INTRODUCTION

In the face of widespread retirements, the reform or removal of civil service protections, increased outsourcing, reduction-in-force layoffs, and other efforts to do more with less, public agencies face the ever-growing challenge to attract and retain skilled managers. Government agencies at all levels face an impending crisis as a large percentage of ‘‘baby boomer’’ managers and professionals approach retirement. The human capital crisis is furthered by the reduced effectiveness of the once popular model of public sector employment— grooming managers from entry-level positions. In response to concerns about impending retirement and hiring restrictions, personnel systems at all levels of government have undergone continuing waves of reform, including adopting special ‘‘flexibilities’’ in hiring and pay (Government Accountability Office 2004; Hays and Sowa 2006; Partnership for Public Service 2005), whereas some state governments have been reforming or I would like to thank Vicky Wilkins for commenting on an initial draft of this article and the three anonymous reviewers who provided detailed, constructive comments for this article. Address correspondence to the author at mkfeeney@uic.edu. doi:10.1093/jopart/mum025 Advance Access publication on September 19, 2007 ª The Author 2007. 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: journals.permissions@oxfordjournals.org JPART 18:465–494 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from eliminating civil service altogether (Hays and Sowa 2006; Kellough and Nigro 2002; Walters 2002). Unfortunately, reforming civil service, making government positions available to external applicants, and expanding at-will-employment does not address an additional challenge facing the public sector, its image. Surveys continue to find that top students and mid-career professionals regard government as offering limited opportunities for challenging and exciting work (Partnership for Public Service 2002) and professionals and managers view working for the government as constraining and lacking independence, advancement, and increased responsibility (e.g., Kilpatrick, Cummings, and Jennings 1964). Governmental personnel systems have been the focus of criticism for decades (Goodsell 2004) and continue to face misconceptions of public sector work. The challenge to attract recent graduates and mid-career professionals and managers to government work (Partnership for Public Service 2002, 2004) is compounded by fierce competition from the private sector. Recent Council for Excellence in Government polls (1997 and 2002) found that adults are more than twice as likely to prefer working in the private sector compared to the public sector. A 1998 Pew Research Center poll found that 70% of respondents would prefer a private business as an employer (Partnership for Public Service 2005). These polls paint a dark picture for the public sector in the competition for talent. In the face of widespread retirements, hiring and promotion restrictions, image problems, and competition from the private sector, state-level public managers’ sector perceptions, and the determinants of those perceptions continue play an important role in the ability to attract and retain talented workers. Many government agencies have shown increasing attention to their employees’ opinions, attitudes, and quality of working life in order to attract and retain talented workers. Researchers have conducted surveys to investigate government workers’ opinions and views about sectors, their own organizations, and many other dimensions of their experiences at work. These developments indicate the value of continuing efforts to analyze and understand perceptions of workers and work life, by sector. Although massive federal surveys provide valuable information about work attitudes and opinions in the federal government, they rarely ask employees about other important matters such as career histories, reasons for making career decisions and choices, and sector perceptions. Furthermore, little public administration research addresses the attitudes and work experiences of state employees, who, in this era of devolution, are increasingly shouldering the burden of administering public programs. The analysis provided here investigates variations in sector perceptions among state managers in Illinois and Georgia. This analysis makes an important contribution to our understanding of sector perceptions, investigating the relationships between job choice motivations, career histories, and sector perceptions. More than ever, perceptions of the public sector play a critical role in attracting and retaining talented people to state government work. The findings are relevant for current and continuing consideration of the attractiveness of the public and private sectors and related challenges in recruitment and hiring. Next, with the support of the relevant literature, I present nine propositions about the relationships between public managers’ work motivations, work experiences, and demographic characteristics and their perceptions of the private and public sectors. I then present the data and models testing these relationships and conclude with a discussion of the results and their relevance for public administration research. 466 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from LITERATURE AND PROPOSITIONS Researchers have documented many distinctions between public and private organizations ranging from organizational differences to variation in employees’ motivations, perceptions, and behavior. Researchers have identified differences in risk taking (Bellante and Link 1981; Bozeman and Kingsley 1998) and perceptions of red tape (Bozeman and Loveless 1987; Bozeman, Reed, and Scott 1992; Buchanan 1975) in organizations. Researchers have also detailed distinctions between individuals’ values and commitments (Lyons, Duxbury, and Higgins 2006), desire for job security (Frank and Lewis 2004), motivations and incentives (Crewson 1997; Jurkiewickz, Massey, and Brown 1998), and job satisfaction (Blunt and Spring 1991; Perry and Porter 1982; Solomon 1986) in the public and private sectors. In sum, the public versus private literature outlines numerous tangible and normative distinctions between the sectors. This article focuses on how statelevel public managers’ perceptions of these differences vary. In particular, how managers’ positive perceptions of the public and private sectors vary due to their motivations for taking their jobs, their previous work experiences in the public and private sectors, and demographic characteristics. Manager Motivations Managers vary in their motivations to pursue a career, accept a job, and retain a position in an organization or sector. A great deal of research, using various methods and samples, has investigated if and how motivations and preferences may differ by sector (Rainey 2003, 237–47). Although diverse, research supports the conclusion that there are stable differences in the motivations of public and private sector employees (Crewson 1997; Frank and Lewis 2004; Perry 1996). Research indicates that, in general, government employees tend to value intrinsic rewards, whereas private sector employees prefer extrinsic rewards. Government respondents, especially in higher ranks, tend to place higher value on challenging and important work (Crewson 1997), which provides the opportunity to have an impact on public affairs or public policy or work on behalf of the public (Frank and Lewis 2004; Houston 2000).1 For example, Federal Human

PROPOSITION

1. Public managers driven by a desire for advancement, compared to those who are not driven by a desire for advancement, will have more positive views of the public sector. Research supports the notion that compared to private sector workers, public sector workers are less financially motivated (Frank and Lewis 2004). For example, in each of the 14 administrations of the General Social Survey (1973 through 1993) a larger percentage of private sector respondents, compared to public sector respondents, ranked ‘‘high income’’ as the most important aspect of a job. In comparison, a higher percentage of respondents working in the public sector rated ‘‘a feeling of accomplishment’’ as the most important aspect of a job (Crewson 1995). In general, public sector employees compared to private sector employees are more concerned with intrinsic rewards such as a sense of accomplishment than monetary incentives (Crewson 1997), whereas private sector workers are more likely to value extrinsic reward motivators such as high salary and shorter work hours (Houston 2000). Frank and Lewis (2004) find that people who place a high value on income are less likely to actually work for the government. Furthermore, among college students who would not seriously consider working in public service the most common (16%) reason was that they felt that the pay and benefits are poor (Partnership for Public Service 2005). Though research indicates that employee financial compensation in state and local government is increasing faster than in the private sector (Silverstein 1995), salaries for administrative and professional positions in state and local government are far lower than salaries in comparable positions in the private sector (Miller 1996). Since government work typically pays less than the private sector, it follows that something else must attract workers to the public sector, be it alternative motivations, a lack of opportunities in the private sector, or the inability to get a private sector job. PROPOSITION 2. Public managers driven by financial motives will have more positive views of the private sector. Another motivation identified by the public versus private literature involves the desire for job security. Although job security motivation may be overexaggerated by stereotypes of government bureaucrats, research suggests that people working in government jobs do value job security and find it an attractive aspect of their work setting (e.g., Frank and Lewis 2004; Kilpatrick, Cummings, and Jennings 1964; Rainey 2003). Surveys that include benefits such as health insurance, retirement, and family leave within the measure of job security are more likely to capture the importance of this motivation in job selection and commitment. Benefits play an important role in the security associated with government work since employer-provided benefits, including paid leave, pensions, and health benefits, are higher in state and local government than in the private sector (Bureau of Labor Statistics 2006; Moore 1991). Karl and Sutton (1998) find that today’s workers place more emphasis on job security, compared to workers of the 1970s and 1980s, but find no significant differences in the desire for job security between public and private employees, on average. This focus on average perceptions may disguise differences in the importance of job security due to rank and position. For example, higher ranking respondents (General Schedule 13 and above) compared to lower grade employees, tend to give lower rankings to job security as a reason to take a job with the federal government (Crewson 1995). Because managers and executives in 468 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from government usually do not have as much job security (e.g., protections against dismissal or reassignment) as do other employees, managers and executives may have less reason to seek or expect security for themselves but may still view job security as an important component of government work. PROPOSITION 3. Public managers driven by a desire for job security will have more positive views of the public sector. Theories of public service motivation (PSM) argue that an individual employee can be committed to a public program due to a personal interest or a personal identification with that program (Downs 1967), a desire to serve the public interest and patriotism of benevolence, a love for all people within our political boundaries (Frederickson and Hart 1985), and loyalty to the government (Buchanan 1975; Mosher 1968). PSM plays an important role in job choice and job performance (Perry and Wise 1990) and motivates government workers by enhancing intrinsic rewards in lieu of extrinsic rewards (Wright 2007). In an attempt to understand why certain individuals are drawn to the public sector, the PSM literature (Brewer, Selden, and Facer 2000; Crewson 1995, 1997; Wright 2001) argues that people with higher levels of PSM will be more likely to seek government employment and will report increased performance as government workers (Perry 1996; Perry and Wise 1990). Research suggests that PSM exists and is more common among workers in the public sector than the private (Houston 2000). Although researchers (Brewer, Selden, and Facer 2000) have identified different types of PSM which are differently associated with extrinsic rewards, this research is concerned with a general desire to serve the public interest and how that motivation relates to perceptions of the public sector. It follows that individuals who have a desire to serve the public interest will indicate more positive views of the public sector than the private sector. PROPOSITION 4. Public managers driven by a desire to serve the public interest will have more positive views of the public sector. Public Sector Experiences A related, very important matter that requires more research involves the experiences of people who choose government jobs. First, encountering and either successfully managing or failing to manage red tape and other procedural and organizational constraints will inevitably shape government workers’ experiences and perceptions of the public sector. Buchanan (1975) suggests that people enter public management positions with the intrinsic motives described above but have frustrating experiences which reduce their organizational commitment, job involvement, and their service ethic. It is possible that government is so inherently burdened by red tape, rigid personnel systems for managers, and other constraints that public managers regardless of their motives will inevitably encounter frustration. Red tape, defined as ‘‘rules, regulations, and procedures that remain in force and entail a compliance burden but do not advance the legitimate purposes the rules were intended to serve’’ (Bozeman 2000, 12), is associated with external laws regulating personnel and procurement (Rainey, Pandey, and Bozeman 1995) and government performance (Brewer and Walker, unpublished data), increases organizational risk aversion (Bozeman and Kingsley 1998), administrative delays, and work alienation (Aiken and Hage 1966; Dehart-Davis and Pandey 2005; Pandey and Kingsley 2000), and decreases benefits to clients (Scott and Pandey 2000). Feeney Sector Perceptions among State-Level Public Managers 469 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from Research indicates that public managers compared to private managers are more likely to report red tape and an excess of rules which constrain performance-based pay increases and promotions (Rainey, Pandey, and Bozeman 1995). Empirical research finds that perceived red tape associated with public personnel systems is negatively associated with organizational commitment and job satisfaction and contributes to alienation among public managers (Dehart-Davis and Pandey 2005). However, red tape is not solely a public sector problem (Bozeman 1993) since it exists in all bureaucracies and is related to organizational structure and size, nor is it an insurmountable challenge facing public managers. Some public managers, particularly those with more positive work attitudes are better equipped to deal with red tape, compared to those with less positive work attitudes (Pandey and Welch 2005). Given the widespread constraints of personnel systems in state government, it seems intuitive that increased perceptions of red tape will be associated with negative views of the public sector. PROPOSITION 5. Increased perceptions of red tape will be associated with decreased positive perceptions of the public sector. Regardless of the presence of red tape and other personnel constraints in the public sector, many employees choose to stay in the public sector throughout their lifetimes. Schneider (1987) suggests that employees who fit well with an organization’s values, leadership, or mission will remain in that organization despite procedural frustrations, whereas those who do not fit will leave. Extensive experience in the public sector may be an indication of shared values with the sector’s values and work environment. Thus, a public manager’s tenure in the public sector helps to shape her perceptions of the public sector. Tenure in the public sector may serve as an indication of satisfaction with one’s career, the assumption being that, regardless of constraints, an individual remains in the public sector because she prefers it. Research testing Schneider’s (1987) attraction-selection-attrition (ASA) framework and the person-organization fit framework (Chatman 1989, 1991) confirms that an individual selects into an organization and remains in an organization that fits the individual’s values. The homogenization of employees’ values, motivations, and goals is furthered as individuals who do not fit in the organization leave thus furthering the similarities of employees who remain (Schneider, Goldstein, and Smith 1995, 749; Byrne 1971; Byrne and Neuman 1992). Following Schneider’s (1987) theory that people who do not fit an organization leave, public managers with longer tenure in the organization will be more similar to one another and be more likely to share positive perceptions of the public sector, compared to the private sector. PROPOSITION 6. Longer job tenure will be related to increased positive perceptions of the public sector. Private Sector Experiences The revolving door refers to the transfer of employees across sectors, either from government service to the private sector or vice versa. Traditional revolving door restrictions and regulations prevented movement across sectors, but as today’s workers become increasingly mobile, adjustments to personnel restrictions are easing movement across sectors. Today due to reductions in public sector employment and increases in privatization, 470 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from public-private partnerships, and outsourcing, public and private managers are regularly in contact with one another. Moreover, some states have reduced revolving door restrictions, making it easier for contracting firms to recruit mid-level managers from the public sector. Experience across sectors can result in more efficient public managers who are able to apply their knowledge of both the private and public sectors to make better decisions. In addition, individuals with experiences in both sectors may have larger social and professional networks which will increase their effectiveness as managers (Kelman 1993). I expect that sector perceptions among public managers will be influenced by respondents’ previous work experiences in the private sector. PROPOSITION 7. Public managers’ whose previous job was in the private sector will report more positive perceptions of the public sector, compared to public managers whose previous job was not in the private sector. One would expect that the state and region in which a public manager works will condition their perceptions of the public and private sectors. The structure, size, and location of state government (Bureau of Labor Statistics 2006) and presence of unions will inevitably influence public managers’ working conditions. For example, a study of implementation procedures for comparable worth/pay equity policies found that seven out of eight states had gender-related disparities which they remedied through payroll adjustments for individuals or entire female-dominated job classes. The most influential factor in the implementation of comparable worth/pay equity was public employee unions (Gardner and Daniel 1998). Unions also play an important role in the frequency of public sector work stoppages. For example, Illinois ranked second among all states in the number of public sector work stoppages between 1982 and 1997, resulting in more than 1.7 million days of idleness for 161,000 workers (Cimini 1998). The availability of public sector unions and the status of civil service protections and other personnel rules vary by state. For example, Illinois grants all state employees collective bargaining rights (as of 1997, according to the American Federation of Labor and the Congress of Industrial Organizations, 56% of Illinois state employees were union members), whereas Georgia offers no collective bargaining to state employees.2 Furthermore, since 1996, all public sector employees in Georgia have been hired without civil service protections (Walters 2002). The presence and high level of activity of the public sector union in Illinois, compared to no unionization and reductions in civil service protections in Georgia, will most likely lead to more positive perceptions of the public sector in Illinois. PROPOSITION 8. Public managers in Illinois, compared to Georgia public managers, will have more positive perceptions of the public sector. Individual Demographic Factors If public managers’ perceptions of the public and private sectors vary, it is logical that individual characteristics including education, age, family structure, gender, and race 2 Out of Georgia’s 248,900 public sectors employees, 3,000 are eligible for union membership. Metropolitan Rapid Transit Authority (MARTA) employees are the only state workers granted bargaining rights. MARTA employees are not included in the NASP-III sample. Feeney Sector Perceptions among State-Level Public Managers 471 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from will play a role in shaping these perceptions. I control for these factors, but also offer a proposition about the relationship between education type and sector perceptions. PROPOSITION 9. Compared to other college graduates, public managers graduating with degrees in public administration, public policy, and public service will have more positive perceptions of the public sector, whereas graduates with business degrees will have more positive perceptions of the private sector. Research indicates that college graduates, compared to less-educated respondents, are not more likely to prefer public sector work but are more likely to have government jobs (Frank and Lewis 2004). Though the public versus private literature investigates the role of education levels in obtaining jobs and salary differences, education type may also be an important determinant of job availability, income, and career success. Roksa (2005) finds that though college graduates earn more than high school graduates, those who major in female-dominated fields are more likely to earn lower salaries than those in maledominated fields. More important, graduates in female-dominated positions are more likely to be employed in the public sector. Given Roska’s findings, it will be important to investigate whether or not graduating with particular training is related to sector employment and perceptions. DATA AND METHODS Table 1 summarizes the propositions and their predicted relationships with the sector perceptions among public managers. Table 1 Sector Perception Propositions Proposition 1: Public managers driven by a desire for advancement, compared to those who are not driven by a desire for advancement, will have more positive views of the public sector. Proposition 2: Public managers driven by financial motives will have more positive views of the private sector. Proposition 3: Public managers driven by a desire for job security will have more positive views of the public sector. Proposition 4: Public managers driven by a desire or to serve the public interest will have more positive views of the public sector. Proposition 5: Increased perceptions of red tape will be associated with decreased positive perceptions of the public sector. Proposition 6: Longer job tenure will be related to increased positive perceptions of the public sector. Proposition 7: Public managers’ whose previous job was in the private sector will report more positive perceptions of the public sector, compared to public managers whose previous job was not in the private sector. Proposition 8: Public managers in Illinois, compared to Georgia public managers, will have more positive perceptions of the public sector. Proposition 9: Compared to other college graduates, public managers graduating with degrees in public administration, public policy, and public service will have more positive perceptions of the public sector, whereas graduates with business degrees will have more positive perceptions of the private sector. 472 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from The propositions are operationalized with variables developed from the National Administration Studies Project (NASP)-III questionnaire. The predecessors to NASP-III also focused on sector difference and state managers’ perceptions. In 1992, NASP-I compared state managers in New York, Colorado, and Florida. In 2003, NASP-II expanded to include managers in 50 states but focused on state health policy and was limited to managers in state departments of health and human services. This newest version of NASP, closed in January 2006, continues to expand our empirical k public sector perceptions). Respondents were assigned one point for each response that is favorable to the public sector and a zero for each response indicating a perception of no difference or favorable private sector perception. The advantage of the Public Sector Perception variable is that we can see variance in the strength of positive public sector perceptions which includes slight, medium, and strong public sector perceptions. The second dependent variable is Private Sector Perception which measures the frequency of positive perceptions of the private sector, from one (slight private sector perceptions) to three (strong private sector perceptions), with zero being no difference and favorable public sector perceptions. The variables Public Sector Perception and Private Sector Perception enable me to investigate the strength of positive public sector and private sector perceptions, respectively. The weakness of these variables is that the comparison group includes those with favorable perceptions for the other sector and those who perceive no differences between the sectors. The third dependent variable, Overall Sector Perception, measures private and public sector perceptions compared to no perceived sector differences. For this variable, I calculated the overall sector positive perceptions for each respondent. For example, if a respondent favored the public sector on two items and the private sector for one item, she was assigned an overall favorable public sector perception. If another respondent indicated no difference for item one, a favorable public sector perception for item two, and a positive private sector perception for item three, he was assigned no overall positive perception. The advantage of the Overall Sector Perception variable is that it captures private and public sector perceptions in comparison to one another and to having perceiving no difference at all. The weakness of the Overall Sector Perception variable is that respondents that split positive sector perceptions will result in a wash out—that is a positive public perception for the first item will cancel out a positive private sector perception for the second item.3 Table 3 illustrates the coding process used to develop the three dependent variables. By including all three dependent variable scales, I hope to gain a more informed perspective of sector perceptions among state government managers. Independent Variables The primary independent variables of interest in this analysis, which focus on job motivations, red tape perceptions, job history, state of employment, and education level, are drawn from previous research investigating factors which shape public manager preferences and experiences. For example, Frank and Lewis (2004) found that education level plays a significant role in shaping government workers preferences for public sector work. This research builds on these findings to investigate the role of having a public administration, policy, or business degree. This research builds on the abundance of research investigating public manager motivations to investigate how manager motivations shape sector perceptions. For example, in a review of the PSM literature, Wright (2001) notes 3 I also calculated the sector perceptions by assigning the three sector perception items the following values: private sector perception (
1), no difference (0), and public sector perception (þ1). I then summed the perceptions for the three items per individual. This dependent variable ranged from strong private sector perception (
3) to strong public sector perception (þ3), with zero indicating no preference. Because this seven category variable is difficult to interpret using multinomial logit, I do not present the results here. Results from the seven category multinomial logit are available upon request. 474 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from that sector employment choice, employee motivations, job satisfaction, job characteristics, and work context shape work motivation. This analysis investigates how job characteristics, work context, previous work experience, work motivations, and individual characteristics shape perceptions. To test the relationships between job motivations and sector perceptions, I have included four items: Advancement Motivation, Financial Motivation, Job Security Motivation, and Public Service Motivation (PSM), which capture respondents’ motivations for taking their current jobs. NASP-III asks respondents to respond to the following directions, ‘‘We are interested in the factors that motivated you to accept a job at your current organization. Please indicate the extent to which the factors below (some personal, some family, and some professional) were important in making your decision to take a job at your current organization.’’ Response options were a four-point Likert scale (responses include very important, somewhat important, somewhat unimportant, and not at all important). (See Appendix for full descriptive statistics and frequencies.) The variables Financial Motivation and PSM4 are responses to the questionnaire items Salary and Ability to serve the public and the public interest, respectively. Security Motivation and Advancement Motivation are scales from a factor analysis of the following items: Opportunity for advancement within the organization’s hierarchy The organization’s pension or retirement plan Desire for increased responsibility Benefits (medical, insurance) Few, if any, alternative job offers The factor analysis using an orthogonal solution and Varimax rotation of the five items above resulted in an optimized distribution of variance along two dimensions: security and advancement (see table 4). The security and advancement dimensions represent 60.5% of Table 3 Coding Process of Dependent Variables Dependent Variables Respondent Item 1 Item 2 Item 3 Public Sector Perception Private Sector Perception Overall Sector Perception R1 Public Public Public 3 0 Public R2 Private Private Private 0 3 Private R3 No difference No difference No difference 0 0 No R4 Public Private No difference 1 1 No R5 Private Private Public 1 2 Private R6 Public No difference No difference 1 0 Public R7 Public Private Public 2 1 Public R8 Private No difference No difference 0 1 Private R9 Public Public No difference 2 0 Public R10 Private Private No difference 0 2 Private 4 I also tested PSM (ability to serve the public and the public interest) as a dichotomous variable (0 5 not important, 1 5 important). Because the four category PSM variable, compared to the dichotomous version, was more strongly correlated with the dependent variables and offers a more detailed understanding of PSM, I maintained the four-point scale. Feeney Sector Perceptions among State-Level Public Managers 475 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from the cumulative variance in the initial correlation matrix. The saved factor scores make up the independent variables, Security Motivation and Advancement Motivation. To test propositions 5 and 6, I have included the variable Red Tape, which is a scale (0 5 almost none to 10 5 great deal) in response to the question ‘‘How would you assess the level of red tape in your organization?’’ and the variable Current Job Tenure, a continuous variable measuring the number of years the respondent has worked in their current position.5 The dummy variable Last Job in Private Sector indicating whether or not the respondents’ previous job was in the private sector tests proposition 7.6 To test propositions 8 and 9, I include the dummy variables Georgia, PP/PA degree, and Business degree which are coded one if the respondent works in Georgia, has a public policy or public administration degree, and has a business degree, respectively.7 Control Variables I have included various controls for individual characteristics such as family structure, gender, race, rank, and age. Family structure, such as being married and having children and parental work experiences may play an important role in public managers’ sector perceptions. Studies of federal employees have found that first, a high percentage of federal managers would not recommend federal employment to their children (National Commission on the Public Service 1989) and second, although individuals who have a parent who worked for the government are more likely to work for the government they are not necessarily more likely to prefer public sector work (Frank and Lewis 2004). Since Table 4 Security Motivation and Advancement Motivation Factor Loadings Matrix Factors Questionnaire Item Security Motivation Advancement Motivation Opportunity for advancement within the organization’s hierarchy 0.255 0.768 Job security 0.773 0.043 The organization’s pension or retirement plan 0.817 0.176 Desire for increased responsibility
0.010 0.851 Benefits (medical, insurance) 0.844 0.149 Few, if any, alternative job offers 0.377
0.290 Eigenvalue 2.346 1.290 Cumulative variance 39.39% 60.61% Extraction method: Principal component analysis, SPSS v13; rotation method: Varimax with Kaiser Normalization. Rotation converged in three iterations. Factors are interpreted in terms of their highest loadings, focusing on those equal to or greater than 6.50. 5 I also tested a variable for tenure at the current organization (including previous jobs at the current organization). It did not add predictive power to the models. 6 I also tested variations of this variable, including number of last three jobs in private sector (count and ratio) and one of any three last positions being in the private sector. I elected to use the variable Last Job in the private sector because it resulted in the strongest correlations with the dependent variables. 7 I also tested a categorical variable for education which had three categories: less than college, college, and graduate degree. This variable was not correlated with the dependent variables and did not strengthen the models probably because 42% of respondents have a college degree and 42% have a graduate degree, thus reducing variance in education. 476 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from the perceptions of state government employees may also be affected by their parent’s careers, I include the control variable Parent, which is coded one if the respondent indicated that at least one parent spent most of his or her working career in the public sector. In their assessment of gender wage gaps between the public and private sectors using 1991 U.S. census data, Gornick and Jacobs (1998) find that the employment rate and salary for women in the public sector are higher than in the private sector. Frank and Lewis (2004) find that, compared to men, women are more likely to have government jobs but not more likely to prefer such work (Frank and Lewis 2004). Though women’s employment rates and earnings, on average, are higher in the public sector, women continue to earn less than men, in both sectors. Women in the public sector earn 19.8% less than men in comparable public sector positions (Belman and Heywood 1991). In addition, women make up 45.5% of public administration employees in the United States (U.S. Bureau of Labor Statistics 2005) but remain underrepresented in high-level management positions and in distributive and regulatory agencies (Miller, Kerr, and Reid 1999) and hold only 11% of city manager positions (Fox and Schuhmann 2001). The probability of women having positions of authority in the public sector remains low (Jaffee 1989), with men’s odds of reaching top managerial positions being 1.8 times higher than those of women (Wright, Baxter, and Birkelund 1995). Regardless of women’s perceptions of or participation in the public sector, research indicates that women may not have positive experiences in public sector employment (Gattiker and Cohen 1997). Given the multiple and often contradictory findings on women’s activity and perceptions of public sector work, I control for gender with a dummy variable coded one if the respondent is female.8 Public manager perceptions of the private and public sectors might also be conditioned by the sectors’ reputation as a friendly work environment for minorities. Historically, the public sector has played a critical role in advancing the rights of women, minorities, people with disabilities, and other marginalized groups in the workforce (Riccucci 2006; Shafritz et al. 1992). Today nonwhites make up 20% of the U.S. population and nearly 30% of all employees in public administration industry, with AfricanAmericans representing more than half of nonwhites in public administration (U.S. Bureau of Labor Statistics 2005). Though nonwhites have increased their presence in the public sector and state level government (Brudney, Hebert, and Wright 2000), they remain concentrated in lower levels of government, where they lack power, status, higher salaries, and prestige (Riccucci 2006) and continue to face the challenge of upward mobility (Cornwell and Kellough 1994). Yet, individuals continue to report being attracted to the public sector because they perceive it to be a friendly employer for minorities (Frank and Lewis 2004). This perception may be particularly important in states with larger minority populations.9 Given these mixed results, I control for race with a dummy variable coded one if the respondent is nonwhite. To control for rank, I include a categorical variable (five categories) for the number of employees the respondent reports supervising in the current position.10 The variable Job 8 I considered using variables for marital status and number of dependent children, but they were not correlated with the dependent variables and were highly correlated with gender and one another. To avoid multicollinearity and for the sake of parsimony, these variables are not included in this analysis. 9 Minorities make up 34% of the population in Georgia and 21% in Illinois. 10 Because the original responses range from 0 to 22,000 (some respondents manage agency sections), I collapsed the variable into four categories (zero, 1–5, 6–15, 16–50, more than 50). Feeney Sector Perceptions among State-Level Public Managers 477 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from Position: Manager is a dummy variable indicating if the respondent’s current job is in a managerial position. Since the NASP-III data set does not have individual salary information, I use the variables Job Position: Manager and Current Job Tenure as proxies for salary and seniority. According to Bridges and Villamez (1994), authority and wage levels in the public sector reflect the interaction between position and time spent in the personnel system. I have included an interaction variable for manager and tenure at the current job. Frank and Lewis (2004) find that a great deal of the difference between federal and private sector self-reported work effort is related to age. For example, younger people are less likely to prefer and have government jobs (Frank and Lewis 2004). An April 2001 Panetta Institute survey found that a mere 26% of surveyed college students are interested in working for the government (Partnership for Public Service 2005) and a Council for Excellence in Government survey of 17- to 24-year-olds found that slightly less than one quarter of respondents would consider working for the federal government and only 19% would consider working for local government. Since age may be related to perceptions of government work, I use a continuous variable to control for age.11 Statistical Approach Though the dependent variables, Public Sector Perception and Private Sector Perception range from zero to three, with one to three ranking from slight to strong positive perception, I do not use ordered logit models.12 I test all the relationships using multinomial logit models which enables the investigation of categorical (nonordinal) dependent variables but has the disadvantage of limiting the analysis to the comparison of alternatives and relative change, one pair at a time (for more information about the technique see: Aldrich and Nelson 1984 and Liao 1994). All analyses were conducted using the statistics package SPSS 14. My discussion of the results and interpretation of the models presents the overall relationships (sign and significance) and estimated odds ratios (exponenitiated b) which relate predictor variables with the three dependent variables. The following discussion focuses on the main effects model which contains the covariate and factor direct effects but no interaction effects.13 RESULTS I present the findings in three tables. Table 5 presents the variations in the strength of favorable public sector perceptions and table 6 presents the determinants of positive 11 I also attempted using variations of age, including age2 , generation X (1964–1981), and nearing retirement (50 years and older). 12 The Private Sector Perception model passed the test of parallel lines, failing to reject the null hypothesis that the beta coefficients are the same across response categories. However, the Public Sector Perception model failed the test of parallel lines (significance 0.016), rejecting the null hypothesis and indicating that the dependent variable is not truly ordinal. I suspect that the low frequency of responses for the final category of Public Sector Perception may have compromised the variance. Due to the mixed results on the test of parallel lines, I use multinomial logit models which have fewer and weaker assumptions than ordinal logit models. Though inappropriately using the multinomial logit model may produce inefficient estimates, it is preferable to the biased estimates produced by the improperly specified ordinal logit model (Liao, 1994, 50). 13 I do not present the full factorial model which includes all main effects and all factor-by-factor interactions because it is highly complex and difficult to interpret. Results from the full factorial models are available upon request. 478 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from Table 5 Strength of Positive Public Sector Perceptions Strong Public Perception Versus No Perception/Private Preference Medium Public Perception Versus No Perception/Private Preference Slight Public Perception Versus No Perception/Private Preference Variable b SE exp(B) b SE exp(B) b SE exp(B) Intercept
55.788 4955 0.861 1.679 2.774 1.143 Red tape 0.247 0.228 1.280
0.105 0.075 0.901
0.105** 0.051 0.900 Security factor 0.463 0.570 1.589
0.022 0.155 0.978
0.037 0.106 0.964 Advancement factor 1.315* 0.713 3.724 0.219 0.165 1.244 0.140 0.111 1.150 Financial motivation 0.108 0.667 1.114
0.288 0.223 0.750
0.175 0.154 0.839 Job tenure 0.102 254 1.108
0.130 0.083 0.878 0.028 0.028 1.028 Rank (job tenure manager)
0.159 254 0.853 0.161* 0.085 1.174
0.014 0.032 0.986 PSM Not at all important 2.801 1.778 16.459
17.463 3997 0.000
0.427 0.593 0.652 Somewhat unimportant
16.334 2710 0.000
0.570 0.552 0.566
0.713** 0.355 0.490 Somewhat important
0.128 0.874 0.880
0.624* 0.347 0.536
0.335 0.230 0.716 Very important 0.000 — — 0.000 — — 0.000 — — Business degree
1.851** 0.920 0.157 0.486 0.418 1.626 0.080 0.241 1.083 PA_PP degree 15.834 0.000 7529090
0.649 0.513 0.523
0.433 0.416 0.648 Female 0.117 0.859 1.124 0.057 0.315 1.059 0.069 0.212 1.071 Last job private 15.709 3066 6643316 0.727 0.648 2.069 0.248 0.333 1.282 Parent in public sector
0.190 0.861 0.827
0.339 0.313 0.713 0.083 0.218 1.087 Manager
18.551 2425 0.000 1.145* 0.631 3.143
0.302 0.380 0.739 Georgia
1.879* 1.003 0.153
0.213 0.330 0.809 0.027 0.220 1.028 No. employees supervised None 17.586 3893 43409804
0.796 0.784 0.451 0.099 0.500 1.104 1–5 employees 15.843 3893 7592554
0.134 0.624 0.875 0.278 0.436 1.320 6–15 employees 17.699 3893 48589942 0.459 0.587 1.583 0.083 0.432 1.087 16–50 employees 16.228 3893 11163790 0.530 0.636 1.699 0.332 0.462 1.394 More than 50 employees 0.000 — — 0.000 — — 0.000 — — Nonwhite 0.156 1.187 1.169
1.009*** 0.364 0.364
0.852*** 0.272 0.427 Age 0.076 0.055 1.079
0.004 0.020 0.996
0.025* 0.013 0.975 Reference category: no difference and private preference. Nagelkerke: 0.2076; x2 (significance 0.000): 104.853;
2 log likelihood intercept only 1075.45; final 970.594. *p ,.10, **p ,.05, ***p ,.01. Two tailed test of significance. 479 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from Table 6 Strength of Positive Private Sector Perceptions Strong Private Perception Versus No Perception/Public Preference Medium Private Perception Versus No Perception/Public Preference Slight Private Perception Versus No Perception/Public Preference Variable b SE exp(B) b SE exp(B) b SE exp(B) Intercept
5.144 2.300
0.585 1.495 1.421 1.321 Red tape 0.352*** 0.107 1.422 0.340*** 0.070 1.406 0.129** 0.057 1.138 Security factor 0.045 0.195 1.046 0.038 0.135 1.039 0.058 0.122 1.059 Advancement factor
0.078 0.194 0.925 0.067 0.139 1.069 0.138 0.126 1.147 Financial motivation 0.263 0.289 1.301
0.084 0.195 0.919
0.018 0.177 0.982 Job tenure
0.055 0.047 0.947
0.004 0.036 0.996
0.042 0.036 0.959 Rank (job tenure manager)
0.124* 0.075 0.883 0.000 0.041 1.000 0.028 0.040 1.028 PSM Not at all important 1.355 0.932 3.878
0.580 0.756 0.560
0.163 0.678 0.850 Somewhat unimportant 1.371** 0.571 3.939 0.517 0.439 1.677
0.189 0.439 0.828 Somewhat important 0.500 0.454 1.648 0.106 0.292 1.112
0.098 0.264 0.907 Very important 0.000 — — 0.000 — — 0.000 — — Business degree
0.325 0.423 0.723
0.090 0.306 0.914 0.246 0.289 1.279 PA_PP degree
0.424 0.803 0.654
0.094 0.541 0.910
0.159 0.454 0.853 Female 0.522 0.408 1.686 0.317 0.271 1.373
0.127 0.243 0.881 Last job private
1.810*** 0.635 0.164
1.131** 0.547 0.323
1.293** 0.521 0.274 Parent in public sector 0.027 0.402 1.027 0.224 0.276 1.251
0.032 0.249 0.968 Manager 0.233 0.683 1.263 0.010 0.494 1.010 0.462 0.450 1.587 Georgia 0.532 0.406 1.703 0.264 0.279 1.302 0.091 0.255 1.095 No. employees supervised None
0.365 0.912 0.694 0.533 0.656 1.704
0.476 0.580 0.621 1–5 employees 0.510 0.805 1.665 0.624 0.579 1.867
0.177 0.493 0.838 6–15 employees 0.149 0.824 1.160 0.329 0.569 1.390
0.121 0.476 0.886 16–50 employees
0.482 0.904 0.618 0.060 0.606 1.062
0.258 0.511 0.773 More than 50 employees 0.000 — — 0.000 — — 0.000 — — Nonwhite 0.196 0.578 1.216
0.060 0.348 0.941
0.165 0.302 0.848 Age 0.049** 0.025 1.050
0.023 0.017 0.977
0.003 0.015 0.997 Reference category: no difference and public support. Nagelkerke: 0.217; x2 (significance 0.000): 117.5;
2 log likelihood intercept only 1350.345; final 1232.877. *p ,.10, **p ,.05, ***p ,.01. Two tailed test of significance. 480 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from private sector perceptions. Table 7 presents overall public and private sector perceptions compared to one another and compared to being having no sector perceptions. Public Sector Perceptions Two variables (1) career advancement motivation and (2) business degree are significantly related to having overall positive perceptions of the public sector compared to the private sector (table 7). First, the odds that a state government manager will have positive public sector perceptions are 3.7 times higher for those who report increased career advancement motivation compared to those with lower motivation for advancement. Second, the odds of reporting positive perceptions of the public sector compared to the private sector for a public manager who has a business degree are 92% as high as the odds for one without a business degree. The variable, Georgia, is negatively related to reporting a positive perception of the public sector. Working in Georgia, compared to Illinois, reduces the likelihood of having positive perceptions of the public sector compared to the private sector. Although the Georgia, advancement motivation, and business degree variables are significantly related to public sector perceptions there are no significant variables in the model comparing overall positive public sector perception to having no sector perceptions at all (table 7, column 1). In summary, there are significant predictors for having positive perceptions of the public sector compared to the private sector but no significant predictors for having positive perceptions of the public sector compared to indicating no perceptions. Table 5 shows the intensity of positive public sector perceptions. The likelihood of having a strong positive public sector perceptions increases for managers who report being motivated by advancement and decreases for managers who have a business degree and work in Georgia. Increased perceptions of red tape and being a nonwhite public manager also reduce the likelihood of positive public sector perceptions. More specifically, for a nonwhite public manager the odds of having a medium or slight positive perception of the public sector are 36% and 43%, respectively, as high as a white public manager holding all other variables at their means. Private Sector Perceptions Red tape perceptions, PSM, and a previous job in the private sector are related to the likelihood of having positive perceptions of the private sector (table 7). In general, increases in the perceptions of red tape are significantly associated with public managers’ positive views of the private sector compared to having no sector perceptions.14 More specifically, a one unit increase in the respondent’s perception of red tape increases the odds of having a favorable opinion of the private sector, compared to having no perception, by 22%. Respondents who report that their last job was in the private sector report lower positive perceptions of the private sector compared to respondents whose previous job was not in the private sector. In addition, reporting that PSM is somewhat unimportant compared to very important increases, the odds that the respondent will report a positive 14 It is possible that red tape perceptions and sector perceptions are not causal but instead, endogenous. I tested for endogeneity and ran the model by predicting sector perceptions on red tape and vice versa. I have concluded that the variables are not endogenous and red tape is a significant predictor for sector perceptions. Feeney Sector Perceptions among State-Level Public Managers 481 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from Table 7 Overall Positive Public and Private Sector Perceptions Positive Public Perception Versus No Difference Positive Private Perception Versus No Difference Positive Public Perception Versus Positive Private Perception Variable b SE Exp(B) b SE Exp(B) b SE Exp(B) Intercept
0.441 1.492
0.861 1.277
57.329 5087 Red tape
0.014 0.065 0.986 0.199*** 0.058 1.220 0.237 0.249 1.267 Security factor 0.092 0.141 1.096 0.109 0.118 1.115 0.408 0.608 1.504 Advancement factor
0.062 0.147 0.940
0.114 0.125 0.892 1.547* 0.805 4.697 Financial motivation
0.043 0.204 0.958 0.065 0.173 1.067 0.177 0.705 1.194 Job tenure 0.035 0.044 1.035 0.036 0.037 1.037 0.144 231 1.155 Rank (job tenure manager)
0.017 0.048 0.983
0.056 0.041 0.945
0.223 231 0.800 PSM Not at all important 0.978 0.932 2.659 1.150 0.826 3.159 2.593 1.910 13.373 Somewhat unimportant 0.379 0.560 1.460 1.208*** 0.461 3.346
16.776 2545 0.000 Somewhat important
0.287 0.305 0.750 0.229 0.251 1.257 0.021 0.943 1.021 Very important 0.000 — — 0.000 — — 0.000 —. — Business degree 0.324 0.342 1.382 0.075 0.270 1.078
2.340** 1.046 0.096 PA_PP degree
0.533 0.480 0.587
0.057 0.470 0.944 15.801 0.000 7284619 Female 0.212 0.281 1.236 0.182 0.235 1.200
0.003 0.926 0.997 Last job private 0.308 0.603 1.360
0.936** 0.427 0.392 16.076 2924 9583966 Parent in public sector
0.030 0.284 0.971 0.240 0.241 1.272
0.212 0.898 0.809 Manager
0.369 0.518 0.692
0.484 0.422 0.617
18.912 2291 0.000 Georgia
0.018 0.293 0.982 0.254 0.245 1.289
1.984** 1.064 0.138 No. employees supervised None 0.565 0.696 1.759 0.383 0.584 1.467 17.572 4163 42800184 1–5 employees
0.233 0.588 0.792
0.172 0.494 0.842 15.279 4163 4320989 6–15 employees 0.153 0.569 1.165
0.135 0.489 0.874 17.644 4163 45991831 16–50 employees
0.631 0.609 0.532
0.944* 0.512 0.389 16.591 4163 16042078 More than 50 employees 0.000 — — 0.000 — — 0.000 — — Nonwhite
0.481 0.331 0.618 0.095 0.302 1.100 0.438 1.259 1.549 Age 0.016 0.018 1.016 0.012 0.015 1.012 0.102 0.063 1.108 Columns 1 and 2: reference category, no difference. Nagelkerke: 0.183; x2 (significance 0.000): 91;
2 log likelihood: intercept only 1058; final 967. Column 3: reference category, private sector. Nagelkerke: 0.245; x2 (significance 0.000): 87;
2 log likelihood: intercept only 603; final 516. *p ,.10, **p ,.05, ***p ,.01. Two tailed test of significance. 482 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from perception of the private sector by 234%. There is a negative relationship between increased PSM and overall positive perceptions of the private sector. Increased perceptions of red tape are consistently associated with positive perceptions of the private sector, holding all other variables at their means (table 6). For example, a one unit increase in the variable, red tape, measuring red tape perceptions increases the odds of having a slight positive private sector perception item by 14% and increases the odds of having a strong positive private sector perception by 42%, compared to having a favorable public sector perception. This comparison shows that in addition to being related to private sector perceptions, red tape perceptions are related to the intensity of positive perceptions and the difference between having a slight or a strong positive view of the private sector. Public manager’s whose last job was in the private sector, overall, have decreased odds for indicating positive perceptions of the private sector, regardless of the strength of that view. For example, the odds of having a strong positive perception of the private sector for someone whose last job was in the private sector are 83% as high as someone whose previous job was not in the private sector. The odds of having a medium and a slight positive perception of the private sectors for those who report a previous job in the private sector are, respectively, 68% and 73% as high as for those who did not have a previous position in the private sector. In summary, public managers who recently left the private sector are less likely than public manager whose previous position was not in the private sector to report positive perceptions of the private sector. The likelihood of reporting positive views of the private sector also are positively related to increased PSM. For public managers who report that PSM is somewhat important, the odds that a manager will have a strong, positive private sector perception are 294% higher than the odds for a manager who reports that PSM is very important. Race and having a business degree, though related to the likelihood of reporting favorable perceptions of the public sector, are not significantly related to having positive private sector perceptions. DISCUSSION The analysis presented in this article investigates whether or not there are variations in state managers’ perceptions of the quality of workers and work life, by sector, and if so, what explains that variation. Table 8 offers a summary of the findings as related to the propositions. In this analysis of the NASP-III data, there are no significant relationships between sector perceptions and gender and financial motivation. Furthermore, having a parent who worked in the public sector is not significantly associated with sector perceptions. The data support proposition 1, that public managers driven by a desire for career advancement will have more positive views of the public sector than those without advancement motivation. This finding extends previous findings that high-ranking federal employees are motivated by challenging work (Crewson 1997; Houston 2000) to managers in Georgia and Illinois state government. These data offer no support for proposition 2 which predicts a relationship between sector perceptions and financial motives. There is also no support for proposition 3, which predicts a relationship between security motives and sector perceptions. I suspect that the overall nonsignificant relationship between security motivation and sector perceptions is driven by an overall desire for job security among the sample respondents. For example, Feeney Sector Perceptions among State-Level Public Managers 483 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from 90% of the NASP-III respondents report that job security was very or somewhat important in making the decision to take the current job. The findings here support propositions 4 and 5 and the larger literatures on PSM and red tape. First, the data indicate that an increase in PSM among these state managers increases the likelihood of positive public sector perceptions and lower PSM, among public managers, increases the likelihood of positive views of the private sector. Second, increases in the perception of red tape among government managers in Georgia and Illinois decreases the likelihood of positive public sector perceptions and increases the likelihood of positive private sector perceptions. This finding points to the significant relationship between perceptions of red tape and dissatisfaction with the public sector and may also point to work alienation among state workers, the inability to develop coping mechanisms for dealing with red tape, and the possible relationship between frustrating work experiences and sector perceptions. Table 8 Summary of Propositions and Findings Propositions Findings Proposition 1: Public managers driven by a desire for advancement, compared to those who are not driven by a desire for advancement, will have more positive views of the public sector. Support Proposition 2: Public managers driven by financial motives will have more positive views of the private sector. No support Proposition 3: Public managers driven by a desire for job security will have more positive views of the public sector. No support Proposition 4: Public managers driven by a desire to serve the public and the public interest will have more positive views of the public sector. Support Proposition 5: Increased perceptions of red tape will be associated with decreased positive perceptions of the public sector. Support Proposition 6: Longer job tenure will be related to increased positive perceptions of the public sector. No support Proposition 7: Public managers’ whose previous job was in the private sector will report more positive perceptions of the public sector, compared to public managers whose previous job was not in the private sector. Support Proposition 8: Public managers in Illinois, compared to Georgia public managers, will have more positive perceptions of the public sector. Support Proposition 9: Compared to other college graduates, public managers graduating with degrees in public administration, policy, and service will have more positive perceptions of the public sector, whereas graduates with business degrees will have more positive perceptions of the private sector. No support: Regarding public administration and policy degree Support: Having a business degree is associated with positive Public sector perceptions 484 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from The data provide no support for proposition 6 which concerns the influence of public managers’ work experiences in shaping sector perceptions. Increased experience in the public sector, measured as current job tenure, is not related to sector perceptions. However, an increase in age is positively related to reporting positive perceptions of the public sector. It is possible that this analysis fails to find a relationship between job tenure and sector perceptions because job tenure is closely related to age. By dropping the control, age, from the models, there is a significant positive relationship between current job tenure and reporting positive perceptions of the public sector. The relationship between job tenure, age, and positive public sector perceptions may be interpreted as weak support for Schneider’s (1987, 442) ASA framework which argues that people are attracted to organizations which fit their values and that those who continue to work in an organization will become more homogeneous over time, whereas those who are dissimilar will leave the organization. It makes sense that increased tenure, or retention of older employees, in public organizations will reinforce positive perceptions of the public sector since people who select into and stay with an organization tend to share values and perceptions (Schneider 1987, 442). There is support for proposition 7. Public managers in Georgia and Illinois whose previous position was in the private sector are consistently less likely to report positive perceptions of the private sector, though they are not necessarily more likely to report positive perceptions of the public sector. This finding is important because it indicates that public managers who have recently exited the private sector may have done so because of dissatisfaction with the private sector. Since previous research indicates a decline in college students and young persons’ interest in the public sector (Partnership for Public Service 2005), it is possible that state governments could benefit from focusing recruitment efforts on the pool of workers in the private sector with positive perceptions of the public sector instead of new graduates with more negative perceptions of government work. The data support proposition 8 that public managers in Illinois, compared to Georgia public managers, are more likely to have positive perceptions of the public sector. Managers in Georgia have fewer positive perceptions of the public sector than public managers in Illinois. This may be the result of the lack of unions and civil service protections in Georgia and the expansion of at-will hiring in Georgia, which is at approximately 72% (Hays and Sowa 2006). On the other hand, the difference in public sector perceptions by state may be an indication that government employment in Georgia is simply not as attractive as it is in Illinois. Finally, it is possible that variation in sector perceptions by state may be the result of the distribution of respondents in Georgia and Illinois by agency type (see table 9). See Appendix 3 for a full description of how respondent agencies were categorized. Although most respondents in both Georgia and Illinois work in a redistributive agency, a larger proportion of Georgia respondents work in distributive agencies Table 9 Respondents by State and Agency Type Total Agencies Total N Georgia Agencies Georgia N Illinois Agencies Illinois N Redistributive 28 425 (54%) 13 178 (41%) 15 247 (69%) Distributive 8 133 (17%) 4 102 (24%) 4 31 (9%) Regulatory 23 231 (29%) 14 152 (35%) 9 79 (22%) Total 59 790 31 432 28 357 Feeney Sector Perceptions among State-Level Public Managers 485 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from (24%) compared to 9% of Illinois respondents. It is possible that the larger percentage of Illinois managers working in redistributive agencies could drive the positive perceptions of the public sector in Illinois compared to Georgia. However, agency type is not significantly correlated with the dependent variables and not a significant predictor of sector perceptions.15 Finally, although having a public policy or public administration degree is not significantly related to sector perceptions, state government managers with a business degree, compared to those without a business degree, are less likely to have positive perceptions of the public sector. This seems to be a plausible finding. Those who pursue degrees in public administration and policy demonstrate a preference for the public sector by choosing those fields of study. In contrast, those who pursue degrees in business but end up working in the public sector are most likely settling for a position in a sector that was not their first choice. CONCLUSION Changes in state governance structures and factors such as increased outsourcing with private firms, reductions in civil service protections, and competition from the private sector and the ever-growing nonprofit sector are altering public managers’ perceptions of the private and public sectors. Public administration theory and research continues to detail the distinctions between the public and private organizations, from differences in political and market authority (Bozeman 2004) to the intensity of principles and commitment to public sector values (Antonsen and Jorgensen 1997; Haque 2001). This research assesses state-level public managers’ perceptions of distinctions between the public and private sectors. Whether these perceptions of sector distinctions are reality or public manager narrative, they are informed by individual work motivation, previous work experiences, current work experiences (red tape perceptions and state), and educational training. This article contributes to our understanding of manager perceptions, in particular, the role of the motivations for job choice, private sector work experience, and manager characteristics in shaping those perceptions among state managers. This article presents empirical support for the important role that employee reward preferences, in particular, PSM and career advancement motivation, play in shaping public managers’ perceptions of the public and private sectors. In addition, these findings support previous research on the relationships between red tape and sector perceptions. State government managers in Georgia and Illinois with an increased desire for career advancement and PSM and previous work experience in the private sector report increased positive perceptions of the public sector. In contrast, state managers who have a business degree or report increased red tape in their organizations report more positive perceptions of the private sector, compared to the public sector. Understanding the relationship between work motivation, training, and previous work experiences will become increasingly important as state governments seek to attract and retain dedicated, talented workers in the face of widespread retirements and increased competition from the private and nonprofit sectors. The initial findings, though complex, point to two new and important relationships which warrant further research. First, state public managers, in this sample, whose previous 15 I included agency type in each of the regression models and found that agency type was not significant and did not significantly affect the beta coefficients for state (Georgia/Illinois) or any of the other predictors in the regression models. These results are available upon request. 486 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from position was in the private sector, are more likely to report decreased positive perceptions of the private sector. It is plausible that these individuals, moving from the private to the public sector, are driven by a desire for challenging work and advancement motivation or a distaste for the private sector. Future research should explore the motivations, perceptions, and characteristics of individuals moving to and from public and private sector work, and how they are related to outcomes and productivity. In addition, research should investigate the advantages of recruiting public managers from the pool of private sector workers. The second important finding is that, in general, managers working in Georgia, compared to Illinois, are less likely to have positive perceptions of the public sector. This significant relationship may be related to the majority of public employees in Illinois being union members and all full-time employees having civil service protections, compared to no unions in Georgia and no civil service protections for Georgia employees hired after July 1, 1996. Given the expansion of at-will employment in Georgia state government, understanding the role of personnel practices and unions in shaping public managers’ work-related perceptions and outcomes will become an increasingly important research agenda as states continue to reform civil service regulations while taking on more responsibility for the administration, funding, and implementation of public programs and services. It is also possible that these differences between public managers’ perceptions, by state, are driven by the culture of public administration and government work in these regions of the country. The findings in this article point to the need for a deeper understanding of regional and cultural differences that may drive perceptions of government work in the United States. Public administration is defined by the theory that public and private organizations, though similar, serve different purposes in society. Because public organizations administer the public will and the public good according to complex accountability structures, employees within public organizations face unique and often complex pressures, incentives, and personnel structures. Though the nature of state government work has changed significantly in the last few decades, understanding the motivations and perceptions of public managers remains an important component of public administration research. Public managers’ perceptions of the public and private sectors are important because they are directly related to organizational and individual morale. The perception that the public sector is an unpleasant, unchallenging place to work will affect the morale of other employees and clients. Since public managers are in leadership positions where they affect the lives of their peers, subordinates, and clients, it is important to understand their perceptions of the public and private sectors. Furthermore, these perceptions may affect retention and turnover in the public sector. As state governments face widespread baby boomer retirements, the reduction of job security through civil service reforms, an increase in reduction-in-force layoffs, and competition from the nonprofit sector which is attracting public service-minded young people, the public sector’s image and public managers’ perceptions of the sectors will become increasingly important in shaping state governments’ ability to attract and retain talented public managers. This research takes an initial step toward uncovering the factors related to variation among state public managers’ perceptions of work and work life, by sector, and reinforces the importance of understanding managers’ perceptions of red tape, PSM, and previous work experiences to explain their current work choices and sector perceptions. Feeney Sector Perceptions among State-Level Public Managers 487 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from APPENDIX 1 National Administrative Studies Project III The NASP aims to increase our empirical knowledge of public management and administration. NASP-III is an attempt to blend the goals of NASP-I and II while addressing a few new themes of its own. NASP-III collected data from a random sample of public and nonprofit managers in Georgia and Illinois. Unlike NASP-II, which focused on a single functional agency (health and human services), the NASP-III sample includes managers from agencies and organizations of numerous functions. The population of managers in Georgia was drawn from the Georgia Department of Audits comprehensive list of state employees who were on state agency payrolls during the 2003/2004 fiscal year. We removed employees at technical colleges, commissions, authorities, the office of the governor, and institutions from the judicial or legislative branch. In addition, we removed employees at institutions with less than 20 employees. The population included any job titles coded as ‘‘director,’’ ‘‘coordinator,’’ ‘‘officials or manager,’’ and ‘‘professionals’’ under the pay grade of 017 and all individuals with a pay grade of 017 or higher. The resulting population included 6,164 Georgia managers. The population of managers in Illinois was developed through a Freedom of Information Act request for a list of all state employees designated as either ‘‘senior public service administrators’’ or ‘‘public service administrators.’’ This list included information on 5,461 state employees, including name, agency, and county. Survey Administration The survey administration included a precontact letter, Wave I survey with letter, followup postcard mailing, Wave II mailing, follow-up contacts by phone call and e-mail, and a final Wave III mailing. The survey was closed on January 1, 2006. Response Rates Though we began with a sample of 2000 public sector respondents, our sample was reduced to 1,849 (912 Georgia, 937 Illinois) because of respondents who had retired (16 cases) or were no longer working for the state (135 cases). The survey was closed with 432 responses from Georgia and 358 from Illinois. The respondents represent a random sample of the population of managers in Georgia and Illinois. Respondents and nonrespondents do not vary significantly by state, gender, job rank, salary (for Georgia), or agency of employment. APPENDIX 2 Frequencies for Dependent Variables: N 5 790 Work is more personally gratifying: Public sector 37%, no difference 40%, business sector 22%. Managers have more work autonomy: Public sector 15%, no difference 23%, business sector 60%. Persons doing similar jobs are more talented: Public sector 9%, no difference 54%, business sector 35%. 488 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from Positive public sector perception, frequencies: No perception and positive private sector perception, 407 (52%); slight positive public sector perception, 260 (33%); medium positive public sector perception, 95 (12%); strong positive public sector perception, 11 (1%); N 5 773. Positive private sector perception, frequencies: No perception and positive public sector perception, 191 (24%); slight positive private sector perception, 313 (40%); medium positive private sector perception, 103 (26%); strong positive private sector perception, 66 (8%); N 5 773. Overall sector perception, frequencies: No perception, 198 (25%); positive private sector perception, 409 (52%); positive public sector perception, 166 (21%); N 5 773. Frequencies for Independent Variables: N 5 790 Red tape rating: How would you assess the level of red tape in your organization? Response categories: 0 (almost no red tape) to 10 (great deal of red tape); mean 5 7.07; standard deviation 5 2.11; N 5 774. Age: age of respondent in 2005, range 23–72; mean 48; standard deviation 9; N 5 782. Gender, frequencies: female, 344 (44%); N 5 784. Public policy or public administration degree, frequencies: Yes, 47 (7%); N 5 673. Business degree, frequencies: Yes, 163 (24%); N 5 673. State, frequencies: Georgia, 432 (55%); N 5 790. Race, frequencies: Nonwhite, 145 (19%); N 5 768. Current job—Manager, frequencies: Yes, 505 (64%); N 5 790. Current job—A promotion, frequencies: Yes, 439 (56%); N 5 790. Current job—Tenure, range 1–39 years; mean 7.81; standard deviation 6.69; N 5 758. Last job was in a private organization, frequencies: Yes, 94 (12%); N 5 790. Number of employees supervised, if any, frequencies: None, 100 (15%); 1–5 employees, 205 (31%); 6–15 employees, 198 (30%); 16–50 employees, 112 (17%); more than 50 employees, 52 (8%); N 5 667. Motivation: ‘‘We are interested in the factors that motivated you to accept a job at your current organization. Please indicate the extent to which the factors below (some personal, some family, some professional) were important in making your decision to take a job at your current organization.’’ (Choice options: very important, somewhat important, somewhat unimportant, and not at all important.) Ability to serve the public and the public interest, frequencies: Very important, 336; somewhat important, 291; somewhat unimportant, 89; not at all important, 35; missing, 9; N 5 781. Salary, frequencies: Very important, 21; somewhat important, 65; somewhat unimportant, 387; not at all important, 308; missing, 9; N 5 781. Feeney Sector Perceptions among State-Level Public Managers 489 at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from Opportunity for advancement within the organization’s hierarchy, frequencies: Very important, 267; somewhat important, 363; somewhat unimportant, 77; not at all important, 76; missing, 7; N 5 783. Job security, frequencies: Very important, 490; somewhat important, 215; somewhat unimportant, 50; not at all important, 28; missing, 7; N 5 783. The organization’s pension or retirement plan, frequencies: Very important, 400; somewhat important, 270; somewhat unimportant, 73; not at all important, 41; missing, 6; N 5 784. Desire for increased responsibility, frequencies: Very important, 252; somewhat important, 379; somewhat unimportant, 93; not at all important, 56; missing, 10; N 5 780. Benefits (medical, insurance), frequencies: Very important, 447; somewhat important, 245; somewhat unimportant, 55; not at all important, 37; missing, 6; N 5 784. Few, if any, alternative job offers, frequencies: Very important, 93; somewhat important, 204; somewhat unimportant, 185; not at all important, 295; missing, 13; N 5 777. APPENDIX 3 Agency Categories I categorized the 59 agencies in the NASP-III sample into Lowi’s typology with three categories: regulatory, distributive, and redistributive, using Kelly and Newman’s (2001) typology which mapped state agencies along the criteria for a federal agency typology outlined by Lowi (1964, 1972, 1985). Below I detail the classification of these agencies. Regulatory Agencies in Georgia Department of Audits and Accounts, Department of Banking and Finance, Department of Corrections, Department of Defense, Department of Industry, Trade, and Tourism, Department of Juvenile Justice, Department of Motor Vehicle Safety, Department of Public Safety, Georgia Bureau of Investigation, Georgia Correctional Industries, Georgia Public Safety Training Center, Georgia Student Finance Authority, State Board of Pardons and Paroles, and Georgia Ports Authority. Regulatory Agencies in Illinois Illinois Department of Financial and Professional Regulation (formerly Dept of Insurance), Illinois State Police, Office of the Attorney General, Office of the State Fire Marshal, Human Rights Commission, Department of Commerce and Economic Opportunity, Department of Corrections, and Department of Human Rights, Environmental Protection Agency. Distributive Agencies in Georgia Department of Agriculture, Department of Natural Resources, Department of Transportation, Jekyll Island State Park Authority. 490 Journal of Public Administration Research and Theory at Old Dominion University on May 15, 2014 http://jpart.oxfordjournals.org/ Downloaded from Distributive Agencies in Illinois Department of Agriculture, Department of Natural Resources, Historic Preservation Agency, and Department of Transportation. Redistributive Agencies in Georgia Department of Administrative Services: State Accounting Office/Georgia Technology Authority, Department of Community Affairs, Department of Education, Department of Labor, Disability Adjudication Services, Department of Revenue, Department of Technical and Adult Education, Department of Veterans Service, Georgia Department of Community Health, Department of Human Resources: Department of Family and Children Services, Merit System of Personnel Administration, Office of Secretary of State, State Board of Worker’s Compensation, Teacher’s Retirement System. Redistributive Agencies in Illinois Department of Aging, Department of Central Management Services, Department of Employment Security, Department of Healthcare and Family Services (formerly Dept of Public Aid), Department of Human Services, Department of Labor, Department of Public Health, Department of Revenue, Department of Veteran Affairs, Emergency Management Agency, Guardianship and Advocacy Commission, Illinois Council on Development Disabilities, Illinois Educational Labor Relations Board, State Retirement Systems, Department of Children and Family Services.