Goal Ambiguity and Organizational Performance in U.S. Federal Agencies

In spite of numerous observations that government organizations have high levels of organizational goal ambiguity that exert major influences on their other characteristics, few researchers have measured goal ambiguity and tested these frequent assertions. In previous research, we developed measures of four dimensions of goal ambiguity: mission comprehension ambiguity, directive goal ambiguity, evaluative goal ambiguity, and priority goal ambiguity. Confirming hypotheses developed from the literature on public organizations, the latter three variables showed relations to such organizational characteristics as organizational age, financial publicness (proportion of funding from government allocations), and regulatory status. This article reports a second analytical step of examining the relations between the goal ambiguity dimensions and indicators of organizational performance based on responses to the 2000 National Partnership for Reinventing Government Survey of federal employees. The performance variables included managerial effectiveness, customer service orientation, productivity, and work quality. Regression analyses with numerous control variables found that directive, evaluative, andpriority goal ambiguity relatednegatively tomanagerial effectiveness. All four performance indicators showed significant negative relationships with evaluative goal ambiguity and directive goal ambiguity. The results provide further evidence of the viability of the newmeasuresof goal ambiguity, support theory-basedbut previously untestedhypotheses, and further indicate the feasibility and value of analyzing goal ambiguity of government organizations. The literature on public organizations contains numerous assertions about the impacts of goal ambiguity on important characteristics of those organizations. Again and again, authors say that vague, hard-to-measure goals influence structural dimensions, attitudes, behaviors, and organizational outcomes in public organizations andmake them different from business firms on these characteristics (for a review, see Rainey 1993). The great frequency of such observations, however, has not generated a comparable amount of empirical research aimed at proving or disproving them. The small number of studies that have tried to provide We wish to thank Jungwook Lee for his valuable editorial assistance with this article. We also wish to thank the anonymous reviewers for rigorous critiques and numerous colleagues who provided comments at the ‘‘Determinants of Performance in Public Organizations’’ seminar, Cardiff University, May 6–8, 2004. Address correspondence to Young Han Chun at [email protected]. doi:10.1093/jopart/mui030 Advance Access publication on February 25, 2005 a The Author 2005. Published by Oxford University Press on behalf of the Journal of Public Administration Research and Theory, Inc. All rights reserved. For permissions, please e-mail: [email protected]. JPART 15:529–557 at R uers U nirsity on A uust 2, 2010 http://jpaordjournals.org D ow nladed fom empirical evidence, moreover, have usually relied on managers’ responses to survey questions about whether their organizations have vague or clear goals (e.g., Rainey, Pandey, and Bozeman 1995). These studies have found that public and private managers do not differ in their responses to questions about whether their organizations’ goals are vague and hard to measure. These findings thus run counter to the typical observation that public organizations have less goal clarity than business firms. They raise the question of whether one would get different empirical results using evidence relying less on managers’ survey responses. The situation calls for more research with better measures of goal ambiguity. This study relates new measures of the goal ambiguity of U.S. federal government agencies to measures of the agencies’ performance. This study is the second report of a larger study reported previously (Chun and Rainey 2005). That first article reported the development of the new measures of the four dimensions of goal ambiguity described below. As described below, it reported evidence that supported hypotheses about the relations between those measures of goal ambiguity and ‘‘antecedent’’ variables that should influence the degree of goal ambiguity. These antecedents included the agency’s financial publicness (the proportion of its funding from government allocations as opposed to sales or user charges), regulatory functions versus nonregulatory functions, complexity of the policy problems the agency confronts, and others. Researchers in the social sciences have paid very little attention to clarifying and measuring the concept of goal ambiguity. These confirmations of hypotheses about relations between the goal ambiguity measures and these antecedent variables justified optimism about measuring organizational goal ambiguity in a meaningful way and about using those measures in research on the many assertions about its antecedents and consequences. The larger study analyzed the relations between goal ambiguity and a variety of consequence variables, including such variables as employee work satisfaction and perceptions about organizational structure. This article reports our analysis of the relations between goal ambiguity and arguably the most important of these consequence variables, organizational performance. Examining these relationships has important implications for the theory of public organizations, since assertions about goal ambiguity figure so importantly inmany scholars’ assertions about the other characteristics of public agencies, including their performance. It also has important implications for public policy and managerial practice, for the assumption that goal clarification will improve organizational performance underlies recent administrative reforms, including the Government Performance and Results Act (GPRA) in the United States and initiatives based on the New Public Management in other nations. This assumption involves a leap of faith, since there is so little empirical evidence of such a relationship and since one can argue that for some government agencies goal clarification might not be feasible and might indeed be dysfunctional. Further development of measures of organizational goal ambiguity and research on their relations to other variables can contribute to better analysis of such matters. Then comes the formidable challenge of measuring the performance of government agencies in a way that provides common, comparable measures for all the different agencies. We use data from the very large 2000 National Partnership for Reinventing Government Survey conducted by the Office of Personnel Management to construct the performance measures from the responses of the employees in the agency. As we describe later, the survey’s very large sample size leads to very sensitive statistical tests that find many statistically significant results but with small effect sizes (small R statistics). Nevertheless, the finding of statistically significant support for the hypotheses offers 530 Journal of Public Administration Research and Theory at R uers U nirsity on A uust 2, 2010 http://jpaordjournals.org D ow nladed fom further support for the usefulness of the measures of goal ambiguity and useful evidence about frequent assertions in the literature. DIMENSIONS OF GOAL AMBIGUITY IN ORGANIZATIONS For this study, organizational goal ambiguity refers to the extent to which an organizational goal or set of goals allows leeway for interpretation,when theorganizational goal represents the desired future state of the organization. An organizational goal loses clear meaning and becomes ambiguous when it invites a number of different interpretations. This definition of organizational goal ambiguity (or clarity) is consistent with some previous conceptions of the construct (DiMaggio 1987; Kelemen 2000; Locke et al. 1989; Zahariadis 1999). Goals can be ambiguous in various ways, however, and along different dimensions. We developed four dimensions of goal ambiguity that refer to communicating the reason for the existence of an organization, directing organizational activities, evaluating organizational performance, and making decisions about organizational priorities (Chun and Rainey 2005). Measuring Organizational Goal Ambiguity The measures for variables usually appear in the method section. Given the newness of the four concepts and measures of goal ambiguity, to aid the reader we describe the measures in this section aswe introduce the concepts. As described below, the data formost of the dimensions of organizational goal ambiguity were collected from the agencies’ strategic plans and performance reports. The Government Performance and Results Act of 1993 requires that virtually every federal agency describe the agency’s goals and performance indicators in the strategic plans and in annual performance plans and performance reports that must be submitted to Congress (U.S. Office of Management and Budget [OMB] 2001). This provides access to information about the formally stated goals of most federal agencies. The measures of the goal ambiguity dimensions we employ do not have a long history of use, so we provide evidence of criterion validity, especially convergent validity, in notes in the following sections. 1 There has been little clarification of the relations among such goal attributes as goal vagueness, specificity, complexity, multiplicity, conflict, tangibility, and measurability. The relationships among these ambiguity-like constructs remain ambiguous themselves. A focus on the level of interpretive leeway involves conceiving goal ambiguity as a general concept incorporating these seemingly interrelated goal attributes. 2 Using the GPRA plans and reports as data sources to measure goal ambiguity in federal agencies has several significant advantages in dealing with methodological complications that have hampered previous research on organizational goals. First, it is a fairly reasonable way to identify organizational goals in government agencies, since the goals of government agencies are essentially mandated by statutes (Lerner and Wanat 1983) and the GPRA explicitly requires federal agencies to develop their goals based on such statutory mandates (U.S. OMB 2001). In fact, researchers on public management have often identified the goals of government agencies by looking at formal mandates (for examples, see Meyers, Riccucci, and Lurie 2001; Perry et al. 1999). The second advantage has to do with a challenge encountered in previous research on organizational goals that has focused on mission statements (Weiss and Piderit 1999) or vision statements (Baum, Locke, and Kirkpatrick 1998). Organizational mission or vision statements are often idealized, symbolic, and brief (Levin 2000). The GPRA plans and reports, however, contain not only the mission statement of an agency but also a description of the agency’s long-term goals and annual performance targets, which convey in-depth information about the agency’s goals. Third, it canminimize the ‘‘whose goal?’’ problem (Grizzle 1982; Vancouver and Schmitt 1991) in organizational goal research. Consultations with key stakeholders are mandatory procedures in preparing the strategic plan of a federal agency (Franklin 2001; Roberts 2000). As the main purpose of the consultations is to draw a workable consensus among stakeholders concerning the goals of a federal agency, the agency goals finally described in the GPRA plans and reports can be considered as agreed-upon goals among key stakeholders. An additional advantage comes from the standardization of terms that otherwise could lead to significant measurement errors when organizational data were collected from archival sources (Van de Ven and Ferry 1980). All of the plans and reports were prepared under the same guidelines provided by the OMB. Chun and Rainey Goal Ambiguity and Organizational Performance 531 at R uers U nirsity on A uust 2, 2010 http://jpaordjournals.org D ow nladed fom Mission Comprehension Ambiguity Mission comprehension ambiguity refers to the level of interpretive leeway that an organizational mission allows in comprehending, explaining, and communicating the organizational mission (Daft 1998; Dess andMiller 1993; Thompson 1997). Organizational leaders often promulgate mission statements to enhance the organization’s legitimacy and in turn to enhance members’ commitment (Gable 1998; Richards 1986; Scott 2003) and ‘‘sense of mission’’ (Campbell and Nash 1992; Wilson 1989). Researchers analyzing mission statements have used the concept of mission statement clarity to refer to the degree to which the mission statement is easy to read and understand (Campbell and Nash 1992; Weiss and Piderit 1999). Consistent with these previous approaches, we conceive mission comprehension ambiguity as referring to the understandability of the mission statement. When the mission statement is easier to understand, explain, and communicate, there will be less leeway for interpretation and more shared agreement about its meaning. Measuring Mission Comprehension Ambiguity. We collected the agencies’ mission statements from their strategic plans during the period of 1997–99 and calculated the Gunning-Fog Index (GFI) for eachmission statement (Gunning andKallan 1994). TheGFI is an established indicator for evaluating the degree of ‘‘fog’’ in a written passage. The GFI has predicted the extent to which a piece of writing would be easily understood by readers. Previous studies have used it as an indicator of mission statement clarity (e.g., Weiss and Piderit 1999). The higher the GFI, the harder the written passage is to comprehend (Gunning 1968). The GFI scores showed reasonable convergent validity with an independent rating of the ‘‘transparency’’ of the performance reports. Directive Goal Ambiguity Directive goal ambiguity refers to the amount of interpretive leeway available in translating an organization’s mission or general goals into directives and guidelines for specific actions to be taken to accomplish the mission (Dess and Miller 1993; Moore 1995; Scott 2003). Other scholars have treated the room for interpretation in translating organizational missions into concrete activities and behaviors as an important facet of goal ambiguity 3 The procedures to get the GFI score of a mission statement were as follows: (a) count the number of words in successive sentences in the statement and divide the total number of words in the statement by the number of sentences, which gives the average sentence length of the mission statement; (b) count the number of words of three syllables or more in the statement and figure the percentage of such hard words out of total words in the statement; and finally, (c) total the two factors. 4 In general, convergent validity is established by demonstrating a high correlation between scores from two different measures of the same construct (Schwab 1999). It is ideal that the criterion and themeasure being validated are collected by totally different methods (O’Sullivan and Rassel 1995). The alternative measure used for validating the measure of mission comprehension clarity usedwas the ‘‘transparency’’ score reported in the Performance Report Scorecards (PRS) released by theMercatusCenter (2000–2) atGeorgeMasonUniversity. Since 2000, the center has annually evaluated the quality,which has been assessed based on three dimensions including ‘‘transparency,’’ of theGPRAperformance reports in twenty-four federal agencies covered under the Chief Financial Officer (CFO) Act. The transparency dimension, which consisted of four evaluating factors including understandability, has been rated on a twenty-point scale by experts in the center. As an agency’s mission statement is supposed to play a key role in preparing the performance report of the agency (U.S. OMB 2001), it was assumed that the transparency score of the performance report should be negatively correlated with the degree of difficulty in comprehending the mission statement. We calculated Pearson correlations between themission comprehension ambiguitymeasure used in this study and the average transparency score in the PRS for the last three years in the twenty-four CFO Act agencies. The two measures were correlated negatively, as expected, andmoderately at .37. Given that understandability is only one of four factors constituting the transparency score in the PRS, which might lower the correlation between the two measures, to find a moderate correlation in the expected direction provides support for convergent validity of the mission comprehension ambiguity measure used. 532 Journal of Public Administration Research and Theory at R uers U nirsity on A uust 2, 2010 http://jpaordjournals.org D ow nladed fom (Ginger 1998; Lowi 1979; Sharkansky 1999; Spicer and Terry 1996). For example, Lerner and Wanat’s (1983) concept of the ‘‘fuzzy mandates’’ of public bureaucracy taps the same construct as directive goal ambiguity when the authors point out that fuzzy terms in legislation provide too little guidance for crisp implementation of the legislative mandates. Measuring Directive Goal Ambiguity. When the terms used in goal statements do not lend themselves to precise definition, goals lose clarity as directives for day-to-day decisions. To translate general and fuzzy formal mandates from legislation into specific guidelines, government agencies often have to refine statutory language by issuing administrative rules (Ripley and Franklin 1991).We used the indicator ‘‘rules to laws ratio’’ (R/L ratio) because it indicates the extent to which a federal agency needs to clarify vague congressional intent or directions by adding specifications. This indicator is the ratio of the number of pages of administrative rules that the agency issues to the number of pages of legislation that the agency administers. Meier (1980) developed this measure in his work on agency power. He used the R/L ratio as an indicator of the autonomy of federal agencies because he posited that more ambiguity in statutes would allow more autonomy for agency officials. Evaluative Goal Ambiguity Evaluative goal ambiguity refers to the level of interpretive leeway that a statement of organizational goals allows in evaluating the progress toward the achievement of the mission. For performance evaluation, the organizational mission should be transformed into performance indicators and targets (Grizzle 1982). Organizations vary in the extent towhich performance targets can be precisely described and in the extent towhich valid and objective performance indicators are available (Gable 1998; Smith 1999). Some organizations can express their performance targets in an objective and measurable manner that allows a minimum level of interpretive leeway. Other organizations have difficulty specifying objective, quantitative, and outcome-focused performance indicators andmay useworkload or process indicators rather than results or outcome indicators in performance evaluation (Bohte and Meier 2000; Grizzle 1982; Merton 1957). 5 The procedures to get the R/L ratio were as follows: (a) count the number of pages of rules in the Code of Federal Regulations (CFR) for each agency, (b) count the number of pages of legislation in the U.S. Code for each agency, and finally (c) divide the first factor by the second factor. Both the CFR and the U.S. Code used for this study were issued in 2000.They covered legislation and administrative rules issued by the end of the previous year, 1999.The rules in theCFR were classified by subdepartmental agency and independent establishments, and the CFR had a parallel table of legal authorities and rules as an appendix, which makes it possible to objectively calculate the R/L ratio for each agency. To provide evidence of convergent validity, an alternative measure of vagueness of congressional directives was used. Appropriation is another important area in which congressional intent often varies in specificity. Some agencies are required to prepare their budget descriptions with great detail, while others are permitted to submit budgets devoid of detail. Ameasure of this directive ambiguity in budgets, whichwas also developed byMeier (1980), is the ratio of the agency’s budget size in dollars to the number of pages the budget takes in the appendix to theBudget of the United States Government. For a randomly selected subset of twenty sample agencies, we correlated the R/L ratio with this measure of budget direction ambiguity. The Pearson correlation was .29. Considering that these two indicators measured related, rather than identical, constructs, i.e., an indicator of the vagueness of congressional intent in legislation versus an indicator of that in budget, finding a modest but positive correlation provides support for the validity of the directive goal ambiguity measure used in the study. 6 As with the previous dimensions, evaluative ambiguity is not a new concept in the goal ambiguity literature (Rainey 1993). For example, role ambiguity researchers viewed the lack of information about the consequences of role performance (Kahn et al. 1964) or performance evaluation ambiguity (Dougherty and Pritchard 1985) as one of the multiple dimensions of role ambiguity. Chun and Rainey Goal Ambiguity and Organizational Performance 533 at R uers U nirsity on A uust 2, 2010 http://jpaordjournals.org D ow nladed fom Measuring Evaluative Goal Ambiguity. Evaluative goal ambiguity refers to the degree of difficulty in objectively evaluating progress toward the achievement of organizational goals. This dimension is measured by the percentage of subjective or workload-oriented performance indicators, as opposed to objective and results-oriented performance indicators, in each agency’s performance plan. Franklin (1999) has used a similar measure in her study of strategic planning in Arizona state agencies. Data were collected from each agency’s GPRA performance plan released in 1998–99 or performance report released in 1999–2000. As with strategic plans, we collected the GPRA-required performance plans and reports either from the web sites of the agencies or by contacting agency officials in planning units. In this study, ‘‘subjective’’ performance indicators refer to measures based solely on individual perceptions about the level of organizational performance and frequently without a numerical target level. Although many subjective indicators are descriptive (i.e., they provide no numerical target level), it is possible for a performance indicator to be subjective and quantitative at the same time. Objectivity in performance evaluation could be obtained by observing tangible conditions or events, such as a moon landing. The performance indicators in the GPRA performance plans and reports are meant to be objective and/or quantitative. Where such indicators are not feasible, however, the GPRA allows use of subjective and descriptive performance indicators, with the U.S. OMB’s (2001) approval. On the other hand, ‘‘workload-oriented’’ performance indicators refer to input and output indicators as opposed to such ‘‘results-oriented’’ ones as outcome and efficiency measures. It should be noted, however, that to distinguish ‘‘results-oriented’’ indicators from ‘‘workload-oriented’’ ones is not as simple as it sounds since the difference between the two is not always crystal clear. To deal with this problem, in classifying the performance indicators reported in the GPRA plans and reports we developed very specific criteria supplemented by examples. The criteria are described in appendix 1. Inter-rater reliability was calculated for a subset of the sample of agencies, and two alternative measures were used for validating this evaluative goal ambiguity measure. 7 Based on the same criteria described in appendix 1, Chun and a research assistant, a doctoral student majoring in public management, independently coded the performance indicators of twenty sample agencies that were randomly selected and calculated the evaluative goal ambiguity scores for the agencies. Rater reliability for the resultant evaluative ambiguity scores was shown by a correlation between raters of .91 (p , .0001). Two alternative measures were used for validating this evaluative goal ambiguity measure. The first was the General Accounting Office’s (GAO) grades on the ‘‘clarity of intended performance’’ in twenty-four CFOAct agencies’ performance plans released in 1999 (U.S. GAO 1999). In assessing the clarity of performance picture provided by the performance plan, the criteria used by the GAO were as follows: the results orientation of performance measures, the possibility of before-and-after performance evaluation, and the presence of performance targets for managerial challenges. For calculating the correlations between the GAO’s grades and the evaluative goal ambiguity measure used in this study, the four grades employed by the GAO were converted to a four-point scale, i.e., 1—clear picture of performance, 2—general, 3—limited, 4—unclear. The Pearson correlation was .52, statistically significant at the .01 level. This result strongly supports the concurrent validity of the evaluative goal ambiguity measure used in the study. Another alternative measure was the ‘‘public benefits’’ score provided by the Performance Report Scorecards of the Mercatus Center, which were described earlier. The public benefits dimension, which was one of the three dimensions that served as the criteria to assess the quality of the GPRA performance reports in twenty-four CFO Act agencies, included such evaluating factors as the results orientation of the agency’s goals and performance measures, the demonstration of a clear relationship between the agency’s work and results, and the presence of cost information. We calculated Pearson correlations between the measure of evaluative goal ambiguity used in the study and the average public benefits score of the PRS for the last three years in the twenty-four CFO Act agencies. The correlation was .67, statistically significant at the .01 level, which provides further support for the concurrent validity of the measure used. 534 Journal of Public Administration Research and Theory at R uers U nirsity on A uust 2, 2010 http://jpaordjournals.org D ow nladed fom Priority Goal Ambiguity Priority goal ambiguity refers to the level of interpretive leeway in deciding on priorities among multiple goals. To indicate priorities means to make decisions about which goals should take precedence over others at a given time or to form a goal hierarchy in which the goals are vertically arranged through means-ends relationships (Richards 1986). The presence of multiple goals without any hierarchical arrangement and prioritization leaves much room for interpretation of such priorities and about which goals take precedence. This dimension is similar to several existing constructs in the goal ambiguity literature, such as goal focus (Weiss and Piderit 1999) and goal complexity (Lee, Locke, and Latham 1989). Measuring Priority Goal Ambiguity. For priority ambiguity, or the degree of imprecision in indicating priorities among multiple goals and performance targets, we used two indicators: (a) the number of long-term strategic goals and (b) the number of annual performance targets. Previous research on goals in public organizations has used indicators such as these to measure the extent to which multiple organizational goals or ‘‘goal equivalents’’ are simultaneously presented without any prioritization (Franklin 1999; Weiss and Piderit 1999). To combine the two indicators, the number of strategic goals and the number of annual performance targets, into a composite measure of priority goal ambiguity, it was necessary to standardize the indicators so that each was based on the same scale (O’Sullivan and Rassel 1995). The Z-scores of each of the indicators were used for this standardization, and the average of the two Z-scores was calculated as the priority ambiguity score for each agency. The Goal Ambiguity Measures and the Antecedent Variables Three of these goal ambiguity measures performed well in the previous study of the ‘‘antecedents’’ that might influence them (Chun and Rainey 2005). The antecedent 8 For example, Lee, Locke, and Latham defined goal complexity as ‘‘the number of different intended outcomes and their interrelationship’’ (1989, 299). Whereas goal complexity pertains to interconnections among goals in a general sense, the concept of goal conflict refers to only contradictory relationships among them. Goals can counteract one another at the same hierarchical level as well as at different levels (Behn 1991, 70). Goal conflict at the same level may have two different forms: direct and indirect conflict. Direct conflict among goals represents explicit trade-offs among multiple goals, and indirect goal conflict refers to potential competition between the goals in acquiring resources. Direct goal conflict may occur when achieving one valued goal directly inhibits achieving another desired goal (Lee, Locke, and Latham 1989). In contrast, indirect goal conflict can be observed whenever there are multiple goals at the same hierarchical level, since more goals means greater need to split limited resources and thereby exposes more points of potential conflict. 9 Data for long-term strategic goals for each federal agency were obtained from the GPRA strategic plans released in 1997–99, and data for annual performance targets, from the GPRA performance plans submitted to Congress in 1998–99 or from the performance reports in 1999–mid-2000. The strategic plan of a federal agency describes longterm strategic goals that define how the agency will carry out its mission over a period of time. Similarly, the performance plan of the agency reports annual performance targets, i.e., specific milestones in achieving the long-term strategic goals. The number of annual performance targets tends to be greater than the number of long-term strategic goals as the progress toward achieving a strategic goal is often demonstrated by several performance measures. 10 To provide some evidence of concurrent validity, we collected data on the number of organizational units, including staff units, that directly report to the agency head in a randomly selected subset of twenty sample agencies. These data were collected from organizational charts posted on the web sites of the agencies. While this indicator does not exactly measure the same construct as priority goal ambiguity, it is a closely related construct in that different organizational units directly reporting to the agency head should represent distinct priorities in the agency. The Pearson correlation between the two measures was .58, statistically significant at the .05 level, which strongly supports the convergent validity of the priority goal ambiguity measure used. Chun and Rainey Goal Ambiguity and Organizational Performance 535 at R uers U nirsity on A uust 2, 2010 http://jpaordjournals.org D ow nladed fom variables, which are used as control variables in the present analysis, showed statistically significant and reasonably strong relations to three of the goal ambiguity measures, in support of hypotheses and in logically explicable patterns. Financial publicness was negatively related to directive, evaluative, and priority goal ambiguity, indicating that those agencies receiving higher levels of funding through government allocations (as opposed to sales or user charges) have higher levels of goal ambiguity (as contrasted with more ‘‘businesslike’’ agencies that rely less on government allocations, and that the results showed to have lower goal ambiguity). Regulatory agencies had higher levels of directive and evaluative ambiguity, as many authors would have predicted, but not of priority goal ambiguity (i.e., the number of different objectives they proclaim). This is consistent with the interpretation that they have more general and vague mandates and goals but not necessarily more different objectives than other types of department. The variable competing demands (the number of different groups that testify about the agency) was strongly positively related to priority goal ambiguity but not to the other types of goal ambiguity. This indicates that the more groups involved, the more different objectives the agency states, but the objectives do not show higher levels of evaluative ambiguity. The measure of mission comprehension ambiguity, however, showed no significant relations to any of the ‘‘antecedent’’ variables, thus raising questions about the value of the GFI as a measure of agency mission characteristics (and indicating implications for interpreting the present analysis). In sum, the mission comprehension ambiguity measure did not perform well, but the other three goal ambiguity measures did. They produced fairly strong results that supported hypotheses, and there were logically reasonable differences among the three measures in their relations to the ‘‘antecedent’’ variables. CONCEPTUAL FRAMEWORK AND HYPOTHESES The present study focuses on four different variables representing organizational performance, including survey responses about managerial effectiveness, customer service orientation, productivity, and work quality. Below we will hypothesize the same relationship between each goal ambiguity measure and each of these performance measures, so for brevity we state one hypothesis about performance for each goal ambiguity measure that refers to all four performance measures. Goal Ambiguity and Organizational Performance Mission Comprehension Ambiguity and Performance Organizational mission statements have become virtually ubiquitous in the United States. This proliferation of mission statements implies the assumption that comprehension of the organization’s mission enhances the organization’s performance. Conversely, mission comprehension ambiguity should be negatively related to organizational effectiveness. Various authors have argued that an attractive mission benefits an organization in numerous ways that link to performance (Rainey and Steinbauer 1999; Scott 2003; Weiss 1996; Weiss and Piderit 1999; Wilson 1989). When it is difficult to understand, explain, and communicate the mission, these benefits should wane. Actually, research findings on this relationship are mixed. Weiss and Piderit (1999) found no significant relationship 536 Journal of Public Administration Research and Theory at R uers U nirsity on A uust 2, 2010 http://jpaordjournals.org D ow nladed fom between a measure of mission statement clarity, the Gunning-Fog Index, and school performance. In contrast, Baum, Locke, and Kirkpatrick (1998) reported a significant and positive relationship between several vision statement attributes, including brevity and clarity, and the financial performance of business firms. Therefore we examine this relationship between mission comprehension ambiguity and performance as indicated by organizational members’ responses to survey items about managerial effectiveness, customer service orientation, productivity, and work quality: H1 Federal employees in agencies with higher levels of mission comprehension ambiguity will perceive lower levels of organizational performance. Directive Goal Ambiguity and Performance There have been two different lines of argument concerning the relationship between directive goal ambiguity and organizational performance. Some scholars view goal clarity in directing organizational decisions and behaviors as a critical element in establishing accountability and meeting the expectations of stakeholders. Others see it as an obstacle to learning from experiments and to flexible adjustments to environmental changes. For instance, many political scientists have considered broadly and loosely defined policy mandates for government agencies as a serious threat to democratic control and accountability (Lowi 1979; Page 1976; West 1997). Similarly, some authors contend that more specific policy mandates make implementation more effective since broadly defined mandates often permit the implementers to interpret them to fit their own values and benefits (Sabatier and Mazmanian 1979). Others, however, assert that specific policy directives can impede political responsiveness to local needs (Lerner and Wanat 1983; Lindblom 1959) and that loosely defined policy mandates allow room for experiments, learning, and adjustments to variations in implementation contexts (Berman 1978; Palumbo 1980). Among management experts, some argue for goals without specific behavioral expectations to support flexible responses to environmental changes (Conger and Kanungo 1998; Kelemen 2000), responses to unexpected opportunities (Moore 1995), avoidance of strategic myopia (Bartlett and Ghoshal 1994), management creativity (David 1989), increased administrative autonomy (Baier, March, and Saetren 1988; Ginger 1998; Kettl 1992; Mahler 1997), and decreased political conflict (Moore 1995). Others emphasize the potential costs of loosely defined policy directives such as the possibility of arbitrary judgments concerning public values (Spicer and Terry 1996) and inconsistency of decisions and activities with the overall organizational mission (Behn 1991). As Sharkansky warns, ‘‘Flexibility is an attraction but ambiguous limits to acceptable behavior invite irresponsible exploitation of flexibility’’ (1999, 11). Empirical evidence to resolve this conflict is limited but tends to indicate that directive goal ambiguity is negatively related to organizational performance in the public sector (Boal and Bryson 1987; Lan and Rainey 1992; Wright 2001, 2004). An outcome in either direction would be interesting, but we hypothesized a negative relationship: H2 Federal employees in agencies with higher levels of directive goal ambiguity will perceive lower levels of organizational performance. Chun and Rainey Goal Ambiguity and Organizational Performance 537 at R uers U nirsity on A uust 2, 2010 http://jpaordjournals.org D ow nladed fom Evaluative Goal Ambiguity and Performance The literature on public management generally holds that evaluative goal ambiguity is negatively related to organizational effectiveness, with some empirical evidence supporting this conclusion (e.g., Boyatzis 1982; Hyndman and Eden 2001; Lan and Rainey 1992); and others similarly contend that clear and measurable performance targets improve performance (Behn 1991). Tullock (1965) suggested that business managers can review such indicators as sales and profits and thus are better able than public sector managers to prevent information distortion in upward communications; such distortions, of course, impede performance. Some executives experienced in the public and private sectors have observed that the lack of a ‘‘bottom line’’ for public organizations causes executives to concentrate on general reputation, media relations, and political relations more than on substantive performance (e.g., Allison 1983; Blumenthal 1983). Thus we have the following hypothesis: H3 Federal employees in agencies with higher levels of evaluative goal ambiguity will perceive lower levels of organizational performance. Priority Goal Ambiguity and Performance Peter Drucker wrote that ‘‘to try to do several things at once’’ in public agencies is ‘‘one of the deadly sins in public administration’’ (1980, 103). Based on her experience, Shalala advises that a government executive must ‘‘set firm goals and priorities and stick with them’’ because ‘‘if you try to do everything, you will accomplish nothing’’ (1998, 287). In a case study on the Federal Deposit Insurance Corporation, Khademian (1995) found that political consensus on a single priority or a ‘‘bottom line,’’ the solvency of the Bank Insurance Fund, played a key role in improving the agency’s effectiveness by letting the managers of the agency concentrate their efforts on the achievement of the ‘‘bottom line.’’ Analyzing local welfare agencies, Meyers, Riccucci, and Lurie (2001) also reported that the existence of a complex set of formal goals could cause goal incongruence in implementation, uncoupling the work priorities of frontline staff from policy goals of offices at higher levels. In their study of AmeriCorps, Perry et al. (1999) concluded that the diversity of goals in the agency allowed implementers to use the situation for the benefit of their own goals (although they also found a positive impact on organizational learning). Thus we hypothesize the following: H4 Federal employees in agencies with higher levels of priority goal ambiguity will perceive lower levels of organizational performance.