Key Information While most of the top public sector issues ...€¦ · Key Public Sector IS Issues...

Key Public Sector IS Issues

Key InformationSystems ManagementIssues for the PublicSector^

By: Sharon L. CaudleSchool of Information StudiesSyracuse UniversitySyracuse, New York 13224

Wilpen L. GorrSchooi of Urban and Public AffairsCarnegie Meiion UniversityPittsburgh, Pennsyivania 15213

Kathryn E. NewcomerDepartment of Pubiic

AdministrationSchool of Government and

Business AdministrationGeorge Washington UniversityWashington, D.C. 20052

AbstractThe public sector has multiple, conflicting, andoften intangible goals. It produces "publicgoods" for problems that should be solved (likecrime and poverty), even though these problemsmay have no known feasible solutions; and it isheavily impacted by politics and bureaucratic redtape. These and other features of the public sec-tor make it potentially a much different setting forIS management than the private sector. This ar-ticle reports on the first national survey of publicsector managers identifying their most importantIS issues. The survey, covering respondents fromfederal, state, and local governments, drew uponprior survey research in the private sector andthe literature on pubiic/private sector differences.

This research was sponsored by the National Association ofSchools of Public Affairs and Administration and by theauthors' affiliated institutions during the time of study: theCenter for Governmental Services. Auburn University; theSchool of Information Studies. Syracuse University: theSchcxjl of Urban and Public Affairs, Carnegie Mellon Univer-sity: and the Department ot Public Administration. GeorgeWashington University.

While most of the top public sector issues alsoappear on the top private sector issue lists, therankings show a tag in public IS development ascompared to the private sector. Perhaps the mostinteresting results of the survey, however, arefrom a deeper analysis. At the main effects-level,we have preliminary evidence that (1) middle-level (instead of top-level) public managers arecritical for IS technology development; (2) smaltgovernment agencies are more interested in IStechnology transfer than targe ones; (3) govern-ments with a lot of red tape tend to have flexibleIS; and (4) local government IS issues are drivenby transaction processing whiie state and federalgovernments have tS more suitable for their over-sight mission.

Keywords; MIS managers, opinion survey, infor-mation systems management, keyissues, management priorities

ACM Categories: CO, H.O, K.6.0

IntroductionMIS Quarterly has published five articles assess-ing the importance of information systems (IS)issues to managers. Three surveyed Society ofInformation Management members (Ball andHarris, 1982; Brancheau and Wetherbe, 1987;Dickson, et al., 1984), one used a sample oforganizations from the St. Louis, Missouri, area(Hartog and Herbert, 1986), and another em-ployed interviews conducted with top managersin 49 Irish organizations (Moynihan, 1990). Thesestudies identified key issues that support deci-sion making by "businesses and governmentagencies . . . [affecting] profitability and effec-tiveness," and the information profession inregard to "commitment of its limited manage-ment, research, and educational resources"(Dickson, et al., 1984, p. 135). The surveys inthese studies are limited, however, to the privatesector. Their public/non-profit sector samplesizes range only from one to 19 respondents persurvey.

The contribution of this article is that it is the firstnational survey oi public managers' ratings of ISissues. Conducted in mid-1988, the surveysampled high-level public managers in federal,state, and county agencies. A comparison of ourimportance-rating results with those from the

MIS Quarterly/June 1991 171


previous private sector surveys shows manysimilarities between the two sectors but alsosome pronounced and critical differences.

The next section reviews the underlying factorscontributing to public/private sector differencesand leads to several uniquely public IS issues.Subsequent sections summarize the survey in-strument, its administration, and overall surveyresults. Next is a comparison of the survey resultswith previous private sector surveys, followed bya close look at the variations within the public sec-tor survey.

Differences in Pubiic andPrivate Sector InformationSystemsRainey, et al. (1976) summarize the literature ondifferences between public and private organiza-tions. Stated in terms ofthe public sector's char-acteristics relative to those of the private sector,these differences include:

1. Environnnental Factors: Less market ex-posure (and therefore more reliance on appro-priations) resulting in less incentive forproductivity and effectiveness, lower allo-cative efficiency, and tower availability ofmarket information; more legal and formalconstraints; and higher political influences, in-cluding impacts of interest groups and needfor support of constituencies.

2. Organtzatlon/Environment Transactions:More mandatory actions due to the uniquesanctions and coercive powers of govern-ment; wider scope of concern and signifi-cance of actions in the "public interest";higher level of scrutiny of public officials; andgreater expectation that public officials actfairly, responsively, accountably, andhonestly.

3. Internal Structure and Processes: Morecomplex criteria (e.g., multiple, conflicting,and intangible); managers with less decision-making autonomy, less authority over subordi-nates, greater reluctance to delegate, and amore political role for top managers; more fre-quent turnover of top managers due to elec-tions and political appointments; difficulties indevising incentives for individual perfor-

mance; and lower work satisfaction andorganizational commitment.

Using these general differences and some propo-sitions on public information systems (Bozemanand Bretschneider, 1986), we proceed to iden-tify IS issues unique to the public sector.

Proprietary versus shared ISPrivate sector firms often treat informationtechnology as proprietary, for use as a com-petitive advantage in the market place (Camillusand Lederer, 1985; Keen, 1988; King, etal., 1988;Lucas, 1986; Marchand, 1985). In contrast,government provides "public goods," such aseducation, public health, public safety, and na-tional security (Savas, 1982). By definition, it isnot possible (nor desirable) to exclude consump-tion of public goods. Thus, while Brancheau andWetherbe (1987) introduced the issue of usingIS for competitive advantage, we introduce apublic sector counterpart.

Technology Transfer: Agencies shouldfind ways to increase sharing of applica-tions and technical assistance with otheragencies having similar functions or pro-viding similar services.

GoalsThe most frequently mentioned difference be-tween the public and private sectors concernsunderlying goals (Self, 1977). Market signals andprofit guide the private sector. In contrast, publicsector organizations face multiple goals, manyof which are intangible or in conflict with oneanother. Problems faced in the public sector areones that "should" be addressed (e.g., reducecrime) even though they may have no knownsolutions; whereas, problems faced in the privatesector are driven by feasibility considerations(Downs and Larkey, 1989).

A consequence of these difficulties is that infor-mation requirements are inherently more difficultand unstable in the public sector (Caudle andNewcomer, 1987; Newcomer and Caudle, 1986).We therefore include the following Issue.

identification of information Require-ments: Agencies should develop syste-matic processes to identify and prioritizeinformation requirements.

172 MIS Quarterly/June 1991


This issue is comparable to, but somewhat nar-rower than, the top-rated issue in all of the privatesector surveys, IS planning.

Bureaucracy and paperworkPartly in response to the inherent difficulties inpublic sector information requirements and theresulting proliferation of forms and records, thePaperwork Reduction Act of 1980 sought to in-stitutionalize information resource managementin the federal government (Caudle, 1988; New-comer and Caudle, 1986). Some states are alsoadopting information resource managementpractices. We therefore include the followingissue.

Information Resource Management:Agencies should adopt informationresource management concepts (e.g., in-tegration of information technology man-agement, records management, and otherinformation management areas) and as-sign a centrai institutional responsibility forinformation collection, processing, dissem-ination, policy, and operations.

Political influencesPolitics disrupts the long-range planningnecessary for information resource management.Downs (1967) points out that elected officialstypically have shorter planning horizons thanbureau officials and are more sensitive to publicpressures. Political cycles cause periodicchanges in top-level management and programpriorities (Nigro and Nigro, 1977; Self, 1977).Bozeman and Bretschneider (1986) state that,"The appropriations process generally operateson an annual basis. Congress turns over everytwo years, and the presidency is subject tochange every four years. Constant pressure ex-ists to achieve quick results—results that canhelp the agency receive a larger share in the nextround of appropriations; results that can help inreelection" (p. 479).

To offset political influences, public sector ISmanagers and users should place importance onplanning mechanisms that provide continuity forlong-range goals. We therefore include the follow-ing Issue.

Political infiuences: New planning andbudgeting procedures should be devel-oped to reduce the Impacts of short-runpolitical priorities and the lack of politicalcontinuity that can undercut comprehen-sive and long-term IS planning.

Another implication of political cycles is that fre-quent leadership and program changes cancause discontinuities in basic data element defini-tions (e.g., as welfare recipient classes change,components are added or subtracted to taxbases). Thus, time series and other data needspecial documentation to allow interpretation andpreserve information value (Gorr, 1986). Wetherefore include the following issue.

Data Discontinuity Between Administra-tions: IS management should find ways toprovide for long-term continuity of datadefinitions and data aggregate classifica-tions, or a historical record of changes, tomaintain the usefulness of historical data.

The Pubiic Sector SurveyTable 1 lists the 37 issues in the survey bycategory, and the Appendix contains the portionof our survey instrument defining these issuesas assertions appropriate for Likert-scaleresponses. We used a five-point scale where 1is "not important at all" and 5 is "extremely im-portant." We defined importance in terms ofmanagement support, attention, and otherorganizational resources required to ensure thesuccess of IS over the typical government budgetplanning horizon of two or three years. Issuesunique to this survey are marked by an asterisk;however, issues like #1, "educating elected of-ficials," which have a comparable private sec-tor issue already identified in the literature(educating senior personnel), are not marked asunique.

Respondent selectionFor federal agencies, we mailed surveys to allexecutive-branch officials identified as "informa-tion resources manager operational contacts" inthe 1987 directory provided by the U.S. GeneralServices Administration. This sample providedcomplete coverage of federal personnel directly



Table 1. Public Survey Information Systems Issues

A. Strategic Management Concerns

Educating elected officialsPolitical influenceLong-term planning mechanisms*

Comprehensive planning integrationInformation resource management*Research and development

B. Operational and Tactical Management Concerns

Identification of information requirementsMeasuring IS effectiveness

Procurement and contracting*Oversight requirements*

C. Applications Development and Maintenance

Technology transfer"Technology transfer mechanisms*Mainframe software development

D. Personnei issues

Civil service constraints*Meeting mainframe personnel needsMeeting microcomputer applications

personnel needs

E. Data Management

Distributed data processingData securityComputer viruses'Data integrity

F. Information Technologies

Integration of technologiesEnd-user computingMainframe computingOffice automationDecision support systems

Database management system impact*Application generatorsSoftware maintenance*

Substitution of technology for trainedpersonnel*

External data sources*Database administrationPublic and private information interests*Data discontinuity between administrations*

Artificial intelligence and expert systemsMicrocomputer software managementTechnology securityTelecommunications technology

* Indicates issues unique to this survey.

responsible for information resource manage-ment programs, policies, control, and oversight.

For state agency representation, we sent surveysto officials in charge of the 50 state data process-ing centers. We aiso had the survey completedby participants of two information systems plan-ning workshops sponsored by the Ohio stategovernment and by members of the New YorkInformation Resource Forum (see Simmons,1988). The latter two groups constitute samplesof convenience made up largely of top-level

managers from the user community who havesubstantial distributed, satellite, or end-user com-puting responsibilities in their own agencies.They nevertheless represent all state governmentbranches {executive, legislative, and judicial) anda great number of program areas.

Finally, surveys were sent to all counties ex-ceeding 250,000 population and to a probabilitysample of counties with less than 250,000population provided by the National Associationof Counties.



Instrument pretesting andsurvey administrationPretesting consisted of mailing the draft surveyinstrument to a limited number of respondentsin two levels of government. In the federal govern-ment, we sent it to four officials, two at the depart-ment level and two at the agency levei. At thestate level, we sent it to the head of a state's datacenter, to an agency head of a department ofanother state who had a satellite computer centerin his agency, and to the director of an informa-tion system consortium of state agencies. Result-ing revisions included sonne changes in thewording of issues and dropping a difficult rank-ing of issues within categories.

The survey was mailed in April 1988. Data col-lection was terminated five months later. Approx-imately one month after the initial mailing, asingle follow-up mailing including another copyof the survey instrument was sent to non-respondents.

Response ratesWith the exception of county officials, well over50 percent of those sampled responded: (1)federal responses were 102 out of a total of 172surveyed for a 59 percent response rate; (2) state-level data processing officials returned 32 out of50 sent for a 64 percent response rate; (3) theOhio workshop members returned 39 out of 60surveys for a 65 percent response rate; (4) theNew York forum members returned 44 out of 55surveys for an 80 percent response rate; and (5)county respondents returned 137 out of 717 sentfor a 19 percent response rate. Forty-nine per-cent of those sampled from counties with a popu-lation of less than 50,000 responded; 10 percentof those sampled from counties with a popula-tion between 50,000 and 100,000 responded; 15percent of those sampled responded from coun-ties with a population between 100,000 and250,000; and 20 percent responded of thosesampled from counties with a population of over250,000.

Findings Regarding theIS Management IssuesTable 2 shows the percentage of "important" or"extremely important" (referred to as "impor-tant" hereafter to simplify exposition) responses

for the total sample. A number of the issues ratedlowest in the total sample (e.g., data discontinuitybetween administrations, research and develop-ment, and artificial intelligence/expert systems)are highly rated by certain respondent segments.A later section of this article presents segmentedrankings. First, the top 10 issues for the totalsample are discussed.

#1 Integration of Technologies: Ninety-one per-cent of the respondents agreed on the importanceof IS management ensuring that current andfuture data processing, telecommunications, andoffice automation technologies are integrated toprevent incompatibility. Integration of tech-nologies is the number one issue across allrespondent categories except for those in lowermanagement offices, small agencies, and cen-tral state agency offices who selected com-prehensive planning integration as their numberone issue.

#2 Comprehensive Planning Integration:Eighty-six percent of all respondents rated as im-portant the issue that there should be a processthat integrates information resource planning withthe agency's overall planning process. This issuereflects the increased emphasis on IS planningin the past several years.

#3 Information Requirements Identification:Seventy-eight percent of all respondents said thatan important issue is agencies developing sys-tematic processes to identify and prioritize infor-mation requirements. Emphasis on requirementsanalysis reflects the difficulty of public sectorproblems; i.e., their multiple and intangiblecriteria, multiple and conflicting interest groups,and lack of feasible solutions.

#4 End-User Computing: Seventy-eight percentof all respondents believe it is important thatagencies provide and increase support, such asInformation centers and standardized hardwareand software, for end-user computing. The ex-pansion of office automation and end-user ap-plications has increased the importance ofsupport for computer and applications literacyand standardization of applications andmachinery.

#5 Office Automation: Seventy-seven percentof all respondents agreed on the importance ofIS managers taking a leadership role in planning,implementing, and managing office automation.No longer are office automation applications



Table 2. Percentage of "Important" or "Extremely Important" Responses for the Total Sample

Issue

Integration of technologiesComprehensive planning integrationInformation requirements identificationEnd-user computingOffice automationData securityLong-term planning mechanismsDatabase management system impactDistributed data processingSoftware maintenance

Telecommunications technologyDatabase administrationMainframe software managementTechnology transferEducating elected officialsTechnology transfer mechanismsSubstituting technology for personnelData integrityMeasuring IS effectivenessInformation resource management

Political influenceCivil service constraintsApplication generatorsMicro software managementProcurement and contractingMainframe computingMicrocomputer personnel needsOversight requirementsMainframe personnel needsExternal data sources

Computer virusesData discontinuity between administrationsPublic/private information interestsDecision support servicesTechnology securityResearch and developmentArtificial intelligence/expert systems

TotalSample

91867878777473707070

69686867666666666663

63635858565353535050

49484747424238

viewed as separate and apart from more tradi- availability through appropriate protocol and ac-tlonal IS in the organization. cess controls. Our discussions with selected

respondents indicated that public sector mana-#6 Data Security: Seventy-four percent o gers would like to find ways to cut through redthe respondents rated as important the issue that ^ ^agencies should balance data security and data

176 MIS Quatierly/June 1991


#7 Long-Term Planning Mechanisms: Seventy-three percent listed as important that agenciesperiodically submit and approve informationresource plans before they submit plan-linkedbudget requests. This results in Improved long-term decision making.

#8 Database Management System Impact:Seventy percent of the respondents rated as im-portant the need for IS to increasingly usedatabase management systems to enhance ap-plication development and maintenance. Theflexibility provided by such systems is importantfor the ill-structured and dynamic set of problemsfaced by public organizations.

#9 Distributed Data Processing: Seventy per-cent also agreed on the importance of the needfor policies that support the advantages ofdistributed data processing while maintaining theintegrity of central IS.

#10 Software Maintenance: Lastly, seventy per-cent of all respondents said that an importantissue is that IS managers need to find ways ofimproving productivity in software maintenance.This result likely reflects the preponderance ofcentralized data processing respondents in thesample.

Comparison With PrivateSector ResearchTable 3 compares the top 15 issues from four ofthe previously published private sector surveysand this article's public sector survey. The fifthprivate sector survey (Moynihan, 1990) usesopen-ended interview questions and thus is noteasily incorporated into an item-by-item com-parison. Discussion will focus on the top 10issues of the public sector survey list. An addi-tional five issues are included in the table (#11to #15), however, to show relationships betweenthe top 10 public issues and private sector surveytrends. For instance, office automation was the#6 issue in the 1982 private sector survey, butthen fell to #12 in 1984, #11 in 1985, and off thetop 15 list in 1986. In contrast, office automationis the #5 public sector issue. Italics indicateissues that are on both private and public sectorsurveys. Issues that have some portion modifiedfor sector differences (e.g., educating elected of-ficials versus educating senior personnel) arecounted as the same. To facilitate comparisons,

some issue labels are changed in individualstudies to make them consistent throughout thetable. At times these changes shade meaning,but we believe that overall meaning is preserved.Issues in our opinion that are unique to any par-ticular study in this table are marked withasterisks.

First, looking down the public sector survey's col-umn, seven out of 10 top issues also appear onthe private sector surveys' top 10 issues. Noneof the top 10 public sector issues, however, isuniquely "public." (We have to go down to the14th-ranked public sector issue, technologytransfer, to find the single uniquely public sectorissue.) Linking IS planning and budgeting (#7),appearing only on the public survey list, couidalso be an issue for private sector firms.Budgeting, however, does play a much more cen-tral resource allocation role in the public sectorbecause there is little equivalent to marketsignals. Software maintenance (#10) is also ap-plicable to both sectors. Thus, although some dif-ferences exist, there is considerable agreementbetween the two sectors on the top 10 issues.The rankings of issues, however, is markedly dif-ferent. The public sector appears to be laggingthe private sector in the development of somemajor IS issues.

Integration of technologies was the #1 public sec-tor issue in the survey. It has been on a declinein the private sector, however, ranking #3 in 1984,#7 in 1985, and #10 in 1986. We speculate thatthe private sector has been getting this issueunder control. The public sector, however, ishamstrung by red tape and procurement rulesleading to long delays and divergent mixes ofequipment.

Aligning IS with agency goals, #2 on the publicsector list, is in line with the private sector rank-ings, which were #7 in 1984 and #2 in 1965. ISplanning, #3 on the public sector list, has con-sistently been #1 on the private sector lists. Thelower rating on the public sector list may be theresult of only recent vigorous activity in ISplanning.

End-user computing (#4), has recently declinedon the private sector lists. It was #11 in 1982, #2in 1984, #5 in 1985, and #6 in 1986. We speculatethat it ranks higher in the public sector becauseof (1) the relative ease of purchasing microcom-puters versus mainframe computers in the public



mCDc!2cta

OCocaroaE13Cra

(0

Ea>*^tnc_ora

oc

raI -

ffl

CO g

^2c c.O) 0)

1 Co ra

— >CO "^_ to

c(0Q.

CO

It lEQ

CO

• = ©5>

to oO) ^

c «

o oo cto o

T3c cra ,o

o-o

Si «8

to -QQ ta

I

Q

to

.9 o

Ifi

If2 ^

- I (0

ECO

CQG

T3 O) g£ E (I) «3 to ^ ^^ CO ( G O )

Is |.?Q a to E

IIo

il< o

Eo

CO

.E>

uct

l£5 i2 S.

CO «— (0

if11

^?J

Si S

roQ.Q.

O g

Cl> j . ^ . ^ "

!= O CO

c.o2E•sto

tj

€O

a3

£

man

3SZ

to

to

Iiif"n to

9 "CO to

s

tca\

c

mm

u

OO<D.Q)

1oc

Q

co

sfra

f/'

,c

CO

c<&EQ.

elo

£2,oto

c

3log

y

co0)1-

P 2

E-2Ci Oo c

,03

Uj o

E «

£ O E

0) O>< tJ

Q.'tO

c .9? 5

ro . ^ 0 )

8 D>CD (DQi Q)

1— "D

5

1 ?O •= tn

2 ^ to"ffl is "(0I- a Q

raEc

co

oatAo

3Q .

• o

CD

0)

ra>a.x:o.acoOl

,E g£ ci> m

a CO

tr o

to Q> £ Q. S n)



sector; (2) the long public sector backlog of ap-plications awaiting development; and (3) the needfor this flexible resource in the paperwork-drivenenvironment of government. As mentionedabove, office automation, #5 on the public sec-tor list, has had a downward trend on the privatelists. It was #6 in 1982, #12 in 1984, and #11 in1985. The private sector surveys tend to bedominated, however, by the manufacturing sec-tor. If there were a private sector survey with ade-quate sample sizes for reliable industry-levelrankings, we might see this issue also rated ashigh as in tha public sector by, say, financialorganizations. Data security, #6 on the publicsector list, is in line with the private sector ratings,which increased to #6 in 1985.

Comparison of Public SectorRespondent CategoriesWe requested additional data from ourrespondents to help explain differences in theperceived importance of IS management issues.The variables included respondents' level ofgovernment, agency size, organizational positionof respondents, agency mission, the degree ofcentralization or decentralization of informationservices, the mix of mainframe versus microcom-puter applications, and IS managementcategories.

Contingency tables and the chi-squared test ofindependence were used to analyze the data.Even though we did not obtain a fully randomsample, we believe this analytical method pro-vides a better guideline than simply reportingaverage scores or percentage comparisons. Be-cause cell sample sizes were often too small ac-cording to a common rule of thumb,^ responseswere aggregated into two groups: "not, slightly,or somewhat important" (scale points 1, 2, and3) and "important or extremely important" (4 and5). Differences between categories that are statis-tically significant at the 0.95 or higher level areincluded in Table 4.

Level of governmentWe expected there would be differences insurvey responses across the local, state, and

^ Cell expected value frequencies less than five under the nullhypothesis ot independence should be avoided.

federal levels of government. For example, as thelevel of government increases, primary workmoves from operation of service delivery systemsto support of oversight and regulatory systems.Local governments are primarily concerned withproviding direct services, e.g,, public safety, solidwaste collection and disposal, and primary/secondary education. These services requiremainframe, transaction-processing systems. Thefederal government, while providing some directservices like national security and national parks,largely has policy making and oversight respon-sibilities. These policy analysis and administra-tion needs require flexible IS technologies likedatabase management systems, end-user com-puting, and computer networks (Caudle andNewcomer, 1987). We reasoned that the statesshould be somewhere in-between the local andfederal governments in regard to transaction pro-cessing versus flexible and distributed process-ing needs.

We asked respondents, therefore, if their agen-cy's primary program mission involved the col-lection, maintenance, and processing of indi-vidual records. The federal and state governmentrespondents were nearly identical, with 38 and35 percent respectively indicating a mission in-volving individual records. This statistic isdramatically higher for county governments: 84percent of county respondents had a mission in-volving transactional records.

As expected and as can be seen in Table 4, localgovernments place more importance on main-frame systems and efficient transaction process-ing while the federal government places moreimportance on distributed data processing,microcomputer systems, and flexible IS. Spe-cifically, both county and state governmentsplace significantly more importance on main-frame computing, while the federal governmentplaces more importance on distributed data pro-cessing and satisfying microcomputer personnelneeds than the other two levels of government.The federal government also places much moreimportance on database management system im-pacts, showing a need for flexible IS developmentand use.

Another factor leading to differences in surveyresponses across the local, state, and federallevels of governments is difficulty in makingchanges (i.e., "red tape"). The size of agencies



Table 4. Statistically Significant Differences in "important" or "Extremeiy important" Responses:Level of Government, Agency Size, and Management Levei (0.05 or iower type I error)

Level of Government

Educating elected officialsResearch and developmentProcurement and contractingOversight requirementsDatabase management system impactCivil service constraintsMicrocomputer personnel needsSubstituting technology for personnelDistributed data processingData securityMainframe computing

Size of Respondent Agency(Federai and State Oniy)

Procurement and contractingTechnology transfer mechanismsData discontinuity between administrations

Management Levei

Comprehensive planning integrationResearch and developmentExternal data sourcesPublic/private information interestsData discontinuity between administrationsArtificial intelligence/expert systems

Federai

5945836982796475796939

Smaii<200 FTE

458060

Middie

976573666568

State

64505449666856706784sr

County

7433404465454155657059

Large>200 FTE

745843

Top

843947454635

increases with level; the bigger the agency, themore rules and layers of management exist, andthe more difficult it is to make large changes(Downs, 1967). Also, the higher the level ofgovernment, the more diverse and widely scat-tered are constituents. These factors increasecontrols and accountability requirements, makeit difficult to hire and retain technical and profes-sional staff, slow down procurement processes,and increase transaction costs. Indeed, as seenin Table 4, the percentages of importance ratingsfor the red tape issues (procurement and con-tracting, oversight requirements, and civil serviceconstraints) increase with increasing level ofgovernment and are remarkably higher for thefederal government than the state and localgovernments.

Lastly, in the level of government section of Table4, note that educating elected officials becomes

increasingly important as the level of governmentdecreases. At lower levels, IS managers exper-ience increased contact and dependence on poli-ticians. Constituencies are smaller and moreintensely focused as th© level of governmentdecreases, placing more pressure on politicians.In turn, local-level politicians exert more influenceon the management of government than theirstate or federal-level counterparts. This attentioncan divert funds away from IS or, if politicans canbe "brought on board," can increase IS support.

Agency sizeThe size-of-agency variable, with categories forless than the greater than 200 full-time equivalent(FTE) employees, again provides some evidencethat red tape increases with size of agency.Seventy-four percent of the managers in large



state and federal agencies^ rate reducing pro-curement and contracting red tape as important,as compared to only 45% in small agencies. Thenext issue, technology transfer mechanisms, ismuch more important for small agencies thanlarge ones. Small agencies likely have limited oreven non-existent software development bud-gets, and technology transfer may be one of thefew ways of obtaining specialized software. Onthe last issue in this group, data discontinuity be-tween administrations, smaller agencies ratedthis more important than large agencies, perhapsbecause small agency respondents are more re-sponsive to transition issues and are closer to ser-vice delivery than their counterparts in largerorganizations.

Management levelTop managers in government organizations havea politicized, larger external role then theircounterparts in private firms. Top government of-ficials must seek appropriations through politicalmeans, are appointed and serve at the pleasureof elected officials or other political appointees,and must deal with political influences and thewants of multiple interest groups (Bozeman andBretschneider, 1986; Lachman, 1985; Marchard,1985; Mintzberg, 1973; Nigro and Nigro, 1977;Rainey, 1983). This means that government of-ficials are much more concerned with policyagenda-setting processes rather than manage-ment functions. Consequently, they do not focustheir attention on information resource manage-ment as much as their private sector counter-parts. We therefore speculated that in the publicsector, middle-level (career) managers fill this gapby placing higher importance on information tech-nology development and pursuit of emerging in-formation technologies than upper-level (electedor political appointee) managers.

To investigate this possibility, we askedrespondents how many management levels ex-isted between the head of their immediate officeand the agency head. The "top" managementlevel category in Table 4 is for a respondentwhose department head is within two levels of

We excluded county data from this analysis because agen-cy size responses are biased high. Relatively few countieswith population less than 250,000 have separate data pro-cessing departments, so respondents tended to provide totalcounty government employment and budgets rather than justdepartment values.

the agency head. "Middle" is three or morelevels below the agency head. As seen in Table4, the middle-level managers place significantlyhigher importance than top managers on fiveissues dealing with IS R&D and new uses of data:research and development, external data sourc-es, public private information interests, datadiscontinuity between administrations, and ar-tificial intelligence/expert systems. This result isas expected.

Organizational arrangements forcomputer servicesWe asked respondents to classify their office interms of organizational arrangements for supply-ing computer services. Table 5's top half showsthat almost three-quarters of the total sample'srespondents were from centralized data process-ing shops. The federal government respondentshave more dispersion among the processingcategories shown; state and county governmentrespondents reported a preponderance of cen-tralized data processing shops. Respondentswere also asked what percentage of their agen-cy's total information system applications aresupported by in-house mainframe staff, in-housemicrocomputer users, outside consulting ser-vices, or by another agency's mainframe and/ormicrocomputer staff. Median responses in eachcategory, shown in the bottom half of Table 5,indicate that 80 percent of the respondents' totalinformation system applications were supportedby in-house mainframe staff; county respondentsare high with 86 percent followed closely by thestates' 80 percent. Only 40 percent of federaltotal information system applications are sup-ported by in-house mainframe staff. Federalgovernment agencies have the highest level ofin-house microcomputer staff, 20 percent, whichdecreases to 10 percent for state governments,and 5 percent for county governments.

We expected that respondents would see issuesaffecting their computing support as more impor-tant. For comparison, we used two variables. Onewas type of processing—either placement in acentralized data processing shop or in a micro-computer shop (defined as 15 percent or moreof applications written by in-house microcom-puter users). The other was type of computerapplications, measured by predominance of in-house mainframe applications (defined as at least



Table 5. Type of Information Processing and Computer Applications (H)

Processing and Applications

Type of Processing (%)Centralized DP shopSatellite of centralized DP shopDecentralized departmental unitUse DP services of othersOther

Total

Median Type of Computer Applications {%)In-house mainframe staffIn-house microcomputer usersOutside consulting servicesOther agency's staff

LevelFederal

530

1628

299%

4020100

of GovernmentState

723

114

11

101%

801000

County

8812a0

99%

86500

TotalSample

731

12125

100%

801000

75 percent of an agency's total IS applicationsare supported by in-house mainframe staff).

As seen in Table 6, managers in "microcomputershops" ("Yes" column) are not interested inmainframe computing, controls that may beplaced on their planning or purchasing ("long-term planning mechanisms"), or advanced IStechnologies like decision support systems. Themainframe-oriented agencies ("Yes" column)want to develop and protect their kind of services(e.g,, "mainframe computing").

Information systems managementcategoryLastly, respondents were asked to pick one offour descriptions that best characterized theoverall state of IS management in their agency.The categories of information system manage-ment were a variation of Nolan's stages (Gibsonand Nolan, 1974; Nolan, 1979; Rainey, 1983;Rainey, et al., 1976) with the addition of informa-tion resource management descriptors (New-comer and Caudle, 1986). The choices were asfollows.

Initiation: Focus on early use of computers bya small number of users to meet basic applica-tions (such as accounting and record keeping);decentralized control and unstructured planning;physical paperwork and records managementare important.

Expansion: Focus on experimentation with andadoption of computers by many users, prolifera-tion of applications; information management isin transition from basic applications to manage-ment of more extensive applications and auto-mated technologies,

Formalization: Focus on increasing efficiency;charging users for IS services under developmentor being implemented; professionalization of in-formation system staff and procedures underway;schedule of prioritized information system pro-jects under development; management of infor-mation technoiogies and technical attributes areimportant.

Integration: Focus on an established set of ap-plications covering priority operational andmanagement information needs; database ad-ministration established; planning well-es-tablished and linked to the budget process;information resource management practicesestablished.

We recognize that this measure has limitationsin that the respondent has to integrate judgmen-tally over all IS components in his/her agency toproduce an overall assessment. Thus, whiie weincluded a four-category measure in the survey,we collapsed it to two categories for analysis: in-tegration stage and less than integration stage(labeled "other" in Table 7). In the total sample,47 percent of the responses were in the integra-



Table 6. Statistically Significant Differences in "Important" or "Extremeiy Important" Responses:Organlzationai Arrangements for Computer Services (0.05 or iower type i error)

Organizational Arrangement

Processing:Microcomputer Shop?Comprehensive planningProcurement and contractingLong-term planning mechanismsMainframe personnei needsMainframe computingData discontinuity between administrationsDecision support systemsTechnology security

Appiications:Mainframe Orientation?Procurement and contractingMainframe software managementOversight requirementsMainframe computing

Yes

8166654444404035

47744564

No

8950795558535247

67616238

Tabie 7. Statisticaiiy Significant Differences in "important" or "Extremeiy Important" Responses:IS Management Category (0.05 or lower type i error)

iS Management Category

Long-term planning mechanismsMainframe software managementDatabase management system impactData integrityExternal data sourcesDatabase administrationMainframe computingOffice automationTechnology securityTelecommunications technology

integration

79767673567961824876

Other

69626660455845723762

tion stage, 28 percent were in the formalizationstage, 21 percent were in the expansion stage,and 5 percent were in the initialization stage.

The 10 issues in this category with significant dif-ferences in Table 7 simply show that already ad-vanced agencies (in integration stage) want evenmore advanced technologies. The demand for IStechnology "snowballs" over time.

ConclusionTwelve of the 15 top-ranked IS managementissues in the public sector survey of this articlealso appeared on the top 15 issue lists of theprivate sector surveys in the literature. Thisreplication of results across sectors, with substan-tial variation in data collection and analysismethods, is gratifying. It suggests that past ef-



forts have developed a valid set of IS manage-ment issues. It also suggests that there are manysimilarities between the public and private sec-tors (a result consistent with Lachman, 1985).There are also notable differences.

First, two of the top 15 public sector issues-linking IS planning and budgeting, and tech-nology transfer—are unique and paramount inthe public sector. The budgeting process is cen-tral to the operation of governments, a processthat substitutes for the market allocativemechanisms of the private sector. Linking ISbudget requests to long-range IS planning is amechanism for overcoming the short-range em-phasis in government on political pressures.Seventy-three percent of the respondents to oursurvey ranked this issue as important or extreme-ly important. Second, technology transfer is anissue derived from the "public" nature of goodsproduced by governments. It is impossible andundesirable to exclude individuals from consum-ing such goods, so their costs need to be shared.Sharing IS development costs through govern-ment consortia and freely transferring tech-nologies from one agency to another contrastsharply with proprietary, private-sector ISdeveloped for competitive advantage.

Second, three issues that have peaked in impor-tance in the private sector and are now on thedecline rank in the five top public sector issues:integration of technologies, end-user computing,and office automation. The literature on pub-lic/private sector differences and the corres-ponding empirical findings found in this articlesuggest this lag to be due to many factors, suchas the complicated nature of public goods, short-run perspective of politicians, and limitationscaused by the government red tape and account-ability requirements. An alternative explanationis that the private-sector surveys largely repre-sent manufacturing firms and are not represen-tative of service organizations, which, likegovernment agencies, are still highly interestedin paperwork-oriented IS.

Finally are some findings on theoretically derivedissues. One is that top-level (political appointee)public managers are less inclined to develop newinformation technologies than middle-level(career) public managers. In contrast, MIS suc-cess in the private sector is clcsely tied to the sup-port of upper-level executives (Bozeman and

Bretschneider, 1986). Apparently, strategies fordeveloping IS technologies in the public sectorneed a middle-up emphasis. Another conclusionis that red tape, increasing with size of organiza-tion and level of government, has major impactson information technologies employed: the morered tape, the more flexible the informationtechnology employed.

In summary, we believe that IS managementissues research can provide overall guidance tomanagers, researchers, and educators. Thevalue of such research can be greatly increased,however, by proceeding to theory-based researchdesigns. Prior research has tended to report sum-mary survey responses at the total sample leveland apparently without much theoretical guid-ance. This article has provided main effects-levelresults, driven by IS and organizational theories.Although the results are still exploratory, theydemonstrate the feasibility and attractiveness oftheory-based IS management issues research.We encourage additional such research in thepublic and private sectors. In the private sector,we would like to see research including marketsegmentation (e.g., one-digit standard industrialclassification code level), organization size andstructure variables, type of IS (transactions pro-cessing, MIS, DSS, etc.), and other factors. Astudy contrasting public and private servicedelivery systems would be extremely interesting,especially one that ties into the literature on"privatizing" the public sector.

ReferencesBall, L. and Harris, R. "SMIS Membership

Analysis," MIS Quarterly (6:1), March 1982,pp. 19-38.

Bozeman, B. and Bretschneider, S. "PublicManagement Information Systems," PublicAdministration Review (46), Special Issue,November 1986, pp. 475-467.

Brancheau, J.C. and Wetherbe, J.C. "Key Issuesin Information Systems Management," MISQuarterly (11:1), March 1987, pp. 23-45.

Camillus, J.C. and Lederer, L. "CorporateStrategy and the Design of Computerized In-formation Systems," Sloan ManagementReview (26.3), Spring 1985, pp. 35-42.

Caudle, S.L. Federal Information ResourcesManagement: Bridging Vision and Action, Na-tional Academy of Public Administration,Washington, DC. 1966.



Caudle, S.L. and Newcomer, K. "Command andControl: Public Program Oversight in the In-formation Age," Information ManagementReview {3:^), Summer 1987, pp. 37-50.

Dickson, G.W., Leitheiser, R.L., Wetherbe, J.C,and Nechis, M. "Key Information SystemsIssues for the 198O's, "MIS Quarterly (8:3),September 1984, pp. 135-159.

Downs, A. Inside Bureaucracy, Little, Brown andCo., Boston, MA, 1967.

Downs, G.W. and Larkey, P.D. The Search forGovernment Efficiency. Temple UniversityPress, Philadelphia, PA, 1989.

Gibson, CF. and Nolan, R.L. "Managing theFour Stages of EDP Growth," HarvardBusiness Review (52:1), January-February1974, pp. 76-88.

Gorr, W.L. "Special Event Data in SharedDatabases," MIS Quarterly {^0:3), September1986, pp. 239-255.

Hartog, C. and Herbert, M. "1985 Opinion Surveyof MIS Managers: Key Issues," MIS Quarter-ly {\QA), December 1986, pp. 351-361.

Keen, P.G.W. Competing in Time: UsingTelecommunications for Competitive Advan-tage, Ballinger Publishing, Cambridge, MA,1988.

King. W., Hufnagel, E., and Grover, V. "UsingInformation Technology for Competitive Ad-vantage," in Information Management: TheStrategic Advantage, M. Earl (ed.). ClarendonPress, New York, NY, 1988, pp. 75-86.

Lachman, R. "Public and Private Sector Dif-ferences: CEOs' Perceptions of Their Role En-vironments," Academy of ManagementJournai {28:3), September 1985, pp. 671-679.

Lucas, H.C., Jr. "Utilizing Information Tech-nology: Guidelines for Managers," SloanManagement Review (28:1), Fall 1986, pp.39-47.

Marchand, D.A. "Information Management:Strategies and Tools in Transition," Informa-tion Management Review (1:1), Summer 1985,pp. 27-34.

Mintzberg, H. The Nature of Managerial Wort<,Harper and Row, New York, NY, 1973.

Moynihan, T. "What Chief Executives and SeniorManagers Want from Their IT Departments,"MtS Quarterly i•\4:^), March 1990, pp. 15-25.

Newcomer, K. and Caudle, S.L. "Federal Infor-mation Resource Management: Embracingthe Concept," Information ManagementReview (2:1), Summer 1986, pp. 29-36.

Nigro, F. and Nigro, L. Modern Pubiic Administra-tion, Harper and Row, New York, NY. 1977.

Nolan, R.L. "Managing the Crises in Data Pro-cessing," Harvard Business Review (57:2),March-April 1979. pp. 115-126.

Rainey, H.G. "Public Agencies and PrivateFirms: Incentive Structures, Goals, and Indi-vidual Roles," Administration and Society(15:2), August 1983, pp. 207-242.

Rainey, H. Backoff, R., and Levine, C. "Compar-ing Public and Private Organizations," PublicAdministration Review (36:2), March/Aprii1976, pp. 233-244.

Savas. E.S. Privatizing the Pubiic Sector,Chatham House Publishers, Inc., Chatham,NJ, 1982.

Self, P. Administrative Theory and Politics, Allenand Unwin, London, 1977.

Simmons, A. "New York's Info Forum," Govern-ment Techno/ogy (1:4), July/August 1988, pp.1-6.

About the AuthorsSharon L. Caudle is dean of research and plan-ning at Riverside Community College in River-side, California. She was formerly with the Schoolof Information Studies at Syracuse University.She is the author of monographs and articles onpublic sector information resource managementappearing in journals such as Public Administra-tion Review, the International Journal of PublicAdministration, and Information ManagementReview.

Wilpen L. Gorr is professor of public policy andmanagement information systems at the Schoolof Urban and Public Affairs, Carnegie MellonUniversity. He has published articles in Manage-ment Science, MIS Quarterly, Decision Sciences,Public Administration Review, The InternationalJournal of Forecasting, and other journals. Hiscurrent research interests include the design, im-plementation, and evaluation of decision supportsystems for drug enforcement and local govern-ment budgeting; development of methods to es-timate spatial diffusion processes; and modelingthe spatial diffusion of the HIV/AIDS epidemic.

Kathryn E. Newcomer is professor and chair ofthe Department of Public Administration atGeorge Washington University, where she teach-



es courses on research design, statistics, andprogram evaluation in the M.P.A. curriculum. Shehas also taught at the University of Nebraska, theUniversity of Denver, the University of Iowa, andGrinnell College. Dr, Newcomer has served asthe principal investigator for a number of evalua-tion projects, including studies related toteleconferencing administrative hearings, publicpolicy implementation, congressional elections,federal regulation, and other topics. She hasdesigned evaluation strategies and then partici-

pated in the subsequent evaluations of programswithin several national non-profit organizationsbased in Washington, D.C. She has also workedas a consultant to the Departments of Army, In-terior, State, and Agriculture. Her most recentbook is Improving Government Performance.which she edited with Joe Wholey. She has pub-lished articles in several books and journals, in-cluding Law and Policy Affairs Quarterly,Evaluation Review. Information Management Re-view, and The Public Administration Review.

Appendix

Survey Questions: Importance of Information Systems Issues

A. Strategic management concernsIssue 1 : Educating Elected Officials. Top political leaders need to learn the role of informationresources, such as information systems, in government.

Issue 2; Political Influence. New planning and budgeting procedures should be developed to reducethe impacts of short-run political priorities and the lack of political continuity that can undercut com-prehensive, long-term information systems planning.

Issue 3: Long-Term Planning Mechanisms. To provide for long-term decision making, agenciesshould submit and approve information resource plans periodically before the submission of cor-responding, plan-linked budget requests.

Issue 4: Comprehensive Planning Integration. There should be a process that integrates infor-mation resource planning with the agency's overall planning process.

Issue 5: Information Resource Management. Agencies should adopt information resource manage-ment concepts (e.g,, integration of information technology management, records management, andother information management areas) and assign a central institutional responsibility for informa-tion collection, processing, dissemination, policy, and operations.

Issue 6: Research and Development. Agencies should devote more resources to monitor andevaluate technological developments for agency needs.

S. Operational and tactical management concernsIssue 7: Identification of Information Requirements. Agencies should develop systematic pro-cesses to identify and prioritize information requirements.

Issue 8: Measuring IS Effectiveness. Agencies should develope better measures and estimationmethods to show the true costs of information and evaluate the net benefits of proposed informa-tion system projects.

Issue 9: Procurement and Contracting. Agencies should modify procurement and contracting pro-cedures to expedite the delivery of information services, hardware, and/or technical assistance.

issue 10: Oversight Requirements. Agencies should balance accountability demands and manage-ment practicalities so external controls do not needlessly hamper technological innovation.



C. Applications development and maintenanceIssue 11: Technology Transfer. Agencies should find ways to increase sharing of applicationsand technical assistance with other agencies having similar functions or providing similar services.

Issue 12: Technology Transfer Mechanisms. On a statewide or national basis, similar agenciesshould avoid "reinventing the wheel" in applications software development and technological in-novation by forming or using cooperative arrangements.

Issue 13: Mainframe Software Development. Mainframe applications software should be developedmore quickly and with consistently higher quality.

Issue 14: Database Management System Impact. IS must increasingly use database manage-ment systems to enhance application development and maintenance.

Issue 15: Application Generators. IS must increasingly use application generators and fourthgeneration languages in place of third generation languages to increase productivity.

Issue 16: Software Maintenance. IS managers need to find ways of improving productivity in soft-ware maintenance.

D. Personnel issuesIssue 17: Civil Service Constraints. Civil service rules that hamper or prevent the recruitment ofskilled information management and technical personnel should be changed.

Issue 18: Meeting Mainframe Personnel Needs. Continued personnel shortages will increase theneed for resources to recruit, retain, and retrain mainframe personnel.

Issue 19: Meeting Microcomputer Applications Personnel Needs. Continued personnel shortageswill increase the need for resources to recruit, retain, and retrain personnel in microcomputer soft-ware and applications positions.

Issue 20: Substitution of Technology for Trained Personnel. Continued personnel shortageswill require IS to depend more heavily on application generators, database management systems,end-user computing, and so on, to meet information needs.

E. Data managementIssue 21: Distributed Data Processing. Policies are needed that support the advantages ofdistributed data processing while maintaining the integrity of central information systems.

Issue 22: Data Security. Agencies should balance data security and data availability through ap-propriate protocol and access controls.

Issue 23: Computer Viruses. Agencies should develop and implement active programs for diagnos-ing and protecting information systems from computer viruses.

Issue 24: Data Integrity. IS must devote additional resources to the accuracy, timeliness, and ade-quacy of data provided to users.

Issue 25: External Data Sources. IS managers should screen external data sources, heighten userawareness of external data valuable to the organization, including interagency and intergovernmentalsources, and provide access to such data.

Issue 26: Database Administration. IS should increasingly use database administration to pro-vide quality assurance of data holdings across many end users and deal with information resourceownership concepts across the organization.

issue 27: Public and Private Information Interests. Agencies should balance agency users' in-formation needs with the requirements of legislation, such as the Freedom of Information Act andthe Privacy Act.



Issue 28: Data Discontinuity Between Administrations. IS maiiagement should find ways to pro-vide for long-term continuity of data definitioiis and data aggregate classifications, or a historicalrecord of changes, to maintain the usefulness of historjcal data.

F. Information technologiesissue 29: integration of Technoiogies. IS management should ensure that current and future dataprocessing, telecommunications, and office automation technologies are integrated to prevent in-compatibility.

issue 30: End-User Computing. Agencies should provide and increase support such as informa-tion centers and standardized hardware and software for end-user computing.

issue 31: Mainframe Computing. Agencies should provide more resources to improve mainframecomputer services.

issue 32: Office Automation. IS managers should take a leadership role in planning, implementing,and managing office automation.

Issue 33: Decision Support Systems. IS should develop and implement decision support systemsthat facilitate their use in an agency's political environment.

issue 34: Artificiai inteiiigence and Expert Systems. IS managers should study artificial intelligenceand expert systems now to plan for their use over the next several years.

issue 35: Microcomputer Software Management. Initial setection of microcomputer software andsubsequent upgrades should be better managed.

issue 36: Technology Security. IS management should implement better physical security overinformation technologies, such as computers and peripheral devices.

Issue 37: Teiecommunications Technoiogy. IS management should ensure that long-term voiceand data telecommunications decisions are made despite continuing technological changes, telecom-munications deregulation, and the scarcity of experienced personnel.


Key Information While most of the top public sector issues ...€¦ · Key Public Sector IS Issues...

Documents

Transcript of Key Information While most of the top public sector issues ...€¦ · Key Public Sector IS Issues...