93

DISCRETIONARY DECISION MAKING IN THEMANAGEMENT OF HUMAN SERVICE ORGANIZATIONS

Therese J. Dent, Ph.D.

Diplomate Clinical Social Worker, Lakeside Center;Adjunct Assistant Professor, st. Louis University,

St. Louis Missouri

Introduction

Human service program directors are faced withguidelines for program delivery usually created at a levelremoved from local conditions. As one consequence, programdirectors are often confronted with a difficult choice--Doyou follow the guidelines or do you deviate in order todeliver a good program? The research reported heredemonstrates that at least some of the "cornmon sense"adaptive behaviors of those in the field charged withprogram delivery are the product of individual leaderqualities and organizational dynamics.

Discretionary decision making is most likely to occurwhen the decision maker has acquired information andknowledge which goes beyond or adds to that whichstructured the program guidelines. Use of discretioninvolves the acquisition and utilization of informationunique to the local context that impels variation from amandated guideline. Several theorists have describedbehavioral decision models founded on the notion of limitedknowledge and bounded rationality (Simon, 1955; Etzionni,1967, 1986; Braybrooke and Lindblom, 1963). These arepragmatic modifications of the classic rational planningmodel which idealistically and impossibly calls forcomplete information and rationality. The ability toperceive and interpret reality from a local implementingcontext adds to the font of information available toprogram decision makers. This added informationfacilitates and encourages the use of discretion regardingcompliance with mandated guidelines at a local level.

Rigid adherence to rules, which may be more typical oftechnological organizations, is frequently experienced ascounter productive in human service organizations. It is acornmon experience in social work practice that the physicalwelfare and/or emotional well-being of certain clients maynot be best served by fitting them into the procrustean bedof program rules and regulations. The state of the art andthe science in human services is not sufficientlysophisticated to program delivery in a manner which takes

94

into account the emerging uniqueness of individuals.Genetic and environmental dynamics of individual clientsinfluence the effectiveness and efficacy of a program foran individual.

In similar fashion, genetic and environmental dynamicsinfluence decision making on an organizational level. Theidiosyncrasies of local cultural and political constraintsand demands foster discretionary decision making. An .examination, therefore, of both human and organizationalfactors which influence discretion is needed. Socialworkers by training may be more aware of the human factorsinvolved in discretion. This research focuses on theorganizational dynamics which also result in the use ofdiscretion.

Focus of the Research

Increased sensitivity to the complexities of programdelivery has been the result of recent implementation andorganizational studies (Pressman & Wildavsky, 1973;Scheirer, 1981; Meltsner & Bellavita, 1983; Palumbo, 1980).The need for accountability and legal compliance requiresstandardization of program design and procedures (Sabatierand Mazmanian, 1980; Williams, 1976). Standardization ofprogram delivery is counterbalanced, however, by arecognition that significant discretionary decision makingnot only occurs but is necessary at the point where staffinteract with clients (Lipsky, 1978). The notion of streetlevel bureaucracy suggests that discretionary decisionmaking can become an organization imperative in order toadjust program goals to local conditions and clients.

The literature on implementation and organizationalchange reveals further that local organizational leadersplay a prominent role in adapting program to context andclient needs (Nakamura & Smallwood, 1980). Katz and Kahn(1978; 532) indicate specifically that leadership emerges"as individuals take charge of .relating a unit or subsystemto the external structure or environment." This followsfrom Rice's (1963; 21) description of it as a break betweenparts of the system.

In general, ·if leadership is a boundaryfunction, the relationship between a leaderand his followers will depend to a majorextent on the leader's capacity to managethe relationship between the external andinternal environments in a way that willallow his followers to perform their primarytask.

i.I,j:

i

j:I

95

Creating a program environment in which 'staff cansuccessfully perform their primary tasks in a mannerconducive to achieving program goals is, therefore, thetask of a flexiole, leader. Policy and program guidelinesmust oe interpreted to adapt programs to local conditions.What is needed is a framework that would permit someordering of how this interpretation oy local leadersoccurs.

Tichy (1980) provides such a framework useful forunderstanding organizational dynamics to which a leadermust adjust. He descrioes all organizations asexperiencing ongoing interactive proolem cycles of apolitical, cultural, and technical nature. Thesesynergistic and simultaneous cycles provide a context forunderstanding organizations as they attempt to performtheir primary tasks and to ensure organizational survival.The problem cycles are interactive and trigger one anothersuch that none is ever totally resolved. A peak in onecycle precipitates a peak in another. Uncertaintyprevails, and important events may not be interpreted in amanner to reduce that uncertainty. Each cycle ischaracterized by a particular set of uncertainty creatingevents and change triggers new uncertainties.

The person attempting to balance these uncertaintieswith the constraints of program guidelines and the needs ofprogram clients is seen to be in a pivotal leadershipposition. He faces the challenges of forging congruentlinkages of policy goals and program guidelines with theidiosyncrasies of a local context and the behavior ofpersons within it. In such a situation, discretion andadaptation strategies become an obvious integral aspect ofdecision making. '

The research problem is to clarify some of the forcesthat result in variation from program guidelines despite amandated program design. '

The study explores three research questions. Given aninitial organizational context into which a program wasintroduced:

1.

2.

What organizational problem cycle(s)appeared dominant, and what is therelationship of problem cycle toadherence to guidelines for programdelivery?Do differences in leader behaviorqualities of a program directormak~ a difference in terms of theprogram's adherence to guidelinesfor program delivery?

96

3. What is the effect of the inter­action among problem cycle, qualityof leader behavior, and leaderactivity on program delivery?,

Methodology

The implementation processesro f el~ youtRemplo¥ment de~enstratioR pregrams sponsored b~the unitedStates Department of Labor were the research sites. Theseprograms were part of a national research project, theVocational Exploration Demonstration Project (VEDP). Assuch, they ,were required to follow a rigorous experimentaldesign. They shared a mandated, highly structured, andclosely monitored set of program guidelines, procedures andprocess. Despite this, variation in the delivery of theseprograms occurred. For purposes of this research, programdelivery is defined as the degree to which adherence toprogram guidelines was obtained by a local VEDP program.

The local context for the implementation of the elevenprograms varied widely. Program directors in these siteswere said to have faced problems imposed by at least twosets of constraints and demands: (1) those imposed by theformal policy mandate of the program which included anintended policy goal of increasing the employability ofyouth and specific guidelines for program delivery; and,(2) those imposed by the complexities of the organizationaland environmental context into which the policy wasintroduced.

The organizational contexts of the eleven programsvaried according to the size and type of sponsoringorganization. Some were sponsored by CETA organizationswhich were larger, well established federally fundedorganizations. Others were sponsored by well establishedprivate non-profit organizations which were larger thanVEDP in size of budget and number of personnel. The thirdcategory consisted of more recently or newly establishedprivate non-profit organizations which were smaller in sizeof budget and number of personnel; some of these evendepended on the program it sponsored for survival.

The planning, assisting and assessing activities forVEDP involved a partnership among three organizations. TheOffice of Community Youth Programs (OCYP) of the UnitedStates Department of Labor, Employment and TrainingAdministration served as the general manager of theproject. Saint Louis University Center for Urban Programs(SLU/CUP) contracted to develop and administer the researchdesign and was responsible for the collection and analysisof data. The National Alliance of Business (NAB)

I' .,,:;1

97

established administrative systems and provided technicalassistance.

Multiple sources of both primary and secondary datadescribing program implementation were collected bySLU/CUP. From multiple data sources available to theauthor, case studies of the eleven sites were constructed.These case studies described the context and process ofprogram delivery. The actions of local program directorswere also a focus of the case studies.

The experience of program delivery was considered tobe particularly appropriate for this exploration ofdifferences in program delivery for two reasons. As aresearch demonstration project, a major on-going effort hadbeen made to standa~dize program guidelines and monitorcompliance. Given the nature of the original federallysponsored research project, essentially an experimentaldesign with control group, a necessary aspect of thatresearch involved an assessment of whether and how the"experiment" was delivered as designed. Particular effortwas made to strengthen the program operations across sitesto facilitate the standardization of program guidelines anddelivery.

Local on-site program monitors were under contract toSLU/CUP for the purpose of routinely describing andassessing program operations. These monitor reportsprovided one of the major data sources for the casestudies. Monitors also completed five separate structuredprocess assessments of program implementation. Theseassessments contained numerous standardized indicatorswhich called for judgmental grading; these yieldedquantitative measures of adherence to program guidelines.

The use of the monitor assessments as the majorsource for this study has two major advantages. First, theopportunity to utilize mUltiple sources of data in a studyincreases the validity of that study through the process oftriangulation of data; in this study, monitor reports andprogram assessments, news reports, VEDP program documents,limited participant observation and unstructureddiscussions with VEDP oersonnel and clients weretriangulated. yin (1982) recommends the above describeduse of multiple data sources to establish validity.Second, because VEDP was a multi-site nationaldemonstration project, multi-site case studies could bedeveloped to explore variance in program delivery andprogram context. Herriot and Firestone (1983) recommendedthe use of multi-site qualitative studies as a ",ay tooptimize description and generalizability in research.

98

The case studies served to provide background data onconEext, Eo ~dent~fy problem cycles, and to describe leaderaction in the delivery of VEDP. Judgments were made, basedon the multiple sources of data available, regarding theexistence and emergence of problem cycles in a programcontext. Notations were also made in the case studiesabout the attempts of program directors to adjust toprogram problem cycles.

Tichy 1 S framework OD 4>l;gsl'liBat:iolial J?rgj;,J em cy£) es was)used as the method for organizing and analyzing dat~ in thecase studies. He concep~ualizedorganizat~ons'as havingthree interrelated adjustment cycles. A technical designor production problem cycle is triggered by uncertaintyover information and cooperative task completion. In VEDP,technical problem cycles were generally low since trainingand technical assistance were readi~y available andresources were more than adequate. Problems in this cycleonly appeared if program staff were not able to workcooperatively or did not obtain necessary information fromthe program director.

cultural problem cycles are triggered by differencesamong personnel over values and ideology. In VEDP, thiscycle arose most commonly with staff changes. Conflict andchange in attitude among staff over "doing a good job forthe kids" versus "going-by the book" were also common,especially given the reality of the short term nature ofthe project. In addition, where a lack of information ordirection from the program director did occur, work groupswhich valued finding solutions based on local conditionsfrequently formed.

Political problem cycles are the result of uncertaintyover who has the power to allocate rewards and status andto decide on goals of the organization. uncertaintyappeared in several sites regarding the autonomy of theVEDP program within its local sponsoring organization.Sponsoring organizatiol)s were prone to attempt to influencethe hiring and firing of program staff and the allocationof resources. Relations with the sponsoring organizationwere often unpredictable, which served to createadministrative crises for the program.

The case studies focused on describing theinterrelationship among problem cycles and the impact onthe organization. Tichy emphasizes that none of theproblems are ever resolved since organizations are dynamicand always undergoing shifts and changes. He suggests thatsuccessful change must rely on the ability to predictcycles and to channel and guide them.

99

A program director as a predictor and manager ofchange would need to demonstrate adaptable leader behaviorin order to channel and guide organizational cycles. Tichy(1980, 174) describes a manager of change:

At times (he or she must) be a politicalbuilder of coalitions, a power broker,and an influence manipulator coping withthe political cycles. At other times, heor she will be solving problems rationally,relying on "scientific- data and principlesto cope with the technical cycle. And atother times he or she will be an ideologicalleader. At still other times he or she maybe doing all three simultaneously.

To measure leader behavior characteristics, aconceptually expansive instrument was thought desirable.The Leader Behavior Description Questionnaire (LBDQ) wasselected.

Historically,~the LBDQ grew out of the Ohio State,ITni"':ll;:si ty Leadersh j p Stnd j es which produced two str'Ollglydefined dimensions of, leader behavior, consideration andinitiation of structure. me resultfr ef leader researchand experimentation by Stodgill (1959) and Halpin and Croft(1962) supported the idea that two dimensions were notsufficient to describe all the complexities of leaderbehavior. Ten additional qualities of leader behavior weresubsequently developed. Form XII of the LBDQ used in thisstudy provides leader behavior measures in each of thefollowing subscales: representation, demandreconciliation, tolerance of uncertainty, persuasiveness,initiation of structure, tolerance of freedom, roleassumption, consideration, production emphasis, predictiveaccuracy, integration and superior orientation.

The LBDQ is suggested for use by peers, superiors orsUbordinates. It is recommended that the questionnaire becompleted by a minimum of four persons familiar with thebehavior of a given leader. The instrument was completedby national and local staff who were familiar with thebehavior of local program directors. The LBDQ was given toas many local program staff as could be located (three tofive per program). It was also given to the followingnon-program staff: the National VEDP Co-Project Directormost familiar with the program in each site, the NationalVEDP Field Coordinator most familiar with the program ineach site, and the local Program Monitor.

100

Appropriate procedures were followed to validate thisinstrument with this population.

Results

In a national multi-million dollar researchdemonstration project where programs were mandated with 'little discretion and closely monitored to follow rules,why did such considerable variance occur? What can belearned about some of the factors that influencednon-compliance with program guidelines?

First, it was found that the size of the organizationsponsoring a program was related to the type oforganizational problem cycle that peaked in that program.Programs which were introduced into sponsoringorganizations that were small were found to be in politicalproblem cycle peaks. Uncertainty over who had the power tomanage the VEDP program was seen. This was especiallycritical in situations where the sponsoring organizationwas dependent on the VEDP program for its survival. Incontrast, programs that were introduced into largernon-profit sponsoring organizations tended to experiencecultural problem peaks. Uncertainty over organizationalvalues and program priorities were common. This was mostprevalent when CETA was the sponsoring organization.Conflicts in values and over the way things ought to bedone frequently occurred. A VEDP program functioned withmost autonomy and ease in a larger private non-profitsponsoring organization. Here VEDP was one of a number ofhuman service programs. Attitudes toward human servicewere compatible and competition for resources and power wasminimal.

Second, the number of problem cycles peaking at onetime is related to degree of adherence to programguidelines. Table 1 indicates that those programs whichwere experiencing mUltiple problem cycle peaks tended toadhere to guidelines for program delivery more than thoseprograms experiencing single peak problem cycles.

101

Table 1

Mean Score of Adherence to Program DeliveryGuidelines by Programs in Singleand Multiple Problem Cycle Peaks

Problem Cycle Peak Adherence to Program Guidelines

single Peak

pittsburghSan FranciscoMemphisTacomaNew Orleans

MUltiple Peaks

HaverhillLansingDulu.thNew York CityAllentownAtlanta

ProgramComposite Score

14.0112.8017.7812.8616.06

14.5120.8119.2219.0015.0217.41

Mean Score forSingle Peak Cycle

14.7

Mean Score forMultiple Peak Cycles

17.7

The data indicate clearly that programs which adheredmore clearly to guidelines also experienced peaks inmultiple problem cycles. Three of the four programs inmultiple peak problem cycles with the highest adherence toguidelines scores were experiencing overwhelming managementproblems. One program faced a political takeover and hadbeen totally shut down for a week by the mayor of the city.Another had severe staff management problems as a result ofthe sponsoring organization controlling personnel mattersand hiring unqualified staff. A third had inadequate andinconvenient classroom space with the result that bothstaff and clients were disgruntled. It appears that whereprogram quality was observed to be most in question,reported adherence to guidelines was high. The less aprogram appeared able to influence problem cycles and

102

adjust program dilemmas, the more security there appearedto be in following the rules.

Third, adherence to program guidelines correlated witha cluster of leadership dimensions in a negative fashion;here the data more resembled follower than leader behavior.Adherence to guidelines was negatively related to the LBDQleader dimensions of consideration, persuasion, productionemphasis and representation. Adherence to programguidelines was positively related only to superiororientation.

Fourth, programs which were able to remain in singlepeak problem cycles had program directors with higherleader behavior scores across all dimensions measured bythe LBDQ. (See Table 2.)

Table 2

LBDQ Dimensions of Leader Behavior Mean Scoresby Program Directors in Single and in

Multiple Peak Problem Cycles

VEDP II Program Problem Cycles

LBDQ Dimensions ofLeader Behavior

RepresentationReconciliationTolerance of uncertaintyPersuasionStructureTolerance of FreedomRole AssumptionconsiderationProduction EmphasisPredictive AccuracyIntegrationsuperior Orientation

Single PeakCycles

N = 5

12.619.328.339.137.034.041. 632.126.418.519.329.5

Multiple PeakCycles

N = 5

10.917.127.532.432.532.535.927.622.215.515.627.2

It is clear that measures of leader behavior among programdirectors are consistently higher across all dimensions inprograms with single peak problem cycles.

103

--.In summary, Tichy's framework of organjg.ationa-J...problem cycles did se~as a useful method for ordering ~

'and analyzing the mult~ple sources of descri €ive da.av ~ e. was oun that compliance with program

'mandates and guidelines appears to be influenced by the twophenomena postulated--the problem cycle context of theorganization and leader behavior characteristics amongprogram directors attempting to manage or cope with thoseproblem cycle contexts. The implications for social workpractice, procedure and assessment would seem substantial.

Implications for Social Work Practice

The information gathering, ordering and analyzingskills useful in social work practice with individuals andgroups are likewise useful for the management of humanservice organizations. They become the bases fordiscretionary decision making which is necessary fororganizational adaptation to environmental conditions. Themore information available beyond that at hand whenorganization and program rules were made, the morediscretion in decision making is likely to occur.

An attitude of going by the book takes place in orderto promote security in the midst of uncertainty. Thecapacity to be flexible and adaptable may enable one to usethe information available in one's environment to respondto uncertainties and prevent organizational problems fromarising. with the existence of relative program stability,discretionary responses to client needs may then also bemore possible.

When problems are high, adherence to rules becomesmore rigid. Following the rules is often equated withcovering one's flanks when mUltiple uncertainties exist.The experience of tightening rules when problems areincreasing is a common one.

The necessity for and advantages of having rules andguidelines is not meant to be ignored. The sense of orderand of a predictable program environment for staff andclients is highly valued, particularly for training newstaff. One may not know when discretionary decision makingmay be called for until one understands the basic valuesand goals intended to be implemented through programguidelines.

Nascent implementation theory is, however, suggestingthat policy is often better made by those charged with itsimplementation. In their study of educationaladministration, Meyer & Rowan (1977) describe howadministrators avoid maintaining controls over day by day

104

practices of staff so that the inconsistencies betweenlocal practices and institutional rules are not discovered.Rowan (1982; 62) explains:

This maintains the legitimacy ofexternal rules, which can work onlyif local personnel are granteddiscretion, avoids the discovery ofdeviance by local personnel, whobend external rules to fit thelocal situation, and keeps upappearances of local conformity.

Rowan (1982;62) further argues that "formal controls are aninappropriate mode for regulating an uncertain technolgy."This concept certainly applies to human serviceorganizations.

The split found to exist between local procedures andformal controls has critical implications for programassessment and evaluation. Procedures for assessingphenomena beyond rule compliance must be utilized.

In conclusion, discretionary decision making is anobvious extention of contingency theory. In the managementof human service organizations, the usefulness ofdiscretionary decision making is to decrease uncertainty,diminish organizational problems and thereby facilitateconditions for responsive human services. To legitimizediscretionary decision making would acknowledge thefrequent rub between rule compliance and responsive humanservice delivery. It ~ould also recognize the knowledge,skills and leadership role required of an effective socialwork administrator.

105

REFERENCES

Braybrooke, D. & Lindblom, C. E. (1963). A Strategy ofDecision. New York: Free Press.

Etizionni, Amitai. (1967). "MixedApproach to Decision Making,"Review, 27:385-392.

Scanning: A ThirdPublic Administration

----:"""'"'c-::-~-::-,-:-:-. (1986). "Mixed Scanning Revisited,"Public Administration Review, 46:8-14.

Halpin, A. W., & Croft, O. B. (1962). The Organizationalclimate of Schools. St. Louis: Washington University(mimeo) .

Herriott, Robert, & Firestone, William. (Feb. 1983)."Multisite qualitative policy research: optimizingdescription and genera1izability." EducationalResearcher.

Katz, Daniel, & Kahn, Robert. (1978), The SocialPsychology of Organizations, 2nd ed. New York: JohnWiley.

Lipsky, Michael. (1978) . "Standing the study of policyimp1ementatibn on its head," in Burnham and Weinberg,(eds.), American Politics and Public Policy.Cambridge, Mass.: M.I.T. Press.

Meltsner, A. J. & Bellavita, C. (1983). The Policyorganization. Beverly Hills: Sage Publications.

Meyer, John W., & Rowan, Brian. (1977). "Institutionalizedorganizations: formal structure as myth and ceremony."American Journal of Sociology, 83:340-363.

Nakamura, Robert T., & Smallwood, Frank.Politics of Policy Implementation.Martin Press.

(1980). TheNew York: st.

Palumbo, Dennis. (1980). "Introduction to symposium onoptimizing, implementing and evaluating pUblicpolicy." Policy Studies Journal, Vol. 8, No.7,special #3.

Pressman, Jeffrey L.,Implementation.Press.

& Wildavsky, Aaron. (1973).Berkeley: University of California

106

Rice, A. K. (1963). The Enterprise and its Environment.London: Tavistock.

Rowan, Brian. (1982). "Organizational structure and theinstitutional environment: the case of publicschools." Administrative Science Quarterly,27:259-279.

Sabatier, Paul, &. Mazmanian, Daniel. (1980). "Theimplementation of public policy: a framework foranalysis." Policy studies Journal, Vol. 8, Special#2, 538-560.

Scheirer, Mary Ann. (1981). Program Implementation:The Organizational Context. Bervely Hills: SagePublications.

simon, H. (1955). "A Behavioral Model of Rational Choice,"Quarterly Journal of Economics, 69:99-118.

Stodgill, Ralph M. (1959). Industrial Behavior and GroupAchievement. New York: oxford University Press.

Tichy, Noel M. (1980). "Problem cycles in organization andthe management of change," in John R. Kimberly, RobertH. Mills &.Associates. The Organizational Life Cycle.San Francisco: Jossey-Bass.

Williams, Walter, & Elmore, Richard (eds.). (1976) SocialProgram Implementation. New York: Academic Press.

Yin, Robert K. (1982). "Studying the implementation ofpublic programs," in Walter Williams et al., StudyingImplementation Methodological and AdministrativeIssues. Chatham, N.J.: Chatham House Publishers.