than we could get if we had more time, and, at the same time, mor e than can be used and more than will make a difference to the decision. A major aim of academic research, in other words, is to reduce the levels of uncertainty enough to justify knowledge c laims. Policy questions, on the other hand, are always answered in the midst of uncertainty, and th ere is always a point beyond which more information--however more excellent--will do little to reduce uncertainty and do nothing to alter the direction of the decision. All this is simply a nother aspect of the claim that policy questions are practical rather than theoretical. They are que stions of the sort that need to be answered, and will be answered, even when we do not know which am ong alternative answers is the best. If these observations are credible, then it is possibl e to understand the claim sometimes advanced that academic research and academic researchers are useful for policy decision inversely to their academic excellence. Policy decisions and ethical decisions : Policies and moral rules or principles differ in their relative durability, but they are distinguished as well as linked in other important ways. Policies are always drawn from within a set of alternative a ctions all of which are either morally indifferent or capable of evoking moral approval. T he set may be established by what are sometimes called "peremptory rules." (See Braybrook e and Lindblom, A Strategy of Decision .) Such rules tend to establish the moral limits withi n which policy can be selected, but, at the same time, they tend to guarantee that policies are chos en from among alternatives, all of which are morally permissible or have worth. In that sense, d efining the set of policy choices is the expression of moral conviction and estimations of v alue, but selecting from within that defined set is not. Two quite different questions are gener ated. One asks what alternatives can be included among the set of choices. The other asks w hich among that set shall we adopt. The first of these questions already narrows the scope of cho ice, and that is partly the reason why, with even the best evaluation data, we are unlikely to a rrive at policy choices substantially different from those we would arrive at without such data. It follows that in policy debate, moral argument i s unlikely to pick out which policies are to be adopted. It is much more likely to aid in det ermining what alternatives can be included among those available for choice. If policy choice were actually determined by moral argument, then policy debate would be transformed more easily than in fact it is, into a debate among competing moral principles. At risk would be the fu ndamental principle that we can agree on what to do without having to agree on our moral pri nciples. (Note 7) It would turn out that what is to be decided in policy debate is not our policy but our moral principles. The matter to be settled would change from a "What should we do?" to a "What should we believe?"


8 of 14 Post-hoc determinations of policy: All arguments ai med at identifying a policy by its consequences are necessarily weak. They are all sub ject to the fallacy of post hoc ergo propter hoc, "after this, therefore, because of this," clea rly a non sequitur. In other words, claims of the form "It must be A's policy to x, because x is the result of his actions" are weak claims. They assume that because certain results follow upon A's action, therefore, it must have been policy to produce those results. If we find that high rates o f failure occur in a particular school system, the inference would be that therefore the production of failure must be the policy of that school system. There are several difficulties in such an inferenc e. One is that it obscures the fact that the identification of policy always implies the identif ication of intent. Failure may occur frequently in a given school system, but it cannot be conclude d from this fact alone, that producing failure is anybody's intent. It cannot follow from this fact a lone that producing failure is a basic policy in such a setting. The more general and decisive diffi culty in this sort of reasoning, however, is that if it is allowed, it will have the consequence of r endering it conceptually impossible ever to speak of failed policies. Every policy has its social res ults however modest and however desirable or undesirable. But we cannot reason from the presence of certain states of affairs that the policy was to produce that state of affairs. The most we c an say, to return to the example, is that frequent failure is the persistent consequence of m easures that are matters of policy. Just as the logic of such post-hoc reasoning is obj ectionable, there are also circumstances in which it would be foolish to press the objection If extreme and undesirable consequences are publicly evident and persist over a long period of time, then the appeal that such effects are unintended, that is, are not matters of policy, bec omes less and less credible to reasonable men and women. Consider the case of a local chemical pl ant engaged in the production of salts which process, over the years, has seriously polluted a n earby lake. To claim "We don't pollute the lake, we simply produce salts" is to stretch the credulit y of the average citizen beyond the breaking point. But how would we describe this situation? I propose that although we would be strongly inclined to the view that polluting the lake is off icial policy of the company, still we would do better to locate the problem in another way rather than adopting the usual post hoc argument. The more complete account would be that in determining policy, the mix of goods included in the "nest" of the policy debate is impoverished. It omi ts matters that ought to be included. In short, such post-hoc arguments, judged simply on the logic of the matter, always fail. From time to time, however, they appear to succeed, and when that happens, it is likely to be not because the argument is sound, but because the init ial mix of goods competing in the policy debate is too limited. We may reject all such posthoc arguments and still retain a decent respect for the opinions of mankind and the normal view tha t "policy" implies intent, and that policies sometimes fail.IV. The Policy Process Between policy analysis, policy formation, policy decision or promulgation, and the political analysis of policy there lie clear differ ences, and the practices of evaluation and policy research will relate differently to each. The tende ncy exists to regard these four activities--analysis, formation, decision, and poli tical analysis--as steps in the policy process. But such a view is misleading, because these activities are never fully discrete in practice and they do not occur in any persistent sequence. Nevertheless, there is a distinction of practice corresponding to each activity, and each practice, moreover, has its distinct kind of theory. Policy Analysis : Policy analysis can be defined as the rational or technical assessment of the net marginal trade-offs between different polic y choices. The question becomes, Which among the implicit set of goods will be advanced, w hich will not, and with what net benefits? This is the same kind of question that we confront, say, in the design of a hand drill. What should


9 of 14be the design? The question is nested' in a set of goods. We want low cost, high safety, ease of handling, and durability. We can ask and rather pre cisely determine what marginal gains in one of these values will produce what corresponding cos ts in the others. If we "go for greatest durability, then we are likely to get higher cost a nd less ease of handling. If we "go for" the lowest cost possible, then we are likely to sacrifi ce something in the way of durability and safety. The design problem is to discover a balance between these competing values. Enter the problem of incommensurabilities! How do we determine which among the competing values is to be given greatest weight? Wh ich has the greatest worth: low cost, safety, durability, or ease of handling? Not even the most refined analysis of the costs and benefits will solve that problem. Such an analysis gives us the p ossibilities or a set of choices, but it does not pick out any preferred answer from within the set. Yet we need some procedure for doing just that. In short, we need a market decision. and gett ing a market decision is, no doubt, going to require a market analysis. Is our market made up of professionals? Or does it consist essentially of amateurs and household craftsmen? If the former, then the proble m is likely to be resolved on the side of durability and safety with a slightly higher price. If the latter, then, by all means, the decision wi ll probably be to minimize cost and sacrifice durabili ty and, to some degree, safety. But then again, the market decision might be to go for the whole ra nge of the market. Produce a variety of designs representing the full range of choices reve aled by the analytic exercise. Something for everybody! These activities are roughly analogous to the disti nctions I want to make in the case of policy. Merely setting forth the marginal costs and benefits of a range of choices is one thing--policy analysis. Selecting one balanced choi ce or a range from within the possibilities is another thing--policy formation. The decision as to which choice or choices to make is still a third--policy decision. And performing the market a nalysis needed for that is yet a fourth-political analysis. Consider an example. Financial educational assista nce can be distributed either to students directly or to institutions. Suppose we entertain t he prospect of distributing it to students and that we resolve to do so on the basis of need. In that c ase we require access to financial information, and not simply on groups, trends, or categories of persons, but on each actual individual. If we propose this kind of policy as more fair or just th an others, then, in the name of justice, individuals will have to reveal personal informatio n that before may have been regarded as privileged Two values conflict. We extend justice but diminish, in some measure, privacy. To secure a definable gain in one, we pay a definable cost in the other. Policy analysis asks, What is the net marginal gain? A truly refined policy analy sis, which rarely exists, would tell us how much we are likely to gain in the advancement of ju stice for some corresponding cost in privacy. But no such analysis, no matter how refined, will t ell us whether the gain is worth it To resolve that question, we need something corresponding to a market analysis and a market decision. We need a political analysis and a political decision. "Policy" implies "polity" and "politics" just as "good industrial design" implies a structure for ma rketing analysis and marketing decision. Consider another example. A Congressman asks wheth er passthrough requirements for allocating Chapter I funds should rest on tests of educational rather than economic need. The answer comes couched not in terms of "whether we sh ould" but in terms of "what happens if we do?" That's policy analysis. In either case, the fu nds would go to roughly the same school districts--but not quite. What's the margin of "not quite"? Is "not quite" "very much"? Is it "enough to matter" And even if it is "not much," wo uld the change create incentives for local districts to pay more attention to educational need in answering allocational questions? And if so, then would the incentives be enough to make a diffe rence? And if so, then (here we are again) how much of a difference? That's policy analysis. But policy analysis does not, and need not stop the re. It can ask not simply what the net


10 of 14consequences would be of doing X, but what those ne t consequences would be compared to doing Y, where Y is either what we are doing alread y or some alternative. The question for policy analysis is not whether doing X is a net imp rovement over doing Y, but simply, what are the net effects? Whether it is better to have a dri ll of low cost instead of high durability will not be determined simply from an analysis of the tradeoffs. It requires a marketing decision. Similarly, whether given the different consequences it is better to do X than Y in public policy will not be determined by a policy analysis. It wil l be determined by a political decision resulting from a political process involving a political anal ysis. In short, policy analysis is that rational, technic al, analytic performance in which the central question is not whether X is a good thing t o do, but simply what are the marginal effects of doing X, and what are the marginal effects as co ntrasted with doing something else instead? Hence policy analysis is simply an activity whose t heory is the theory of marginal utilities. It is, by all accounts, an activity that consists in the e xercise of theoretical, rather than practical, rationality. It assumes that the policy question is "nested" in a conflict of values present as objective states of affairs in the society. It is a n activity in which evaluators and other researcher s may take a leading role provided that they do not s uppose they are actually evaluating policy, as opposed to merely recording--either in prospect or in retrospect--the consequences of doing X or Y. Policy Formation : Policy formation is an activity of a contrasting genre. It is that activity by which we seek to gain agreement on what form a s pecific policy can or will take, as opposed to what form it ought to take. Not even by the most refined policy analysis will we have actually formed a policy statement. Indeed, policy analysts are not typically in a position to actually formulate policy. For the latter, we need to engage in conversation, persuasion, argument, and in (seemingly) endless meetings with those who will ac tually pen the regulation, mark up the bill, establish the procedures, write the guidelines, etc The theory of policy formation can then be discerned as one aspect of the theory of government al management and rhetoric. At the Federal level, it usually turns out to be the theory of int er-agency politics. "Don't fight over turf; just ta ke up space" is a rule for the conduct of policy forma tion. I include here the theory of rhetoric because clearly it makes a difference what things a re called. The same policy that under one name may never see the light of day will, under ano ther name, pass without objection. Calling it "school aid" may defeat it; but calling the same th ing "national defense" may insure its acceptance. "If it matters what you call it, then c all it something that matters" is another guiding rule in the theory of policy formation. Call it Gov ernment spending and you may be in for political difficulties. Call it "public investment" and you may have more success. Policy Decision : Policy decision can be described as the authorita tive action of some office, administrative or legislative, by which a l ine of action, for the moment at least, is established. Policy decision is not so much an acti vity or process as it is a momentary end point in the continuing conduct of government. It is that end point that is sometimes supposed by the naive to entirely capture the policy process, as th ough making policy could be reduced simply to an act of will or the result of divination. The the ory of policy decision is simply the theory of the polity itself, the political and legal theory by wh ich authority is distributed, obligations for decision are assigned throughout the structure of p olitical institutions, and agents of authority are enjoined to act. Political Analysis : Unlike policy analysis, political analysis is con cerned not with determining the net benefits of a given course of a ction, but with measuring their political weight. The aim is not so much to determine the net social benefits of a particular policy, but to determine its constituency. If policy analysis is c oncerned with establishing what course of action has greatest worth in the utilities of public goods then political analysis is concerned with estimating who will vote for it. It is concerned wi th determining whether the best thing to do is the same as the best thing that can be done. Often it is not. The theory of political analysis s the


11 of 14theory of political behavior. We may gather these thoughts in a brief culminating summary. The theory of policy analysis is the theory of marginal utilities. It mi ght rank policy choices according to the estimated net utilities of each. The theory of policy formati on is the theory of inter-agency politics. It is th e governmental process by which a course of action co mes to be selected and actually framed. The theory of policy decision is nothing less than the theory of the polity itself, the theory underlying the placement of authority. And finally, the theory of political analysis is the theory of political behavior. When we view all of these activities toge ther, not as discrete steps in the policy process, but as distinct facets of a social process --now one feature predominating and now another--then we can discern more clearly where the practices of evaluation and policy research fit, what may be their relevance to the creation pr omulgation and implementation of public policy, and what virtues are required for the actor s in this drama. Evaluators, evaluation and policy research can cont ribute to each of these activities, but not to each in the same way or in the same degree. The rational standards of policy analysis are the standards of theoretical reason, but the rational s tandards of policy decision and political analysis are the standards of political judgment. These are practical activities. This difference may help to explain why it is that when the question, "What sho ld we do?" is given a policy analysis, we mayget one answer, and when given a poitical analys is or when rendered in a policy decision, we may get an entirely different answer. In short, the exercise of political judgment is a practical activity, also an evaluational activity. But the re sult of that activity may differ from or even contradict the results of policy analysis. What we should do--even the best thing to do--may turn out to be one thing by policy analysis and a very d ifferent thing when it comes to political decision. The professional evaluator and policy researcher ca n contribute in the context of government, but he or she will contribute to all of these activities only to the extent that they become also politicians and political advisors. Con sider, for example, the case of policy formation. The evaluator, as professional, can cont ribute, but that contribution will be most substantial to whatever extent he or she becomes a student of bureaucracy and a trusted counselor to authoritative leadership. So the dilemma is this. Each of these activities in volves evaluation and research in some broad sense of those terms. Each involves evaluatio n and research in the sense, say, that buying a camera does. But only in the case of policy analysi s is the evaluator's role, as professional, undiluted by the need to take on other roles. The e valuator, as evaluator, is likely to make a contribution only to the conduct of policy analysis But in government as elsewhere, the possession of knowledge can bring with it a certain kind of power. To the extent that the evaluator and policy researcher goes beyond his pro fessional practice and with superior knowledge also earns the confidence of political le aders, exercises political judgment, and acquires the additional skills of a practiced polit ical observer of the present bureaucracy and an undertain future, then she will contribute to every facet of the policy process. But in doing so, he or she will also become less an evaluator or resear cher in any limited professional sense and more a political leader or public servant in a quit e old-fashioned and conventional sense. He mail characteristic will not be the possession of techni cal skill. It will be the possession of a particularly important kind of civic virtue not dif ferent in kind from that virtue we have always called upon citizens to perform.NotesTitle Note. This paper is an adaptation of an earli er paper, "Policy and Evaluation: A Conceptual Study," #52 Paper and Report Series, Northwest Regi onal Educational Laboratory, Research on Evaluation Program, 1981, pp. 162-187. It will even tually find its way into one or the other of


12 of 14two works in progress: VOICES, on the educational f ormation of conscience, or WALLS, on education in communities of text and liturgy both s acred and secular. The intent of that original paper remains, namely to display the conceptual fea tures of policy questions generally and to display the tendencies to view such questions as es sentially technical. The present version, however, is much elaborated and amended.Note 2. Can we suppose that human beings will agree on what to do provided they agree on what is good? It does not seem likely. Where does this l eave us? The argument is, on the one hand, that we can agree on a course of action even when w e disagree on what is good; and, on the other hand, that when we do agree on what is good, when w e share the same goals, it does not follow that we will agree on a course of action. If these are the facts, then it does not seem obvi ous in what ways theoretical discussions about the nature of the good and the good society a re relevant to the determination of public policy. Think, for a moment about political coaliti ons. The formation of such coalitions is one demonstration of people agreeing on a course of act ion without having to agree on what is good. Coalitions come and coalitions go, and their coming and going manifests one feature of the fact that an agreement on what to do is always a practic al question and is never determined by answers to theoretical questions of "value" and "pu blic good." Note 3. Notice that "rationally more persuasive" ma y not imply either "politically more weighty" or even "politically more wise," a point that, as w e shall see later, figures prominently in understanding the different stages of policy format ion. Note 4. A compilation of eschatological visions wou ld reveal a great deal about this human tendency to form images of perfection, how the incl ination persists, even though the content changes. The parameters of such a collection, the b order types, might be found in two paradigmatic cases, the one a 12th Century version coming from southern France, and the second an understanding found in Marx's Lectures on Politi cal Economy. By the first of these, heaven is understood as a kind of life in which one works and works, but does not weary, a life in which the rest and fellowship at day's end are sweet like candy, but one is never sated. From these conditions, it is easy to discern what were perceiv ed to be the principal failures of the world, the things most fervently longed for, by those standing within this tradition. On the other hand, when Marx observes that bourgeois society is the last st age of Prehistory we know that we have entered a world entirely different from the France of the m edieval peasant, a world in which the eschaton comes not at the end of history, but at the beginni ng, a world in which the conditions of human beings as the objects of social forces has yielded to a day in which they are the actors. Even in Marx's secular world of the Enlightenment, images o f paradise prosper. Note 5. What will become more apparent later should be stated briefly here. The view of human ''"goods (read 'values' if you must) is that such g oods as security, health, privacy, liberty of many sorts, work, and the like are not so much attitudes or dispositions of persons as they are objective social structures, the arrangements of social life. H public policy aims at an enlargement of liberty, then it must nurture those economic, socia l, and educational, arrangements by which liberty of movement, of conscience, of speech, of i nquiry and the like are nurtured. It is these arrangements that conflict in such a way that they cannot be jointly accessible. Note 6. The fact that human discontent is endemic t ogether with the hope and expectation (not merely wish) that it needn't be, may provide a clue not only to the apparently ineradicable human tendency toward utopian thinking, but also a clue t o the peculiar American impulse to engage in complaint, even perhaps to the point of elevating l amentation to the level of an academic


13 of 14 discipline, or, if not that, then a literary genre. There is probably no society in the world whose people are so given to lament as the American. For Americans, perhaps more than any other peoples of the world, there is always something wro ng, seriously wrong, that needn't be. The solution always seems to be more fidelity, more ben evolence, more liberty, more education, more pride, or stronger leadership. Lamentation is a nat ional recreation. Americans seem to find pleasure in it. This peculiar mark of national char acter makes no sense until we see it possibly as just the underside of American optimism that some h ave supposed is nurtured by life in a land of a material abundance sufficient to provide for endl ess moral possibilities. Sacvan Bercovitch has sought the origins of this disposition in what he h as called The American Jeremiad (University of Wisconsin Press, 1978)Note 7. Enter an important caveat! The policy choic e is likely to occur in consequence of a kind of utilitarian calculus refined through several sta ges of the policy process. But it seems odd to regard this utilitarian calculation as in any sense a process of "moral reflection," so thoroughly infused is it with considerations of aggregate effi ciency and net benefits. The so-called "right-to-life" position is not grounded in such co nsiderations. It is grounded in convictions concerning what is morally right, wrong, or evil, a nd hence is concerned not so much with what policies should be adopted as with what courses of action cannot be included among viable choices. Here is a clear case where the moral delib eration occurs outside the realm of policy debate. Which is where it belongs.ReferencesBercovitch, Sacvan. (1978). The American Jeremiad Madison, WI: University of Wisconsin Press.Braybrooke, David & Lindblom, Charles E. (1963). A Strategy of Decision. Policy Evaluation as a Social Process NY, NY: Free Press of Glencoe.About the AuthorCopyright 1994 by the Education Policy Analysis ArchivesEPAA can be accessed either by visiting one of its seve ral archived forms or by subscribing to the LISTSERV known as EPAA at (To sub scribe, send an email letter to whose sole contents are SUB EPAA y our-name.) As articles are published by the Archives they are sent immediately to the EPAA subscribers and simultaneously archived in three forms. Articles are archived on EPAA as individual files under the name of the author a nd the Volume and article number. For example, the article by Stephen Kemmis in Volume 1, Number 1 of the Archives can be retrieved by sending an e-mail letter to LISTSERV@a and making the single line in the letter rea d GET KEMMIS V1N1 F=MAIL. For a table of contents of the entire ARCHIVES, send the following e-mail message to INDEX EPAA F=MAIL, tha t is, send an e-mail letter and make its single line read INDEX EPAA F=MAIL.The World Wide Web address for the Education Policy Analysis Archives is Education Policy Analysis Archives are "gophered" at To receive a publication guide for submitting artic les, see the EPAA World Wide Web site or send an e-mail letter to and include the single l ine GET EPAA PUBGUIDE F=MAIL. It will be sent


14 of 14to you by return e-mail. General questions about ap propriateness of topics or particular articles may be addressed to the Editor, Gene V Glass, Glass@asu.ed u or reach him at College of Education, Arizona Sta te University, Tempe, AZ 85287-2411. (602-965-2692)Editorial Board John CovaleskieSyracuse UniversityAndrew Coulson Alan Davis University of Colorado--DenverMark E. Thomas F. GreenSyracuse Alison I. Arlen Gullickson Ernest R. Aimee Craig B. Howley u56e3@wvnvm.bitnet William Richard M. Jaeger Benjamin Thomas Dewayne Mary P. Les Susan Bobbitt Anne L. Hugh G. Richard C. Anthony G. Rud Dennis Jay Robert Robert T.

xml version 1.0 encoding UTF-8 standalone no
mods:mods xmlns:mods http:www.loc.govmodsv3 xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govmodsv3mods-3-1.xsd
mods:relatedItem type host
mods:identifier issn 1068-2341mods:part
mods:detail volume mods:number 2issue 7series Year mods:caption 19941994Month April4Day 1515mods:originInfo mods:dateIssued iso8601 1994-04-15

xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam a22 u 4500
controlfield tag 008 c19949999azu 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E11-00022
0 245
Educational policy analysis archives.
n Vol. 2, no. 7 (April 15, 1994).
Tempe, Ariz. :
b Arizona State University ;
Tampa, Fla. :
University of South Florida.
c April 15, 1994
Policy questions : a conceptual study / Thomas F. Green.
x Research
v Periodicals.
2 710
Arizona State University.
University of South Florida.
1 773
t Education Policy Analysis Archives (EPAA)
4 856