The Effect of Definitions of Policy on the Nature and ... · The Effect of Definitions of Policy on the Nature and Outcomes of Policy Analysis Varying interpretations of the word

The Effect of Definitions ofPolicy on the Nature andOutcomes of Policy Analysis

Varying interpretations of the word policy greatlyaffect how and where particular policies arecreated and implemented and, ultimately,whether their results are as intended.

EGON G. GUBA

in the literature on policy analy-sis, one can safely conclude that

the term policy is not defined in anyuniform way; indeed, the term is rare-ly defined at all. Authors of policypapers, monographs, and books as-sume that their readers know what theterm means. And, of course, theseauthors have their own implicit or tacitdefinitions in mind which shape whatthey have to say.

Yet these implied definitions varyconsiderably among authors and, mostlikely, among readers. This variationcan lead to miscommunication andconfusion, but there is another farmore serious problem: the particulardefinition assumed by the policy ana-lyst determines the kinds of policyquestions that are asked, the kinds ofpolicy-relevant data that are collected,the sources of data that are tapped, themethodology that is used, and, finally,the policy products that emerge. Un-fortunately, if the reader and the ana-lyst operate from different definitions,the reader will find the policy prod-

Egon G. Guba is Preswr, Sdool qOfEdu-cation, Indiana Urnr-sit, Bloom-groIndiana

This artice is based on a pap-periedat the annual meeurg of the AmevanEducational Resarce Asociation in NewOrleas in April 1984.

Author's note: I am indebted to DennisJPalunbo, ubo aarendy beads the Morri-son Insitaue for Public Policy at ArizonaState Unisrsty, for harng prwidd methe opporfunky during the smmer of1982 to develop the ideas eressed in thisanrie. His encourgement and s4porta gre fully acknouledged

OcroBEs 1984

1

63

". .. the analystcan predeterminewhat a policyanalysis willproduce by thesimple act ofchoosing theguidingdefinition."

ucts irrelevant at best and perniciousat worst Moreover, analyses thatemerge from different definitions dif-fer sufficiently among themselves sothat the analyst can predetermine whata policy analysis will produce by thesimple act of choosing the guidingdefinition. The political and ethicalimplications of this possibility are im-pressive.

I propose, first, to note the variousdefinitions ofpolicy that are implied inthe literature and, second, by applyingthese definitions in three policy are-nas, to note their differential effect.

The Three Exemplar PolicyArenasI have chosen three widely differentfields for these analyses:

1. 7be community correctons program witbin the criminal justice field(CJCC). The concept of communitycorrections is based on the assump-tion that retaining offenders in theirhome communities is more effectivethan sending them to prison. Felonswho have committed less serious felo-ny crimes (typically referred to asClass C felonies) are eligible for sen-tencing to programs such as drug re-habilitation, alcohol treatment, job

skills development, personal/psycho-logical counseling, marital/familycounseling, work release, victim resti-tution, halfway house, parole, proba-tion, vocational training, communityservice, or employment/placementcounseling. The assumption is that of-fenders who stay in close touch withtheir families are more likely to berehabilitated. Moreover, court caseloads and prison overloads may besubstantially reduced; and punish-ments may more closely fit the natureof the crime. Possible negative effectsinclude increased risk to the generalpublic, more plea bargaining (sincelesser offenders become eligible), andincreases in the web of control, that is,the number of persons who remainunder active supervision as comparedto a prison or release situation.

2. The program of education for allhandicapped students under PublicLaw 94-142 (SPED-special educa-tion). P.L. 94-142, passed in 1975, re-quires that all students, regardless ofhandicap, receive a free, appropriatepublic education in the least restrictiveenvironment consistent with the pu-pil's handicap. The least restrictiveenvironment is usually interpreted tomean that the pupil shall, insofar aspossible, be assigned to a regularclassroom or mainstreamed. Whenstudents are suspected of having ahandicapping condition, they are re-ferred, tested, diagnosed, and assignedin an elaborate form of due process inwhich parents must be intimately in-volved. If a child is found to be handi-capped, an individual educational pro-gram must be prepared that stipulatesin detail the instruction the child is toreceive, including placement, objec-tives, methods, timing, and so on. Theindividual educational program is re-viewed at least once every three years.Possible negative effects are stigma forthe child, excessive legalism, excessivepaperwork for teachers and adminis-trators, and instruction within the reg-ular classroom by teachers ill-pre-pared to deal with handicappedyoungsters

3. The process of accreditation ofcolleges and unie.rsities by regionalaccreditation associations such as theNorth Central Association, the Mid-States Association, and the like(HEAC). Colleges and universitieswishing to be accredited by a regional

64

association must apply for accredita-tion and undergo close scrutiny toensure that they meet certain definedstandards. Previously accredited insti-tutions must undergo inspection atleast every ten years; inspections maybe scheduled more frequently if nec-essary or if the institution applies for anew level of program. Candidate insti-tutions must prepare a lengthy self-study, which provides information rel-evant to the standards (such as facultyqualifications), and must submit toinspection by a site-visit team. A groupof evaluator-consultants retained bythe accrediting agency, a team is com-posed of experts competent to makejudgments in the field in which ac-creditation is sought, typically the fullgamut of colleges and professionalschools. The team leader prepares areport based on the individual find-ings of team members. This report,which contains an accreditation rec-ommendation, may be challenged bythe institution if it is seen as inaccurateor ill-founded A variety of grievancesteps exist for the airing of such ap-peals. Possible negative effects of theaccreditation process include externalcontrol over institutional policies, de-velopment of self-study reports thatput the institution's best foot forwardwhile covering up faults, and emphasisby the team on providing leverage toinstitutional administrators rather thanon overall accreditation issues.

Definitions of PolicyLet us consider eight different defini-tions of policy:

1. Policy is an assertion of intents orgoals

2. Policy is the accumulated stand-ing decisions of a governing body bywhich it regulates, controls, promotes,services, and otherwise influencesmatters within its sphere of authority.

3. Policy is a guide to discretionaryaction.

4. Policy is a strategy undertaken tosolve or ameliorate a problem.

5 Policy is sanctioned behavior, for-mally through authoritative decisions,or informally through expectationsand acceptance established over (sanc-tified by) time.

6. Policy is a norm of conduct characterized by consistenct and regularityin some substantive action area.

7. Policy is the output of the policy

EDIU(:ATI()NAI. L.FAI)FRSHIP

r

1

L

Figure 1. Policy Types, Determiners, Appearances, and Proximity to Point of Action for Bght Definitions of Poc.

Promily to PoitPolicy Types Policy Detenniners Definllon of Policy of Acl Policy Lodo LUe

Policy-in-intention CICC: State legislature 1. Goals or intents Distant EndsSPED: U.S. CongressHEAC: Commission on institutions

of higher education of aregional accrediting associ-ation

CICC: State criminal justice divi- 2. Standing decisions Intermediate Rulession and local supervisors

SPED: State education agency andlocal supervisors

HEAC: Regional association ad-ministrators

3. Guide to discretionary Intermediate Guidelinesaction

4. Problem-solving Intermediate Sets of tacticsstrategy

Policy-in-action CJCC: Street-level bureaucrats 5. Sanctioned behavior Close Expectationsand local supervisors

SPED: Street-level bureaucratsHEAC: Street-level bureaucrats

and regional associationadministrators

6. Norms of conduct Close Norms

CICC, SPED, HEAC: All of theabove 7. Output of the policy- NA Effects

making system

Policy-in-experience CICC, SPED, HEAC: Clients 8. Constructions based on Inside Encountersexperience

making system: the cumulative effectof all the actions, decisions, and behav-iors of the millions of people whowork in bureaucracies. It occurs, takesplace, and is made at every point in thepolicy cycle from agenda setting topolicy impact

8. Policy is the effect of the policy-making and policy-implementing svs-tem as it is experienced by the client.

These eight definitions are abbrevi-ated in Figure 1. which makes severalother points:

* Three polio' Otpes are implied bythese eight definitions Policv Defini-tions 1 through 4 treat policy aspolic-in-intention, that is, as statementsabout policy De)hnitions 5 through 7treat policy as polio'1-in-inplemnenta-tion, that is, as behaviors or activitiesthat are displaved in the process ofimplementing policy l)elinition 8treats policy as policT-in-a:pri'ience,that which is actually experienced bythe client

* Different pol/ic detIerine7-, are

also implied. Examples are given inFigure 1 for the three exemplar areas.Thus, for Definition 1--Goals or In-tents-the policy determiners areseen as high level agents-membersof the several state legislatures thathave enacted communitv correctionsprograms; the U.S. Congress in thecase of Public Law 94-142; and theComnmission of Institutions of HigherEducation within the various accredit-ing agencies. For Definitions 2through 4, policy determiners tend tocomprise lower level subordinateagents-the state criminal justice divi-sion together with the supervisors ofactual implementers (tspically re-ferred to as street-level bureaucrats')for CJCC: the state education agencywith local supervisors for SPED; andregional association administrators inthe case of HEAC For Definitions 5and 6. policy determiners are the op-erational implementers together withtheir immediate supervisors in allthree areas. For Definition 7. all levels

are involved, while for Definition 8 itis the clients' constructions that count

0 Different definitions have policyformulated at different distances fromthe point of action. Thus, per Defini-tion 1, policy is created far from thepoint of action; per Definitions 2through 4, at an intermediate point;and per Definitions 5 and 6, dose tothe point of action. Since Definition 7includes policy made at all levels, theconcept of proximity is not relevantDefinition 8 places policy formulationwitbin the point of action, since policybv this definition is the constructionsthat clients make out of their experi-ences.

* Finally, it is clear that these differ-ent definitions will produce policies ofdifferent complexions. In the case ofDefinition 1. policy looks like a set ofends. Definition 2 results in policystatements that look like ru/es, whileDefinition 3 results in guidelnes withbuilt-in discretion. Definition 4 resultsin a set oftactics. Definition 5 results in

(k:TOB)R 1984

eaectaons; 6 in norms; 7 in efects; The Impact of Various impact of each of the eight definitionsand 8 in encounters. These differences Deflnitions in terms of five categories.in policy appearance are further eluci- Figure 2 (in eight parts) summarizes 1. Policy question. What is the essen-dated below. the essential information about the tial question that the policy analyst

Figure 2. Polcy Question, Data Colected, Data Sources, Methodology, and Policy Products forBght Definitions of Policy in Three Exemplar Areas: CJCC, SPED, and HEAC.

Defhition 1. Poy is an assertion of oitenlt or goat

Item CCC SPED HEAC

Policy Question: What are the overall intents or What are the overall intents or What are the overall intents orgoals of the community correc- goals of PL 94-142? goals of accreditation in highertions program? education?

Data Collected: State and county level criminal jus- National, state, and local level spe- National, state, and local collegetice statistics, such as crime rates, cial education statistics, such as va- and university statistics, such asarrests, incarcerations, paroles, re- riety and rate of handicapping con- admission data, program selectioncidivism; plus: ditions; plus: data, employment data; plus:

needs assessments, strengths assessments, resource assessments, opinions of experts

Data Sources: Stakeholding audiences: general Stakeholding audiences: handi- Stakeholding audiences: facultypublic, victims, offenders, judges, capped students, parents, school and administrators, boards of trust-parole officers; state and county boards, principals, teachers, spe- ees, students, parents, employers;records; relevant documents: such cial service providers; national, position papers; national, state, Io-as white papers; court decisions; state, local education agency rec- cal college and university records;experts such as university profes- ords; position papers; court deci- court decisions; experts such assors of political science and foren- sions; experts such as professors university professors of highersics and leading practitioners. of special education and leading education.

practitioners.

Methodology: Statistical data analyses, interviews, observations, questionnaire studies, documentary analyses, hearings.

Policy Products: Listing of goals to be served; prior- Listing of goals to be served; prior- Listing of goals to be served; prior-itized goals, such as reduction in itized goals, such as free and pub- itized goals such as assisting insti-costs of operating the criminal jus- lic education for all, education in tutions to improve their programtice system, reduction in incarcera- the least restrictive environment, quality, inserting a quality floor intions, lowered recidivism, better fit reduction in stigma suffered by the higher education, protecting theof punishment to crime. handicapped, increased general ac- public against inferior institutions.

ceptance of the handicapped.

Definition 2. Pocy is a governng body's standing decisions by which it regulates, controls, promotes, services, and otherwiseinfluences mnatters within its sphere of authority.

Item CICC SPED HEAC

Policy Question: What rules should govern the deci- What rules should govern the deci- What rules should govern the deci-sions in implementing the commu- sions regularly made in implement- sions regularly made innity corrections program? ing PL 94-142? implementing higher education ac-

creditation standards?

Data Collected: Decision-making roles and func- Decision-making roles and func- Decision-making roles and func-tions such as state division of crim- tions such as state departments of tions such as review boards, evalu-inal justice, judges, probation offi- education, superintendents, direc- ator consultants, review teamcers, service providers. Types of tors of special education, class- chairpersons. Types of decisionsdecisions made by each. Timing of room teachers. Types of decisions made by each. Timing of deci-decisions. Impact of decisions made by each. Timing of deci- sions. Impact of decisions such assuch as public safety, recidivism, sions. Impact of decisions such as failure to accredit, loss of funds byrelief of prison overcrowding, plea least restrictive environment, stig- institution, appeal, court suit.bargaining. ma, grade achievement.

Data Sources: State and county community cor- State and local education agency Examining team reports, self-studyrections administrators, state and administrators, state and local edu- reports, consultant evaluators, cli-county statistical records, operat- cation agency records, operating ents.ing personnel, clients, other oper- personnel, clients (including par-ating community corrections pro- ents), classroom experiences.grams.

Methodology: Questionnaire studies, records Questionnaire studies, records Analysis of self-study and team re-analyses, site visits. analyses, site visits, evaluation re- ports, interviews with consultant

ports. evaluators and clients, question-naire studies.

Policy Products: Identification of needed decisions; Identification of needed decisions; Identification of needed decisions;recommendations for rules such as recommendations for rules such as recommendations for rules such asClass C felons are eligible to be parents must be afforded due institutions must prepare an ade-sentenced to community correc- process before children are placed quate self-study report, institutionstions programs, offenders who in any special education program, have the right to veto a proposedcommit another felony shall be im- local education agencies must es- evaluator-consultant.mediately resentenced to the state tablish a child-find program.penitentiary.

would try to deal with, having definedpolicy in a particular way? Policy ques-tions are given for each of the threeexemplar areas. That is, given Defini-tion 1, "Policy is an assertion of intentsor goals," the CJCC analyst would ask,"What are the overall intents or goals

of the community corrections pro-grams?" At the other extreme (Defini-tion 8), the analyst would ask, "Howwould the clients describe the policyof the community corrections pro-gram as a result of their encounterswith it?"

2. Dam colecaed. What kinds of datawould the policy analyst collect, giventhe policy question posed by a particu-lar definition of policy? For example,given Definition 1--goals and in-tents-the analyst might well collect,for CJCC, state and county criminal

Definition 3. Policy is a guide to discretionary action.

Item CICC SPED HEAC

Policy Question: What is the scope of discretionary What is the scope of discretionary What is the scope of discretionaryaction that can be tolerated in im- action that can be tolerated in im- action that can be tolerated in de-plementing the community correc- plementing PL 94-142? termining the accreditation statustions legislation? of a higher education institution?

Data Collected: Action roles and functions such as Action roles and functions such as Action roles and functions such asthose in Definition 2 above. Types those in Definition 2 above. Types those in Definition 2 above. Typesof actions taken by each. Impact of of actions taken by each. Impact of of actions, such as those in Defini-actions such as those in Definition actions, such as those in Definition tion 2.2. 2.

Data Sources: As in Definition 2.

Methodology: As in Definition 2.

Policy Products: Identification of key action roles; Identification of key action roles; Identification of key action roles;recommendations for expectations recommendations for expectations recommendations for expectationswith respect to each, recommen- with respect to each, recommen- with respect to each, recommen-dations for discretionary limits, dations for discretionary limits, dations for discretionary limits,"meta-rules," such as a probation "meta-rules," such as the school "meta-rules," such as evaluatorofficer may recommend termina- psychologist may recommend consultants may recommend non-tion of sentence with the approval placement in the gifted program accreditation if they believe the in-of the judge, a county criminal jus- for students whose l.Q. falls be- stitution will be unable to providetice administrator may determine tween 115 and 125, parents may funds for a viable program in thethe county's degree of involve- elect to keep an EMR child in a future, institutions may appeal if itment in the program: full, partial, regular classroom even though is believed that their situation hasor none. school staff recommend otherwise. been misassessed.

Definition 4. Policy is a strategy undertaken to solve or ameliorate some problem.

Item CICC SPED HEAC

Policy Question: What strategies can be followed to What strategies can be followed to What strategies can be followed tominimize the problems that arise minimize the problems that arise minimize the problems that arisein implementing community cor- in implementing PL 94-142? in implementing higher educationrections programs? accreditation processes?

Data Collected: Types and frequency of problems. Types and frequency of problems. Types and frequency of problems.Impact of each type of problem, Impact of each type of problem, Impact of each type of problem,such as prevents locating commu- such as assigns blacks dispropor- such as tends to misassess specialnity corrections facilities in estab- tionately to the EMR category and purpose institutions such as biblelished neighborhoods, encourages whites to the LD category, over- colleges and chiropractic schools,offenders to accept as valid the loads teachers with paperwork. biased toward re-accrediting for-classification imposed on them by merly accredited institutions eventhe court. in the face of program determina-

tions to the contrary.

Data Sources: Implementers, such as parole offi- Implementers, such as teachers, Implementers, such as drafters ofcers, judges; clients, other stake- administrators, special services self-study reports, evaluator con-holding audiences such as the gen- providers; clients; other stake- sultants; clients; other stakehold-eral public, legislators; state and holding audiences such as parents, ing audiences, such as students,local records, such as recidivism, state department of education employers; national, state, localsentence completion, plea bargain- monitors; national, state, local rec- records such as faculty credentials,ing; actual experience. ords, such as increases in the num- enrollment data, placement data.

ber and type of handicappedserved, amount of paper perserved client.

Methodology: Statistical data analyses, interviews, observations, questionnaire studies, records analyses (including exceptionreports and monitoring reports), "hearings."

Policy Products: Identification of common, recur- Identification of common, recur- Identification of common, recur-rent problems; sets of tactics for rent problems; sets of tactics for rent problems; sets of tactics fordealing with each, such as con- dealing with each, such as moni- dealing with each, such as usingducting community forums to toring black-white distribution of special purpose faculty as evalua-overcome fear of corrections facili- cases over the several special edu- tor consultants, emphasizing in-ties in neighborhoods, counseling cation categories, using computer structions to evaluator consultantsto overcome negative self-image. facilities to provide as much rou- to ignore previous accreditation in

tine paperwork as possible. reaching judgments.

(Continued on page 68.)

Delouil;m SI Pky saconed behaWrr, foinai through authoritative decisions, or infonnrlly through expectations andacc ace, esuablshed over (sanctied by) lmne.

Ner o CICC SPED HEAC

Policy Question: What expectations should be set What expectations should be set What expectations should be setfor implementers of community for implementers of PL 94-142, ei- for implementers of higher educa-corrections programs, either as a ther as a priori prescriptions or as tion accreditation processes, eitherpriori prescriptions or as modified modified through experience? as a priori prescriptions or as modi-through experience? fied through experience?

Data Collected: Implementation roles and func- Implementation roles and func- Implementation roles and func-tions such as parole officers, tions such as regular and special tions, such as evaluator consul-judges, special service providers; education teachers, administrators, tants, review board members; "ex-"expert" opinion about expecta- special service providers; "expert" pen" opinion about expectationstions reasonable for each role, that opinion about expectations reason- reasonable for each role, that is,is from counterpart implementers able for each role, that is from from counterpart implementersfrom other sites; constructions of counterpart roles at other sites; who have carried out other accred-role incumbents. constructions of role incumbents. itations; constructions of role in-

cumbents.

Data Sources: Role incumbents and counterparts; Role incumbents and counterparts; Role incumbents and counterparts;site observations; activity reports. site observations; activity reports; site observations; activity reports;

evaluation reports. evaluation reports; team reports.

Methodology: As in Definition 4.

Policy Products: Role definitions and corresponding Role definitions and corresponding Role definitions and correspondingexpectations, such as parole offi- expectations, such as regular class- expectations, such as evaluatorcers shall be at all times responsi- room teachers shall consult with a consultants shall inform universityble for knowing the whereabouts resource teacher when they are officials of the accreditation statusof their charges, service providers unsure about how to handle a to be recommended but not dis-shall follow the program plans they case, administrators should be cuss it, review boards shall nothave filed with the county commu- knowledgeable about every special overturn the recommendation ofnity corrections administrator. education child assigned to their the visitation team without input of

buildings. additional evidence.

Deibiion 6. Policy is a norm of conduct, charactenrized by consistency and regularity, in some substantive action area.

tem CICC SPED HEAC

Policy Question: What forms of conduct are consis- What forms of conduct are consis- What forms of conduct are consis-tently acceptable in agents who are tently acceptable in agents who are tently acceptable in agents who areimplementing the community cor- implementing PL 94-142? implementing higher education ac-rections program? creditation processes?

Data Collected: Implementation agents, such as Implementation agents, such as Implementation agents, such asjudges, probation officers, special regular and special education evaluator consultants, reviewservice providers; tried-and-true teachers, administrators, special board members; tried-and-truesatisfying practices of local agents service providers; tried-and-true satisfying practices of local agentsand counterpart agents elsewhere. satisfying practices of local agents and counterpart agents elsewhere.

and counterpart agents elsewhere.

Data Sources: As in Definition 5.

Methodology: As in Definition 4.

Policy Products: Descriptions of satisfying behavior Descriptions of satisfying behavior Descriptions of satisfying behaviorand corresponding norms, such as and corresponding norms, such as and corresponding norms, such asparole officers should check the teachers should keep paperwork evaluator consultants should checkwhereabouts of parolees sufficient- up-to-date to satisfy any monitor, the data presented in the self-ly often to be comfortable about administrators should reassign bor- study at random as a means of vali-the possibility of absconding, derline cases as needed to manage dation, evaluator consultantsjudges should accept plea bargains teacher loads or to take account of should be less concerned with ac-as necessary to keep their case special circumstances of the child. creditability than in providing le-load manageable. verage to the institution in solving

difficult problems.

justice statistics such as crime rates,arrests, incarcerations, paroles, and re-cidivism. He or she might conductneeds assessments, strengths assess-ments, and resource assessments aswell, and finally, might solicit the opin-ions of various experts about what theintents or goals should be. Paralleldata would be collected for SPED and

HEAC. At the other extreme, in follow-ing Definition 8-client construc-tions-the analyst would be interestedin collecting as many such construc-tions as possible

3. Data sources The data stipulatedabove cannot be found just anywhere;knowing what data are wanted makesit possible to stipulate where they

might be found. Sources may include(as they do for CJCC by Definition 1),groups of stakeholding persons, rec-ords, documents, position papers,court decisions, and groups of experts;others will be useful for other defini-tional situations. In the case of Defini-tion 8, the sources will be almostexclusively clients since It is their con-

68 ISOUCA rioNAt. LF.ADI�RSHiP

68 EDUCAFI()NA. ILEADERSHIP

Definiion 7. Poicy i the output of the polymalih system: the cumlaive elecd of al thke acdes, deci md hehekm ofthe mfnions of people who work in rbueaurcies. t occrn, takes plce, and is made at eery pot in the politcyde, from agenda se, to polcy impact. As such, poicr s an anJl c categor.

Item CICC SPED HEAC

Policy Question: What are the collective effects of What are the collective effects of What are the collective effects ofthe behaviors of the several imple- the behaviors of the several imple- the behaviors of the several imple-menters of the community correc- menters of PL 94-142? menters of higher education ac-tions program? creditation processes?

Data Collected: Relevant laws, court decisions, ex- Relevant laws, court decisions, Relevant standards, guidelines, ex-ecutive orders, administrative state department of education dir- emplar field reports and recom-guidelines, regulations, memoran- ectives, monitoring reports, guide- mendations, minutes of reviewda, instructions, behavior of street- lines, exemplars of common pa- board sessions, appeals, behaviorlevel bureaucrats; responses to the perwork, behavior of the several of administrators, evaluator consul-above by clients; opinions of ex- implementer categories; responses tants, and review board members;perts. to the above by clients and other responses to the process by stake-

stakeholding audiences such as holding audiences within the insti-parents; opinions of experts. tutions; opinions of experts.

Data Sources: Law libraries, state and local rec- Law libraries, national, state, and Professional literature, accreditingords, practices of agents at state local education agency records, association records, practices ofand county sites and counterpart practices of agents at state and lo- agents within the association andagents elsewhere, "experts" such cal levels and counterpart agents at the candidate site, "experts"as professors of political science elsewhere, "experts" such as pro- such as professors of higher edu-and forensics and leading practi- fessors of special education and cation and leading practitionerstioners. leading practitioners. such as university administrators.

Methodology: Statistical data analyses, documentary analyses, interviews, observation, unobtrusive measures, investigative in-quiry techniques.

Policy Products: Descriptions of collective effects, Description of collective effects, Description of collective effects,such as widening the net of social such as determining who is "teach- such as stigmatizing nonaccreditedcontrol to include offenders who able," placing more emphasis on institutions, diminishing the rangemight otherwise be released, in- keeping auditable records than on of creativity in programming thatcreasing the range of acceptable overcoming student problems. institutions will be willing to risk.plea bargaining.

Definition 8. Policy is the impact of the polcymaking and podicy-implementing system as it is experienced by the cient.

Item CdCC SPED HEAC

Policy Question: How would the clients describe How would the clients describe How would the clients describethe policy of the community cor- the policy emanating from PL 94- the policy of the accreditationrections program as a result of 142 as a result of their experiences process as a result of their experi-their experiences with it? with it? ences with it?

Data Collected: Client constructions

Data Sources: Clients

Methodology: Interviews, observation, unobtrusive measures, investigate inquiry techniques.

Policy Products: Client constructions, such as coer- Client constructions, such as stig- Client constructions, such as nega-cion, control, reluctance to pro- ma, babysitting rather than help, tivism, overquick judgment, inade-vide entitlements such as employ- reluctance to provide entitlements quate understanding, reluctance toment opportunities, addiction ther- such as adequate diagnosis, thera- provide entitlements such as fairapy. py, prosthetic devices. hearings study of special circum-

stances.

structions that define policy.4. Methodology. Different data

sources imply different methodologiesin tapping them. Thus, personal con-structions may require interviews andobservations, records may requireanalysis, statistics may require numeri-cal analysis, and so on. Hearings are acommon form of input device to col-lect data from interest groups.

5. Polio' products. This row of Fig-ure 2 is by far the most interesting,

OCTOBER 1984

since it is here that one can bestappreciate the vast differences that dif-ferent policy definitions can produce.In CJCC, for example, a definition ofpolicy as intents or goals results in apolicy analysis yielding recommenda-tions or prioritizations among goalssuch as reduction in costs of operatingthe prison system, lowering recidivismrates, providing a better fit of punish-ment to crime, and so on. This productis quite different from, say, Definition

2, which results in recommendationsof rules (for instance, Class C felonsshall be eligible for consideration forthe community corrections program);,Definition 3's guidelines (a probationofficer mav recommend terminationof sentence with the approval of thejudge); Definition 4's sets of tactic(conducting communin, forums toovercome fear of having correctionsfacilities placed in neighborhoods);Definition 5's expectations (judges

69

should accept plea bargains as neces-sary to keep their case lod manage-able);, Definition 6's norrs (judgesshould accept plea bargain4 as neces-sary to keep their case loads manage-able); Definition 7's effects (wideningthe net of social control to includeoffenders who might otherwise be re-leased); or Definition 8's client con-structions (coercion and control).

A Test for Policy ProductsThe examples appearing in the PolicyProducts row of Figure 2 can reason-ably be construed as policies. Some ofthe examples may sound more likepolicies to the reader than others,although this apparently greater valid-ity is as much a function of what thereader normally takes to be policy as it

.is of any property intrinsic to theexamples themselves. Nevertheless, itis reasonable to suggest a test; and thereader is invited to apply the followingone. Using the sentence stub, "It is thepolicy of this entity to ... " completethe sentence by adding the variousexamples of Figure 2. The term entitycan, of course, be understood to meana program, a unit, an agency, a legisla-ture, or any other body or institution.To illustrate, let us take one examplefrom Figure 2 for the special educa-tion (SPED) definition of policy:

Definition 1: Intents or Goals. It isthe policy of this entity to provide afree, appropriate public education forall.

Definition 2: Standing Decisions. Itis the policy of this entity to affordparents due process before placingany child in a special education pro-gram.

Definition 3: Discretionaiy Guide toAction. It is the policy of this entity topermit a school psychologist to rec-ommend placement in a gifted pro-gram of any child whose I.Q. fallsbetween 115 and 125.

Definition 4: Problem-Solving Strat-egy. It is the policy of this entity to usecomputer facilities to provide as muchroutine paperwork as possible.

Definition 5: Sanctioned Bebavior.It is the policy of this entity to haveregular classroom teachers consultwith a resource teacher when unsureabout how to handle a case. (Thispolicy statement differs from numbers2 and 4 above in that it is devised by

implementers close to the point ofaction; it is less a rule or strategy than aself-developed rule-of-thumb; the be-havior is sanctioned by virtue of thesuccessful experience of imple-menters.)

Definition 6: Norms of Conduct. Itis the policy of this entity that adminis-trators should reassign borderline cas-es as needed to manage teacher loadsor to take account of the special cir-cumstances of the child. (This policystatement differs from number 5above in that norms are peer-devel-oped accommodations to reality thathave been worked out through mutualnegotiation, even if only implicitly.)

Definition 7: Output of the Policy-Making System. It is the policy of thisentity to place more emphasis onkeeping auditable records than onovercoming student problems. (Thisanalytic statement deduces policyfrom observations of behavior, analy-sis of records, and so forth. It may, asin this case, be an unwritten policythat is deduced but one that is opera-tive in controlling the behavior ofpolicymakers and implementers at alllevels.)

Definition 8. Client Constructions.It is the policy of this entity to denyentitlements such as adequate diagno-sis, therapy, or prosthetic devices.(This statement, like number 7 above,is a deduction, based not on analysisbut on client experience. Clients hav-ing made such constructions will act asthough these deductions were in factthe policy of the organization.)

ConchsionsSeveral conclusions may be drawnfrom the above analysis:

* It is nonsense to ask the question,"What is the real definition of policy?"No one of the definitions given above(or any others, for that matter) canstake a legitimate claim to priority. Alldefinitions are constructions; nonecan claim tangible reality. Virtually anypolicy definition must be admitted solong as its proposer can make a ratio-nal case for his or her particular usage.

* Nevertheless, not all definitionsare equal in their consequences forpolicy analysis. Each definition callsfor its own data, sources, and methods,and produces unique outcomes. Ex-plicitly or tacitly, each different defini-

tion has an enormous impact on theprocesses and products of policy anal-ysis.

*The particular definition that ananalyst elects to use should depend onthe purpose of the analysis. From theperspective of the analyst some defini-tions will always be better than others.When Congress is seeking to set broadgoals for the education of handi-capped children, a definition of policyas goals or intents is clearly moremeaningful than any other; indeed,certain others would not even be pos-sible. And there can be no client con-structions unless a program is alreadyin place, for example.

*What constitutes a better defini-tion is of course a matter of values.Thus, the selection of a particular defi-nition is a value choice. Certain defini-tions of policy are likely to be moreconsonant with the analyst's valuesthan others. The political implicationsof this choice should not be over-looked; that is, policy analysis may beused to advance certain values and tocheck others. The authoritarian policyboard is not likely to be much interest-ed in implementer norms or clientconstructions, for example. Policyanalyses are not value free; the out-come depends heavily on the defini-tion of policy that the analyst explicitlyor implicitly follows.

* To be ethical, the policy analyst isobliged to point out the particulardefinition used in an analysis and tocharacterize its consequences for thevariety of stakeholding audiences con-cerned with the analysis. The best wayto do that is to show, as in Figure 2,what might happen if different defini-tions of policy were adopted. The besttime to make this disclosure isprior toundertaking the analysis so that allstakeholding audiences may knowwhat is involved and arrive at someconsensus before the analyst pro-ceeds. If consensus is not possible, theanalyst may need to develop differentanalyses for different audiences; rec-onciliation and accommodation thenbecome the audiences' problemE

'Michael Lipsky, Street-Level Bureaucra-cy (New York: Russell Sage Foundation,1980)

EDUCATIONAL LEADERSHIP

Copyright © 1984 by the Association for Supervision and Curriculum Development. All rights reserved.