The theory of representative bureaucracy suggests that organizations perform better if their workforces reflect the characteristics of their constituent populations (Andrews, Boyne, Meier, O'Toole, and Walker, 2005). According to the U.S. Bureau of the Census Community Survey 2006-2008  , U.S. population is becoming more diverse in Race and Ethnicity, Education, Origins and language, etc. It indicates that the components of organizational workforces should have been increasingly heterogeneous as the development of diverse trend in total population. In reality, the demographic changes in the American workforce affected by civil rights legislation and affirmative action programs have indeed led to unprecedentedly high diversity or heterogeneity within organizations (Choi, 2008).
The increasing workforce diversity poses some of the most challenging questions for the management of public service (Riccucci, 2002). For example, diverse components of population and workforce bring employees into contact with people who may have very different racial, education, language, training, skills, functional background, cultures, and even values. As organizations increasingly operate in a multinational and multicultural context, understanding how diversity in the composition of organizational groups affects outcomes such as satisfaction, creativity, and turnover will be of increasing importance (Milliken and Martins, 1996).
Although inconsistent on the extent to which increased job satisfaction leads to improved performance (Brayfield and Crockett, 1955; Petty, McGee, and Cavender, 1984; Iaffaldano and Muchinsky, 1985), Petty, McGee, and Cavender (1984) revealed a strong relationship between job satisfaction and performance. Moreover, previous studies also demonstrated that satisfied employees are more likely to have lower absenteeism and turnover than those unsatisfied one (Locke, 1976; and Carsten and Spector, 1987). Therefore, to study employees' job satisfaction is of great significant meanings.
Nevertheless, the researches on job satisfaction either focus on its impacts on job/organization performance, absenteeism and turnover, etc. (Brayfield and Crockett, 1955; Petty, McGee, and Cavender, 1984; Iaffaldano and Muchinsky, 1985; Locke, 1976; and Carsten and Spector, 1987), or other determinants of job satisfaction, such as empowerment, participative management, quality of work life, the role of managers, individual characteristics, work environment, etc. (Bruce and Blackburn, 1992; Rainey, 1997; Kim, 2002). A relatively small literature appears to have examined the potential effects of diversity on employees' job satisfaction.
Furthermore, among a few studies seeking to the impacts of diversity on job satisfaction, most concentrate on employees on federal level (Choi, 2008), rather than in state or local government. Moreover, previous researches measure employee job satisfaction as a whole, but didn't study the impacts of diversity on employees on different facets of job satisfaction, such as the different impacts on friendliness and assistance received in work and on work itself. Besides, the dimensions most studies used for demographic diversity include race/ethnicity, education, age and sex, and disability (Wise, 2000; Andrews et al, 2005; Pitts, 2005; Choi, 2008; Pitts, 2009), but few of them use origins and language as a measure of demographic diversity. However, with the development of globalization, more and more first generation immigrants, whose first language is not English and is not born in the US., are becoming U.S. citizens and attend public sector. Â According to U.S. Census Bureau Data  released on February 22nd, 2005, the nation's foreign-born population numbered 34.2 million in 2004, accounting for 12 percent of the total U.S. population. This number is 2.3 percent higher than it was in 2003. Therefore, a study on the effects of demographic diversity of origins and language on job satisfaction is very important. In addition, most studies used static data at one time point. It couldn't reflect the development trend of demographic characteristics of public employees.
This paper aims at exploring the development trend of public employee diversity and the impacts of demographic diversity of local government on affective outcomes, job satisfaction (Figure 1), through a ten-year period. How different dimensions of demographic diversity, such as race/ethnicity, education, gender, origins, language, and education, affect employees' job satisfaction? What's the discrepancy of the impacts of demographic diversity on different aspects of job satisfaction? First, I review the previous literature on diversity focusing on five important demographic dimensions of diversity-race/ethnicity, origins, first language, gender, education, and job satisfaction. In the next section, I explore the theoretical framework and present hypotheses of diversity and its impact on different facets of job satisfaction. In the last section, I discuss research design/methodology.
Figure 1 Diversity and Job Satisfaction in Local Government
Development and Achievement
Friendliness and Assistance received
The term diversity often provokes intense emotional reactions from people who, perhaps, have come to associate the word with ideas such as "affirmative action" and "hiring quotas"; yet it is a word that simply means "variety" or a "point or respect in which things differ" (Milliken and Martins, 1996). Actually, demographic diversity refers to "the degree to which a unit (e.g., a work group or organization) is heterogeneous with respect to demographic attributes," such as race/ethnicity, sex, age, organizational tenure, and social status (Choi, 2008). This study limits its discussion to four demographic dimensions-race/ethnicity, origins, language, and gender.
Some scholars have developed several approaches to categorize various dimensions of diversity (e.g., Cummings, Zhou, and Oldham 1993; Jackson 1992; Jackson, May, and Whitney 1995; Maznevski 1994; Milliken and Martins 1996; Tsui, Egan, and O'Reilly 1992). One common distinction is between diversity on observable or readily detectable attributes such as race or ethnicity, age, or gender, and diversity with respect to less visible or underlying attributes such as education, technical abilities, functional background, tenure in the organization, or socioeconomic background, personality characteristics, or values (Cummings et al., 1993; Jackson et al., 1995; Tsui et al., 1992). According the above categorization, race/ethnicity, origins, gender, and first language should be included into observable attributes; education should be considered as underlying attributes.
Employees' satisfaction with their jobs is an important criterion for evaluation organizational effectiveness (Choi, 2008). According to Hoppock (1935: 47), job satisfaction is any combination of physiological, psychological, and environmental circumstances that causes a person truthfully to say, "I am satisfied with my job". Most common definitions are those that assume existence of needs (in varying forms) and generally view job satisfaction as resulting from the fit between these individual needs and the job and its environment (Hopkins, 1983). However, Lock (1969) rejected the concept of needs and defined job satisfaction as the pleasurable emotional state resulting from gratification or satisfaction about one's job. The interaction of one's values and one's perceptions of the job and its environment are two main causes of job satisfaction. In this study, I primarily rely on a modified need-satisfaction model, which combines individual needs and work environment.
There are two basic ways of measuring job satisfaction, facet-free and facet-specific measures. It is easy to administer and understand (Kalleberg, 1974), and it correlates highly with more complex measurements of job satisfaction to use facet-free approach (Robinson, Athanasious, and Head, 1969). In the process of responding to facet-free questions, employees tend to provide their own means of summarize a variety of different aspects of their job into a single answer (Johns, 1980). However, this approach has been severely criticized for a number of reasons (Hopkins, 1983). Most importantly, job satisfaction is unidimensional and multidimensional rather than a single dimension (Kalleberg, 1974). One general question couldn't measure all aspects of a job.
However, facet-specific job satisfaction measures ask the respondent to assess his/her satisfaction with a series of job facets. And faceted measurement is advantageous in that it coincides with the multidimensional character of job satisfaction, provides comparability across the respondents, and permits a degree of control and direction by the researcher (Hopkins, 1983). In previous researches, there are twenty-three facets studied (Hopkins, 1983). In this study, twenty-one facets will be used except salary facet and I categorize the twenty-one facets into three aspects: friendliness and assistance received in work; development and achievement; and work itself. Friendliness and assistance received in work, and development and achievement facet of job satisfaction will be used to test the research question.
Theoretical Framework and Hypotheses
Previous literature has demonstrated that different dimensions of diversity may lead to different consequences (Jackson, Joshi, and Erhardt, 2003; Jehn, Northcraft, and Neale, 1999; Milliken and Martins, 1996). Less visible or underlying attributes may provide a broad range of ideas, skills, and insights, and ultimately results in organizational performance improvement by increasing organization's creativity and problem-solving capabilities (Cox, 1993). To increase organization's creativity and problem-solving capabilities, and contribute to organization performance improvement is a crucial foundation and measure to individual development and achievement. Therefore, underlying attributes is assumed to be positively related to development and achievement facet of employees' job satisfaction.
H1: Education diverse is positively related to development and achievement facet of employee job satisfaction.
Nevertheless, observable attributes, such as demographic backgrounds, may lead to high levels of emotional conflicts between individuals in organizations. Heterogeneous groups may also experience a lack of communication or miscommunication, disabling conflict, and the burden of the high costs of coordination and resolution of conflicts (Jehn, Northcraft, and Neale, 1999).
According to Byrne's (1971) similarity-attraction theory, people prefer interacting with similar others and find interactions with them easier, positively reinforcing, and more desirable compared with interactions with others who are different. Individuals in diverse groups tend to feel less safe and to trust each other less. Lower trust is more likely to lead to higher conflicts within groups (Choi, 2008). Furthermore, for employees, who were not born as U.S. citizens and whose first language is not English, may have communication obstacles with his/her colleague. This communication problem may affect their development and promotion within the organization we well, which ultimately leads to decreased job satisfaction. Therefore, observable attributes are supposed to be negatively related to friendliness and assistance received facet of job satisfaction:
H2: Race/Ethnicity diverse is negatively related to friendliness and assistance received facet of employee job satisfaction.
H3: Origin diverse is negatively related to friendliness and assistance received facet of employee job satisfaction.
H4: Gender diverse is negatively related to friendliness and assistance received facet of employee job satisfaction.
H5: Language diverse is negatively related to friendliness and assistance received facet of employee job satisfaction.
Dependent Variable: public employees' job satisfaction is measured by two dimensions, friendliness and assistance received from work and development and achievement. It is measured on individual level based on responses to relevant survey and interview questions.
Friendliness and assistance received could be measured by the following questions:
Chance to make friends
Friendliness and helpfulness of co-workers
Enough help and equipment to get the job done
Development and achievement could be measured by the following questions:
Chance for promotions
Opportunity to develop special abilities
Independent Variable: observable and underlying attributes of diversity of public employees in local government in NJ. In recent years, three kinds of diversity measures have been used most frequently: the coefficient of variation, the Blau's index of heterogeneity, and the entropy index of diversity (Choi, 2008). In this study, the levels of each agency's demographic diversity-race/ethnicity, origins, gender, first language, and education-are calculated through the Blau's index of heterogeneity or Entropy index of diversity, which are appropriate for categorical variables, rather than the coefficient of variation, which is proper for continuous variables. The final selection between the Blau's index of heterogeneity and Entropy index of diversity depends on the data collected and data analysis results.
Race/Ethnicity is classified into five groups: White; Black or African American; Asian; Native Hawaiian and Other Pacific Islander; and Others, based on the U.S. Bureau of Census's categorization.
Origin is classified into two groups: born naturalized as a U.S. citizen and foreign born not U.S. citizen, based on the U.S. Bureau of Census's categorization.
Gender of public employee in local governments includes both male and female.
First Language is categorized into five groups: English; Spanish or Spanish Creole; Other Indo-European Languages; Asian and Pacific Island Languages; and Other Languages, based on the U.S. Bureau of Census's categorization.
Education is measured by two indicators. The first indicator is to measure the diversity in education level. The diversity in education level is categorized into six groups: less than high school graduate; high school graduate; some college or associate's degree; bachelor's degree; master's degree; Ph.D. degree. The second indicator is to measure the diversity in major. It is categorized into nine groups: social science and humanities; public affairs; library and information; business; education; law; engineering; medical science; and others.
Control Variable: there are some other variables are supposed to have significant impacts on public employees' job satisfaction, according to previous research. Those factors include employee position tenure, age, organization population, city/township population, etc.
A mixed research method will be used to examine the research question and hypotheses. I plan to collect ten-year period (2000-2009) data in employee race, gender, education, origins, first language, and job satisfaction from each local government in New Jersey, using unobtrusive resources, survey , and individual interview. The reason for collecting data in ten-year period is firstly to track the trend of the employee demographic components in local government in NJ during the past ten years, and to measure whether diversity is a manifest character of local public employee in NJ. Secondly, I want to test the change of demographic factors' impacts on job satisfaction, rather than a static effect at one time point, because it is supposed that the demographic characters of local public employees are changed to be more diverse.
Unobtrusive resources: Archive and Official Documentation to collect public employees' demographic components, including race, education, gender, origins, and first language, in each city/township government in NJ during the past ten years. Based on this data, descriptive statistics could be used to describe the development trend of employee demographic factors.
Survey and Sample: I will conduct a mail survey to collect job satisfaction data. 10% employees from each city/township government will be randomly selected and the questionnaire will be sent to them. In order to assure a good respondent rate, I will conduct a three-round survey. The follow-up survey may be conducted by phone or email.
Semi-Structured Individual Interview: in-depth data is needed to unearth some deep information. The main content of individual interview is to investigate employees' job satisfaction and how diverse working environment affect their job satisfaction. What are the most important factors that impact their job satisfaction? What are the employees' perceptions or suggestion regarding to diversity management?
Above all, content analysis will be used to examine written documents. Content analysis is "any technique for making inferences by systematically and objectively identifying special characteristics of messages" (Holsti, 1968). First, is to decide the levels and units of documents analysis. Sampling may occur at any or all of the following levels: words, phrases, sentences, paragraphs, sections, chapters, and books. Second, is to combine content analytic elements, and coding. Coding frames are used to organize the data and identify findings after open coding has been completed. The whole process is to divide collected archive and documentation data and fill them into five independent variables.
For some quantitative data, job satisfaction survey data are Likert Scale data, 5 scales. Recode each factor to make sure every question and data is in the same direction from 1 (strongly dissatisfaction) to 5 (strongly satisfaction). Mean score higher than 3 is considered satisfaction, and lower than 2 is considered dissatisfaction.
Furthermore, descriptive statistics is used to calculate mean, media, standard deviation, standard error of collected data, including documentation, survey, and interview data. For The documentation/archive data is a pooled data in ten years. To compare data in different year, we can conclude whether the component of employee in local government is more and more diverse. What's the development trend, to speed up or slow down?
If it is necessary, Cronbach's Alpha can be used to test measurement reliability before creating index variables. Multiple regression could be used to test the significance level and coefficient of correlation between demographic diversity and employees' job satisfaction. Meanwhile, Variance inflation factor (VIF) can be used to test whether there is multicollinearity problem and Breusch-Pagan test is used to examine whether the model has heteroskedasticity problem or not.
Some software programs may be useful to analyze data, for example SPSS, Office Excel, and Stata. The analyzing software should including the following functions: word processors, text retrievers, textbase managers, code-and-retrieve programs, code-based theory builders, and conceptual network-builders.