AUTOMATED INFORMATION-SYSTEMS AND URBAN DECISION-MAKING

Abstract This paper assesses the value of automated information systems for urban decision making and investigates alternative strategies for enhancing the value of this information. It is based on intensive survey and case study data in 40 U.S. cities with populations greater than 50.000. The findings indicate that although automated information systems are attributed significant usefulness by some types of decision makers for certain arrays of decisions, they are not generally useful to most urban decision makers. However, the findings suggest that they might be made more useful through management strategies which stress sensitive integration of these users with the technology.

Abstract-This paper assesses the value of automated information systems for urban decision making and investigates alternative strategies for enhancing the value of this information. It is based on intensive survey and case study data in 40 U.S. cities with populations greater than 50.000. The findings indicate that although automated information systems are attributed significant usefulness by some types of decision makers for certain arrays of decisions, the) are not generally useful to most urban decision makers. However. the findings suggest that they might be made more useful through management strategies which stress sensitive integration of these users with the technology.

awareness
of the complexity and interdependence of urban problems has created a need for better information to aid decision making [4,21]. At the same time, research suggests that decision makers seldom make particularly effective use of available information [6,7,25.27]. As a consequence, much concern has been focused on achieving more effective utilization of available information resources. This concern has contributed to the development and expansion of automated information systems (AIS), which aim to make better information more accessible to decision makers [ 18.201. The benefit of investments in AIS is a matter of considerable debate. First, research suggests that information is often used to serve dominant organizational and political interests rather than to formulate more rational decisions [2,5,251. And it seems unlikely that the new automated information systems will alter these barriers to rational decision making [14]. In fact, automation might extend an organization's capability to use information in ways that reinforce existing structures of influence [19].
Second, even if information sometimes does enhance the rationality of decision making processes. there is considerable disagreement regarding those factors which have promoted the use of such information for decision making. The most prominent alternate explanations suggest that the effective utilization of information is a function of informational content, or the organizational role of decision makers, or the nature of AIS technology, or the organizational environment, or the style of decision makers, or the level of socio-technical integration. Some view the content of inforrmtion as a primary determinant of its utilization. Decision makers will use information to the degree it is 'powerful'-important, comprehensive, understandable, sophisticated and from a credible source [6]. From this perspective. automated information will be used differentially, depending upon its direct relevance to particular decision makers. The use of automated information might also be a function of an individual's ovya~izationul role. On the one hand, computers might alter the flow of information so that higher level officials will get more extensive, less filtered information, which can be used to enhance their control [13,15,281. On the other hand. computer-based information is likely to be most usefu'l to those with the greatest expertise in the use of information and of computer technology.
As a consequence, one might predict that technical t Authors are listed randomly to denote equal contribution. staffs will gain more information benefits than will top managers. who in tur-n will gain more information benefits than \viIi top executives [9]3 Many view the utilization of information as 21 function of tc,c,hrlo/or/ic,tr/ rl~,r.c,/oprlfc~r~r. Automated information systems are expected to increase the availability of better information (that is. higher qualit>. greater quantity. more managcablc) to top decision makers [I. 26. 2X]. Some research suggests that urban decision makers tind considerable value even in rather unsophisticated computer applications [IO]. It is assumed the development of more sophisticated systems should further increase the \,aluc of automated information for decision makers [IX. 241,lLll cw~~iu~r~r~~v~~ as an important determinant of the inlpact and value of an organization's information system to the organization itself [20]. Generally.
the value of an automated information system is expected to bc higher in organizations with a managerial climate that is favorable to the use of modern professional management practices and to rational decision making processes; the value of AIS is expected to be lower in ;I climate fa\orablc to more traditional and more personal decision making processes.
Finally. two sorts of explanations of the usefulness of AIS are based on indi\,idual differences among decision makers. One explanation is dependent on the d~,~.i.sio/l UIU~\PV'.S s~J~/c. According to this view. automated information will be of most value to ;I new class of decision makers who are relatively professional, cosmopolite. educated. young and rational in their approach to decision making. In contrast. the old class of urban decision makers is relatively political. local in orientation. less educated. older and intuitive in their approach to decision making 193. The second explanation posits that the value of automated information will \'ary across individuals. depending on the degree of soc.io-tc,c.llrlic.crl i~tcqrcitim that is. the degree to which users are involved. trained and familiar with the information systems [3. 12. 251. It is important to understand which of these altcrnativc explanations best account for the value of automated information to urbl;n decision makers. Knouledge of the political. administrative. and individual factors which constrain or enhance the value of automated information to decision makers might aid in the development and implcmentation of more effective systems. While the literature offers ;I t-ich array of theoretical expectations.
few have been empirical11 :ind systematicall) examined outside of case studies. This paper examines alternative explanations for the value of automated information for decision making within the context of American local government. Each hypothesis is cxamincd in light of survey data on the usefulness of automated information for urban decision makers. The findings are then discussed in terms of both their theoretical and policy implications.

AND DATA
Our strategy for an empirical assessment of the factors which contribute to the value of automated information systems is to analyze the perceptions of a variety of key decision makers in the setting of urban local government. First. we specify the level of usefulness attributed to different kinds of automated information in order to evaluate our first hypothesis, which suggests that the value of information is a function of informational content. Second. wc examine whether differences in the perceived usefulness of information is contingent upon the organizational role of the decision maker. Given the organizational role hypothesis. there should be important between-role differences in the value attributed to automated information.
Third. we assess the remaining hypotheses by examining the relationships between the perceived usefulness of automated information systems and ;I variety of independent variables which represent the degree of technological development. the organizational environment. the style attributes of the decision makers. and the degree of soaio-technical integration.

Thr smnple
This study is based on data collected in 1976 in 42 U.S. cities with populations greater than 50,000. The cities were selected by a stratified sampling procedure such that all cities in the sample were automated and the cities varied in the sophistication, extensiveness. and organizational arrangements of computing.? Thus. the average sample city is somewhat more automated than the typical U.S. city over 50.000 and the computing environments of these cities have somewhat more variation than would be the case for a random sample of automated cities.

The data
Within each city. data was collected using self-administered questionnaires, field coded questionnaires and case studies of the use of computing by local government officials. This paper relies most heavily on the self-administered questionnaires, which are treated as empirical data on the way in which computing is used by local government officials. The pretested. self-administered questionnaire was completed by about 75 respondents within each city. An 82% response rate was obtained by using extensive follow-ups and personal visits to pick up questionnaires.
In addition to the self-adminis'tered questionnaires, the field work involved each of six investigators, including the authors, in case studies in at least eight cities. with field work averaging three person-weeks per city. Each site visit provided rich case study material as well as systematic judgmental ratings. The systematic ratings were derived from a series of structured questionnaires which were completed by the investigators during each site visit and were based upon numerous interviews as well as archival research.

Thr respondents
Since decision makers for urban governments are many and varied. we chose to focus on seven important roles within local government itself: mayors. councils, top managers.
central management staff, department heads, administrators. and analysts. Mayors are the top elected officials in local governments and are often the primary executive policy makers. Elected council members have a legislative policy role but are usually less likely than the mayor to be involved in executive policy making. Top managers (here. the city manager or the chief administrative officer within a mayor-council city) are the focal point for administrative as well as many policy decisions in local government.
Central management staff are those professional staff of the mayor and manager who play important analytical and advisory roles. Department heads, such as the police chief and public works director, have key roles in making and administering policy within their organizational domain. Administrators. in our classification, are those personnel who serve department heads in such tasks as budget monitoring, staff supervision, and project management.
And analysts serve various decision makers by maintaining and analyzing data generated from city operational files, field surveys. and other sources. The specific kinds of officials dlassified in these roles are listed in the Appendix.

Limitations
There are several important limitations to our analysis. First, this study is based on a sample of U.S. cities with at least a moderate level of automation.
Consequently, we cannot generalize to all current U.S. cities; but, given our sample, we can generalize to some future state which most cities will experience. Second. the effective utilization of information is a controversial issue and its measurement is complex. Our reliance on individual perceptions of the value of automated information is necessary given our sample size and our focus on individual differences. Hopefully. other studies, using more objective measures, will complement our efforts.
t A detailed description of this sampling procedure is contained in [17].
The value of information to decision makers should increase kvith its importance. comprehensivcncss.
sophistic~~tion. and credibility. Thus, the value of computer-based information systems should vary, given that urban data systems are not able to address a broad range of questions equally. For a variety of political and administrative reasons, the kinds of computer applications adopted by American local governments tend to he heavily oriented to finance and budgeting [Xl. Financial datu generally is: the most comprehcnsivc within the government. because it is clrganiLation-wide in scope; the most crcdiblo because it is based on continually audited accounting systems: and the most important because it is undergirds many nonfinancial as weil as financial decisions. Fewer applications contain information about the community environment.
Environmental data is less potent because: the data usually is fragmented in planning, building and public works files: its meaning is frequently clouded by the sophisticated analysis used to distill and interpret it; and its importance tends to be limited to physical dcvclopmcnt decisions. Even fewer applications contain information. such as performance data or monitoring data. which can be used directly for management control [X. Ih]. Performance data tend to be fragmentary and the least standard of all. It is easily manipulated by those being monitored and easily misinterprctcd by those seeking to control; therefore, it lacks credibility for management decisions. Given these differences in informational content. we expect that urban decisionmakers might find that computer-based information systems are the most useful for bL]dgeting, somewhat less useful for policy and decision making. and least useful fol managerial control. In general, these expectations are consistent with the ratings of municipal oflicials. Computer-based data are not rated as 'very useful' overall. But they are rated as relatively more useful for budgeting than for either community decision making or for rn~~~l~gen~e~lt control.
Across all officials, computer-based data is rated bctwccn *useful' and 'somewhat useful' for budgeting decisions ( Table I ). Automated data is most useful during the annual budget cycle. with one-fourth of all officials viewing computer-based data as very useful during this period. One-fifth or less of all ofikials rate computer-based data as very useful for day-to-day expenditure decisions, salary questions and negotiations. or cost accounting (Table I). Computer-based data is rated 'somewhat useful' for decision making about the community el~vironment (Table 2). Only about one-tenth of all officiuls rate automated data as very useful for identifying city problems or for providing indicators of community conditions.
Less than one in twenty officials believe that automated data is wry useful for determining solutions to city problems (Table 2). Automated data is reported to be least useful for managcmcnt control. On the average, the oficials rate automated data as less than 'somewhat useful' for manpower allocation, setting realistic goals for subordinates. monitoring subordinates, or evaluating subordinates' performance (Table 3). In general. then. the usefulness of automated data does vary as a function of the match between the information content of AIS and the type of decision to which the automated tiles arc applied.
Since different organizational roles are likely to create different information needs and demands. the value of automated files might vary with the role of a decision maker. In fact. the value of computer-based information across roles does tend to reflect the different i~if~~rrn~ition~il needs and demands of the occupants of varied roles. Managers tend to find computer-based information more useful for budgetary decisions than do other officials (Table 1). consistent with their responsibility for formulating and executing the city budget. Elected officials and analysts tend to attribute more value than other roles to c~~Inpilter-b~lscd ii~forin~tioll for cornrn~li~ity policy making (Table 2). This pcr- (2) (1) (3) (4, In no cases In a few cases In many cases In nearly all cases Average score ception reflects the broad policy making perspective of the mayor and council and the importance of community-based data (e.g. data on land use and population) to analysts in meeting reporting requirements. preparing grant proposals. and developing planning documents.
And administrators find computei--based data more useful for managemcnt control than do other officials (Table 3). This assessment is consistent Lcith the particular need of department heads and higher level oficials to utilize information concerning interdepartmental. inter-governmental and communit! matters as opposed to intradepartmcntal management--the job of administrators. However. the similarities among officials in the perceived value of computer-based information are more striking than the differences. This is reflected in the data in Tables 1 3. but is more clearly summarized in Table 4. which lists an average index score for each type of role's rating of the usefulness of automated information systems for budgetary decisions, community policy decisions. and management control.? Interestingly, managers and elected officials tend to attribute more utility to computer-based information than do other professional personnel such as department heads and top staff. But. the differences among these average scores arc small and are not statistically significant.
In short. the organizational role of decision makers is not a critical determinant of the value attributed to automated information.
There is often more variation within a role-type than bctwecn different roles. Yet role is not irrelevant. There is some support for our expectations that AIS provide information of more value to higher level officials. And. more generally, the perceived value of computer-based information tends to reflect the varied information needs of different decision makers.
The technological development hypothesis suggests that computer-based information will be more useful in organizations with more highly developed AIS technology. The development of AIS technology in the 42 cities is indicated by measures of the number of automated data banks, the number of operational data systems. and the sophistication of computing at the local government's computer installation (Appendix). According to this hypothesis. officials in cities with more data banks, more operational systems and greater computing sophistication should rate computer-based information as more useful than do officials in less technologically developed sites. However. this hypothesis is not supported.
Officials in highly developed cities are neither more nor less likely to pcrccive computer-based information as more useful (Table 5A). Tables I 3 were

Organizational environment
The organizational environment hypothesis posits that computer-based information will be judged more useful by government actors in cities with organizational climates that are more favourable to professional management and rational decision making. The organizational climate of the 42 cities is indicated by measures of the city's size, of the use of the city-manager form of government, of the use of professional management practices, and of political partisanship (Appendix). The organizational environment hypothesis suggests that officials in the smaller, reformed, professional and nonpartisan cities will make greater use of computer-based data and, as a consequence. will attribute greater utility to such data. This hypothesis is weakly supported, since there is a tendency for computer-based data to be rated as somewhat more useful by government officials in council manager cities and in cities with a greater use of professional management practices (Table 5B). City size and political partisanship have virtually no systematic relationship with the usefulness attributed to AIS.
Decision maker's style The decision style hypothesis suggests that computer-based information will be judged more useful by those governmental actors who are more professional, cosmopolite, more educated, younger, less tenured, and characterized by a rational (vs. intuitive) decision style-in short, who are the 'new' urban decision makers. This hypothesis receives some support in the data, because several of these characteristics of individual decision makers are weakly associated with the usefulness they attribute to automated information.
Specifically, officials who are more professional, cosmopolite, educated, and less intuitive tend to rate computer-based information as more useful (Table 5C).

Socio-technical integration
The socio-technical integration hypothesis suggests that the value to decision makers of a 'high' technology like automated information systems is dependent on the degree to which those individuals have been involved with, trained about, and linked to the contact with data processing personnel, and USC' of computing experts for the analysis and int~rpr~t~~tio1~ of data (Appendix).
Generally. the expectations about s~~cio-tcchnic~~l integration are more strongly supported than an> alternative explanation of the value of computer-based information. Individuals attributing greater usefulness to AIS tend to have a higher level of computer utilization. more experience with computing. more EDP training. more interest in coml?Liting. mot-c involvement in design. and they make greater use of computer experts (Tohlc SD). The importance of socio-technical integration is further supported by multiple regression analysis. Such analysis indicates that only four variables from Table 5 have :I  statistically  independent direct association with the usefulness of automated information systems indcx.t Three of these four variables are aspects of socio-technical integration. and the fourth variable is the decision style of the decision maker. The path analysis in Figure 1 reveals that the four variables explain about 20"(, of the variance in the usefulness of AIS index. Morcovcr. the figure suggests two interesting 'routes' for enhancing the value of computer-based information to urban decision makers. First. the USC of an expert as an integrating mechanism appears to be a fruitful means to enhance the value of available information resources. This is indicated by a direct relationship between the use of experts and the perceived value of computing. The second route involves a more complex network of variables which represent the individual's relationship to the computing milieu. This route suggests that the active USC of computing by an individual is promoted by the individual's interest in computing and by the individ~1ai.s tendency to employ a rational rather than an intuitive decision style (Fig. I). If one's objective is to increase the value of computing to urban decision makers. one must involve them in the use of computing, stimulate their interest in the technology. and promote their utilization of systematic information in decision making. Each of these aims concerns an aspect of socio-technical integration. Therefore. it is important to identify those factors which might lead to increased computer utilization. interest. and rational decision making.

Interest in computing
also is predicted well by certain socio-technical integration strategies, particularly the involvement in the design of computer applications and also by training in EDP (Table 7). It is also evident that individuals with greater interest in computing tend to be more cosmopolite. more professional and more rational in their decision style (Table 7).
Finally, the decision makiny orientation of an individual, unlike computer utilization and interest. is not strongly associated with any other variables employed in this analysis. Thus. such orientations might be difficult to shape through management strategies other than recruitment practices.

SUMMARY AND DISCUSSION
This study supports the findings of other studies which note the quite low levels of utilization of systematic information in governmental decision making. Specifically, we have shown that most urban decision makers attribute only marginal usefulness to automated information.
However, this study has identified important variation among urban decision makers in the utilization and perceived value of automated information for decision making. Furthermore.
the findings have indicated that there are specific strategies which might enhance the value attributed to automated information by urban decision makers.
The usefulness of automated information is shaped by the degree an individual decision maker has been 'integrated' with the technology. In particular.
the value attributed to automated information by urban decision makers tends to increase where the decision maker utilizes such information more extensively, is interested in computing. is involved in the design of automated systems. and makes greater use of AIS experts. While socio-technical integration appears to be the most important explanation of the value attributed to automated data, other explanations are also supported. First, the content of automated information is related to its usefulness to different officials. A considerable number of the automated information systems in city governments center in financial data, which is of prime benefit to those concerned with the allocation and control of resources. Second, the value of automated information varies with the organizational role of an individual. While higher. level decision makers tend to attribute more value to automated information. other officials also find computer-based information useful when it corresponds to their specific areas of responsibility. Also, the decision style of an individual tends to be important in two respects. First, one's decision making orientation affects the perceived value of automated information. Second, a professional and cosmopolite decision maker is more likely to have an interest in computing, and this, in turn. promotes the individual's utilization and assessment of the value of automated information systems. There is little support for other alternative explanations for the value attributed to AIS. Technological development does not seem to enhance the value of automated information as much as some research suggests. This might be due in part to the fact that much current utilization of computing for urban decision making involves the pragmatic use of rather unsophisticated computer applications [lo. 111. The amount of automated information is important to the extent of its utilization ( Table 6). but the sophistication of AIS technology is not. Likewise. the nature of the decision makers' organizational environment in general is not an important determinant of the value they attribute to automated information. However, two characteristics of reform governments---the city manager form of government and use of professional management practices-do associate positively with the usefulness of AIS. This finding is consistent with earlier studies [ 10,11,18) which indicate that reformed governments have greater management use of computing and greater perceived benefits from that usage. These findings have implications for the management of automated information systems in local governments and. possibly, in other organizations. Specifically, the value of automated information for decision making can be enhanced by management policies which promote the integration of users with the technology. Socio-technical integration might be approached by involving users in the design of information systems. creating extensive opportunities for training and experience with the technology. providing information analysts to assist decision makers. and recruiting more professionally oriented. rational decision makers to key organizational positions.
Collectively. these policies might substantially increase the utilization and usefulness of information systems not only to individual decision makers. but to the organization as a whole.