Content uploaded by Steve Bradley
Author content
All content in this area was uploaded by Steve Bradley
Content may be subject to copyright.
Job autonomy and job satisfaction: new evidence
Anh Ngoc Nguyen, Jim Taylor and Steve Bradley*
Department of Economics
Management School
Lancaster University
Lancaster, England
a.nguyen@lancaster.ac.uk
jim.taylor@lancaster.acuk
s.bradley@lancaster.ac.uk
August 2003
Abstract: This paper investigates the impact of perceived job autonomy on job satisfaction.
We use the fifth sweep of the National Educational Longitudinal Study (1988-2000), which
contains personally reported job satisfaction data for a sample of individuals eight years after
the end of compulsory education. After controlling for a wide range of personal and job-
related variables, perceived job autonomy is found to be a highly significant determinant of
five separate domains of job satisfaction (pay, fringe benefits, promotion prospects, job
security and importance / challenge of work).
JOB AUTONOMY SATISFACTION PAY GENDER PROMOTION
JEL classification: I31, J28
* Anh Nguyen would like to thank the ESRC for financial support under its Post-doctoral Fellowship
Programme. Communications to be addressed to Professor Jim Taylor, Management School,
Department of Economics, Lancaster University, Lancaster, LA1 4YX, England.
1
1. Introduction
Traditionally, economists have shied away from investigating well-being because of its highly
subjective nature and have considered 'personal judgements of satisfaction and other
subjective opinions as a black box that should be opened only by psychologists and
sociologists' (Levy-Garboura and Montmarquette 1997, p.1). The main concern has been that
no two people will use the same scale to answer questions about their well-being.
1
Nevertheless, the analysis of subjective well-being has become a topic of increasing interest
among economists and is now regarded as something worthy of empirical investigation (Clark
and Oswald 1994; Blanchflower and Oswald 1999; McBride 2001; Frey and Stutzer 2002a,
2002b; van Praag et al. 2003).
2
The driving force behind this increasing interest in personal well-being is the growth in the
number of large-scale labour market surveys that include questions about how much workers
are satisfied with their job. Several studies have consequently attempted to identify and
measure the determinants of job satisfaction (Clark and Oswald 1996; Hamermesh 1977,
2001; Borjas 1979; Sousa-Poza and Sousa-Poza 2000). As Freeman (1978) has pointed out,
‘the answers to questions about how people feel toward their job are not meaningless but
rather convey useful information about economic life that should not be ignored’ (p. 135).
There are several compelling reasons why economists should care about job satisfaction.
First, job satisfaction has been found to be a strong predictor of a worker's behaviour and
performance. For example, reported job satisfaction has been used to predict separations, quits
and labour productivity (e.g. Hamermesh 1977; Freeman 1978; Akerlof et al. 1988; Clark et
al. 1997; Clark 2001; Shields and Price 2002; Levy-Garboura et al. 2001; Tsang et al. 1991).
3
Secondly, job satisfaction is an important predictor of overall well-being (Argyle 1989; Clark
1997; Sousa-Poza and Sousa-Poza 2001; van Praag et al. 2003). If the answers by individual
workers to job satisfaction questions only contained white noise, it is unlikely that such
correlations would have been found (Clark 1997).
1
Layard (2003) reports evidence from research in neuro-science that feelings (e.g. happiness) can be measured
accurately through brain activity and that happiness can be compared between people.
2
See Frey and Stutzer (2002) for an extensive account of economic research on happiness and on why
economists should and could use subjective data on human happiness in general and job satisfaction in
particular.
3
It should be noted that some earlier studies have found only a low correlation between job satisfaction and
worker performance (Iaffaldano and Muchinski 1985).
2
The determination of job satisfaction has therefore become a focus of numerous recent
studies. Previous studies have explained job satisfaction as dependent on a number of factors
such as gender (Clark 1997; Galdeano 2001), own wage or income, relative wages (Clark and
Oswald 1996; Hamermesh 1977, 2001), union activity (Borjas 1977) and mismatches between
education and skill (Allen and van der Velden 2001).
Among other factors believed to influence job satisfaction is job autonomy. More autonomy is
expected to be associated with greater job satisfaction because workers have more freedom to
determine their own effort and work schedule. Previous research in this area has been
confined to the disciplines of psychology and sociology, and has been either qualitative in
nature or relies on small, unrepresentative, samples of respondents (Anderson et al. 1992;
Bhuiman et al. 1996; Birdseys and Hill, 1995; Landersweerd and Bousmans 1994;
Schienman, 2002). Much of this research also ignores the issue of ‘how much’ job autonomy
increases job satisfaction. Furthermore, very few of the studies that investigate the impact of
job autonomy on job satisfaction fail to control for other determinants of job satisfaction, such
as personal characteristics. The economics literature on job satisfaction has also ignored the
multi-dimensional nature of job satisfaction, focusing instead on overall satisfaction with a
job.
In this paper we intend to fill these gaps in previous research. We investigate the effect of
different levels of job autonomy on several dimensions (i.e. domains) of job satisfaction using
a large-scale survey of young adults in the US. Specifically, our data refer to the fourth
follow-up of the National Educational Longitudinal Study conducted in 2000. Respondents to
this survey are observed eight years after the end of high school, and those in employment at
the time of the survey were asked not only whether they were satisfied with different aspects
of their job but also how much freedom they had in deciding how their job should be done.
Together with a wide range of other variables, this allows us to investigate the impact of job
autonomy on job satisfaction.
The remainder of the paper is structured as follows. Section 2 provides a review of the
determinants of job satisfaction. A framework of analysis is set out in section 3. Section 4
describes the data and variables. The results of the empirical analysis are presented in section
5. Section 6 concludes.
2. Previous studies of job satisfaction
3
Previous studies have attempted to explain a worker’s job satisfaction as a function of the
individual's personal characteristics and the characteristics of the job itself. Variables such as
age, gender, education, marital status, hours of work and earnings figure prominently in these
previous studies.
One of the main findings is that women are more satisfied with their jobs than men, even after
taking into account many observed characteristics and sample selectivity (Clark 1996, 1997;
Groot and Brink 1999; Sanz de Galdeano 2001; Blanchflower and Oswald 2001).
4
Clark
(1996, 1997) and Galdeano (2001) explain the existence of a positive relationship between
being a female and job satisfaction as reflecting women's lower expectations from their job,
which arise from the poor position in the labour market that women have traditionally held
(Clark 1997).
The observed relationship between job satisfaction and age suggests the existence of a U-
shaped relationship, which is captured by a quadratic term in age in the regression equation
(Clark et al. 1996; Sloane and Ward 2001; Blanchflower and Oswald 2001).
5
Marital status is
also believed to influence job satisfaction, married individuals being more likely to report a
higher level of job satisfaction (Blanchflower and Oswald 2001; Clark 1997).
Previous research also suggests that higher levels of education are associated with lower
levels of job satisfaction (Clark 1997; Clark and Oswald 1996; Sloane and William 1996).
One explanation of this result is that job satisfaction depends on the gap between outcomes
and aspirations and that aspirations increase with the level of education. Individuals with a
higher level of education consequently tend to be less satisfied with their job because they
have higher expectations than those with lower levels of education.
Other variables correlated with job satisfaction include earnings, hours of work, job tenure,
union membership, size of establishment, and self-employment status (Freeman 1978; Borjas
1979;
Clark and Oswald 1996; Clark 1997; Belfield and Harris 2002; Shields and Price 2002).
An interesting result from these earlier studies relates to the estimated effect of a person's
earnings on their job satisfaction. Although, theoretically, income is believed to influence
4
Ward and Sloane (1999) find for the UK academic profession that there were no significant differences
between males and females regarding job satisfaction. Moguerou (2002) finds females are less satisfied with
their job than males. Clark (1997) notes that this difference must diminish over time as the position of females in
the labour market improves.
5
But Shields and Price (2002) find that job satisfaction increases progressively with age.
4
individual worker’s job satisfaction,
6
empirical evidence testing this hypothesis gives mixed
results. Clark (1997) and Shields and Price (2002) report that income is important for
worker’s "satisfaction with pay” and for “overall job satisfaction”. On the contrary, Clark and
Oswald (1996) find that a worker’s reported level of well-being is at best weakly correlated
with their income. Similarly, Belfield and Harris (2002) find no evidence that job satisfaction
depends on income among those working in higher education. Some studies argue that it is
not own income, but relative income, that is important. The idea that job satisfaction is
dependent on relative income has been suggested and tested by Hamermesh (1977, 2001),
Clark and Oswald (1996), Neumark and Postlewaite (1995), Sloane and Ward (2001) and
Shields and Price (2002). Most studies have found some effect of relative income on job
satisfaction, though the effect is generally rather small.
3. Job satisfaction: an analytical framework
In a recent study by van Praag et al. (2003), an individual’s general satisfaction is
hypothesised to depend upon several individual domain satisfactions, which include work,
home, wealth, leisure and the environment. Each of these domain satisfactions in turn depend
upon a set of explanatory variables. The utility derived from having a job can be regarded as
one of several sub-utility functions that together determine an individual's general utility
(Clark and Oswald 1996).
7
In a similar vein, we argue here that job satisfaction is also likely
to be multi-dimensional. Overall job satisfaction (S) is determined by satisfaction in several
job domains (S
k
) such as pay, job security, promotion prospects, fringe benefits and the
importance attached to the job: S = g(S
k
). Previous studies have focused almost exclusively on
a single aggregate measure of job satisfaction and have therefore ignored its multi-
dimensional nature.
8
It is of interest to see how the impact of the determinants varies between
the different job domain satisfactions.
We illustrate this idea in Figure 1. An individual’s personal background and job
characteristics affect overall job satisfaction through the various domain satisfactions. We
assume here that all individual domain satisfactions are determined by the same set of
explanatory variables: S
k
= f(i, j), where i refers to personal characteristics and j refers to the
characteristics of the job itself. Personal characteristics include factors such as gender, age,
6
A robust and general finding is that richer people on average report higher general subjective well-being (Frey
and Stutzer 2002).
7
Van Praag et al. (2002) represent the subjective well-being as a general utility function which includes
satisfaction over a number of domains such as work, home, health, wealth, leisure and the environment.
5
ethnicity and educational attainment. Job characteristics include variables such as earnings,
hours worked, skill level, occupation and the industry in which a person is employed.
A further variable that may be expected to influence job satisfaction, and its various domains,
is the degree of perceived autonomy that workers enjoy in the way they do their job. It is
expected that a higher degree of job autonomy will lead to greater satisfaction. Several studies
that have investigated the influence of perceived job autonomy on job satisfaction are
primarily qualitative in nature and fall within the field of business research (Bhuiman et al.
1996; Birdseys and Hill 1995), sociology (Schienman 2002) and psychology (Anderson et al.
1992; Landersweerd and Bousmans 1994; Weaver 1977). Most of these studies are
descriptive in content, employ small and unrepresentative samples and often do not control
for other variables such as personal characteristics. For example, Bhuiman et al. focus on the
influence of job autonomy on the job satisfaction of ex-patriots in Saudi-Arabia while
Landerweerd and Bousmans focus on the influence of job autonomy on the job satisfaction of
nurses. This research in the field of organisational behaviour suggests a positive relationship
between job autonomy and job satisfaction.
The next section describes the data and the variables included in the model and the statistical
methodology used to estimate the impact of job autonomy on job satisfaction.
4. Data, variables and model
We use data from the National Educational Longitudinal Study (NELS:2000). The study
began in 1988 with a cross-sectional survey of eighth graders, and continued with four follow-
up interviews in 1990, 1992, 1994 and 2000. The first three follow-ups provide detailed
information about a respondent's family background, academic record and their activities
before, during and after high school. The 2000 follow-up survey provides detailed
information about their labour market activities eight years after the end of compulsory
education. The sample selected for the present study includes only those employees who were
in a full-time job in 2000. Part-time workers are excluded since we wish to focus here on the
attitudes to work only of those whose primary activity is working for pay. Self-employed
persons are also excluded since our measure of job autonomy is relevant only for employees.
Modelling job satisfaction
8
An exception is Clark and Oswald (1996), who examine overall job satisfaction and satisfaction with pay.
6
Several individual domains of job satisfaction are identified in the National Educational
Longitudinal Study. These include satisfaction with pay, fringe benefits, promotion prospects,
job security and importance / challenge of work. The correlation matrix for these five job
satisfaction domains (Table 1) indicates that although the individual domains are significantly
positively correlated with each other, the correlation coefficients are low. The logit regression
reported in Table 2, however, indicates that overall job satisfaction is highly significantly
related to all five individual domains of job satisfaction. This result is consistent with the view
that overall job satisfaction is a multi-dimensional construct.
In order to estimate the influence of perceived job autonomy on job satisfaction, it is
necessary to control for the personal characteristics of each respondent as well as for the
characteristics of the job itself that are likely to influence job satisfaction. Empirically,
satisfaction with pay can be described by the following latent variable model:
S* = x
i
’
β
+
ε
i
where S* is a latent variable that is assumed to be linearly related to the vector of explanatory
variables, x
i ,
which influence an individual’s utility from being in a job. In our data, job
satisfaction is described as a binomial response variable, indicating whether individuals are
satisfied (S=1) or dissatisfied (S=0) with their job.
9
We therefore estimate a binomial logit
model and report the marginal effects for ease of interpretation (Green 1997).
Explanatory variables
Two groups of explanatory variables are identified: (1) those relating to the personal
characteristics of the respondent, such as age, gender, race, marital status and number of
children; (2) those relating to the job itself, such as job autonomy, income earned and hours
worked. Although all respondents were in 8
th
grade in the first sweep of the NELS in 1988,
their year of birth varies from 1972 to 1975. Race distinguishes between white, black,
Hispanic (non-black), Asian and American Indian.
10
Our primary aim in this paper is to estimate the extent to which the degree of job autonomy
influences various dimensions of job satisfaction. The variable of interest is the degree of
perceived autonomy that workers enjoy in the way they do their job. Four levels of job
9
The respondents were asked to answer the question: “Overall, would you say you are satisfied or dissatisfied with
your job as a whole?”
10
It should be noted that the sample size of American Indians is small (1.2% of the respondents).
7
autonomy are identified in the NELS data: zero autonomy, limited autonomy (a worker is told
what to do but has some control over how to do it), some freedom in deciding what to do, and
virtually complete autonomy (i.e. basically one's own boss). Zero autonomy is included in the
base category so that the effect of the various degrees of autonomy on overall satisfaction and
its domains can be estimated. More autonomy is expected to be associated with greater
satisfaction simply because workers have freedom to determine their own effort and work
schedule. Table 3 provides some prima facie evidence that a worker's satisfaction with
various aspects of their job increases as perceived job autonomy increases.
11
The greatest
increase in job satisfaction, however, occurs at the low end of the job autonomy scale.
Increasing job autonomy from 'no freedom in job' to 'limited freedom in the way a job is
done’ is associated with by far the biggest increase in job satisfaction between the various
categories. We investigate this further in the empirical analysis below.
Two other job-related variables that are included in our analysis of job satisfaction are
earnings and hours worked. We follow previous studies by using annual earnings (from the
respondent's main job) rather than the earnings for the most recent week or month in order to
reduce the problem of measurement error (Bound et al. 1999; Hamermesh 1999). For hours
worked, we use the average number of hours worked per week as stated by the respondent.
Other job-related variables included in the regression model are occupational status and the
industry in which the respondent works. Occupational status distinguishes between
professional, managerial, skilled non-manual, skilled manual and unskilled / semi-skilled
workers. The latter is used as a characteristic of the base group. The industry variable is
divided into ten main industry groups.
5. Empirical results
This section presents two sets of results. The first set investigates the determinants of job
satisfaction as a whole and provides the results for females and males separately as well as for
the total sample. The second set provides the results for each of the five domains of job
satisfaction referred to above. Although the primary focus of the empirical investigation is to
estimate the impact of perceived job autonomy on job satisfaction, the influence of other job-
11
We recognise that there may be a two-way relationship between job satisfaction and job autonomy in so far as
workers who are happy in their job are more likely to perform well and consequently get promoted to jobs with
greater autonomy. Since the NELS database does not contain sufficiently detailed information about a person's
job history, we cannot attempt to control for the potential endogeneity of job autonomy here.
8
related factors and personal factors on job satisfaction is also discussed. The estimated
regression equations include several controls that are in general statistically insignificant and
these results are not reported (see note to Tables).
From the results presented in Table 4, it is clear that the degree of job autonomy is highly
statistically related to overall job satisfaction. As job autonomy increases from 'no freedom' to
'limited freedom', for example, the probability of a worker being satisfied with his or her job
increases by 0.13 for females and 0.11 for males. And as the degree of freedom increases
from 'no freedom' to 'basically one's own boss', the probability of being satisfied with one's
job increases by 0.24 for females and 0.17 for males. It is also interesting to note that the
differential between the different threshold measures of job autonomy increases almost
linearly for females but the rate of increase begins to fall for males at higher levels of
autonomy. In sum, even when a wide range of other job-related and personal factors are taken
into account, the degree of autonomy that workers have in their job has a substantial impact
on their overall job satisfaction, and there are differences between males and females.
Two other job-related variables are significantly related to job satisfaction. As expected,
current income has a positive effect on overall job satisfaction, with the estimated impact of
income being nearly twice as great for males as for females. Being employed in a professional
occupation also increases the probability of being satisfied with one's job (by 0.11 for females
and 0.05 for males compared to the base group of unskilled and semi-skilled workers). Rather
surprisingly, the number of hours worked is not correlated with overall job satisfaction.
Job satisfaction does not appear to be related to personal factors in most cases. There are,
however, two exceptions. Black workers have a lower probability of being satisfied with their
job as a whole than white workers (whereas other ethnic groups are not significantly different
to white workers); and there is a some evidence that single workers are less satisfied with
their job as a whole than married workers. Taken together, being black and single therefore
has a substantial negative impact on overall job satisfaction.
Although there is clear evidence that the degree of job autonomy has a substantial and highly
significant influence on job satisfaction, further analysis reveals that the magnitude of this
impact varies between different aspects of a worker's job. Table 5 shows the regression results
for satisfaction with pay, fringe benefits, promotion prospects, job security and importance /
9
challenge of job. The main result is that the degree of job autonomy is highly significantly
related (positively) to all five aspects of job satisfaction.
The estimated impact of job autonomy on job satisfaction varies considerably, however,
between these five aspects of job satisfaction. The estimated impact is much greater for pay,
promotion prospects and importance / challenge of job than for fringe benefits and job
security. The most striking result is that the probability of being satisfied with one's
promotion prospects increases by 0.24 as job autonomy changes from 'no freedom in job' to
'basically one's own boss.' The regressions estimated separately for females and males
(Tables 6 and 7) indicate that the estimated impact of job autonomy is similar for the two
groups, though there is some indication that the estimated impact of job autonomy on job
satisfaction is slightly greater for females than for males, especially with respect to
importance / challenge of work.
Two other results relating to the characteristics of the job are of interest. First, current income
is significantly positively related to four aspects of job satisfaction, with the greatest impact
being on satisfaction with pay as would be expected. Current income is not statistically
related, however, to importance / challenge of work. Hours worked is also significantly
related (negatively) to satisfaction with pay and the estimated coefficient (a semi-elasticity) is
about the same magnitude (0.20) as the estimated coefficient on earnings (but with the
opposite sign). Very similar results are obtained for females and males with respect to these
two variables. This suggests that an equal proportionate change in earnings and hours at the
mean value of these two variables would leave workers at the same level of satisfaction.
Black workers are consistently less satisfied with all aspects of their job than white workers,
though their dissatisfaction over pay is much greater than their dissatisfaction over other
aspects of their job. This is consistent with the view that black workers perceive that they are
discriminated against in the labour market, especially with respect to pay. The probability of
being satisfied with their pay, for example, is 0.13 less for black workers than for white
workers. Evidence that other ethnic minorities are less satisfied with various aspects of their
job than white workers is generally rather weak. Very similar results are obtained for ethnic
minority groups when the regressions are run for males and females separately.
Finally, single persons are invariably less satisfied with all aspects of their job than married
workers. Similar results are obtained for females and males separately. There is also some
10
evidence that divorcees are less satisfied with certain aspects of their job (particularly for pay
and fringe benefits) than married workers, though the results are statistically much stronger
for females than for males.
6. Conclusion
Job satisfaction is a topic of considerable interest to employers since it is likely to influence a
worker's, and hence the firm's, performance. Previous studies suggest that firms are likely to
benefit through lower job turnover and higher productivity if their workers have a high level
of job satisfaction. It is also important for workers to be happy in their work, given the
amount of time they have to devote to it throughout their working lives.
In view of the potential importance of job satisfaction to both the employer and the employee,
this paper has investigated the determinants of several domain satisfactions using data from
the National Educational Longitudinal Study. This survey, which traces the school-to-work
transition of a representative sample of youths from eighth grade through the following twelve
years, provides an opportunity to investigate aspects of job satisfaction not previously
possible. Specifically, it allows us to investigate the potential impact of a range of job-related
and personal factors on five different domains of job satisfaction for a large and nationally
representative sample of workers.
The focus of the present paper, however, has been on a hitherto under-researched aspect of
job satisfaction, namely the impact of job autonomy on several different domains of job
satisfaction. Since a key characteristic of job satisfaction is its multi-dimensional nature, we
have investigated the extent to which job autonomy is related to satisfaction with pay, fringe
benefits, promotion prospects, job security and importance / challenge of job. The main
finding is that the degree of job autonomy is significantly related to all five aspects of job
satisfaction. As a worker’s control over how a job is done increases, the level of job
satisfaction also increases. The increase in job satisfaction between ‘no freedom in job’ and a
‘small amount of freedom in job’ is particularly striking. This is especially the case for
satisfaction with pay, promotion prospects and importance / challenge of job.
A further result of interest is that there is some evidence that job autonomy has a greater
impact on all domains of job satisfaction for females than for males. This is especially the
case for the impact of job autonomy on the importance / challenge of job. Female workers
11
with a high level of job autonomy have a significantly higher level of satisfaction with respect
to the importance / challenge of the job than is the case for males.
Finally, we find that most personal characteristics have little effect on the various domains of
job satisfaction. Two notable exceptions are that black workers and single workers are
substantially and significantly more dissatisfied with various aspects of their job compared to
the base group. Black and single workers, for example, are substantially more dissatisfied
with their pay than white workers.
12
References
Akerlof. G., Rose, A., Yellen, J., 1988. Job Switching and Job Satisfaction in the U.S. Labor
Market, Brookings Papers on Economic Activity 2, 495-582.
Allen, J., van der Velden, R., 2001. Educational mismatches versus skill mismatches: Effects
on wages, job satisfaction and on-the-job search, Oxford Economic Papers 53, 434-452.
Anderson, L., Tolson, J., Filed, M., Thacker, J., 1992. Job autonomy as a moderator of the
Pelz effect, Journal of Social Psychology 130, 707-708.
Argyle, M., 1989. The Social Psychology of Work, 2nd edition. Harmondsworth: Penguin.
Belfield, C., Harris, R., 2002. How well do theories of job matching explain variation in job
satisfaction across educational level? Evidence for UK graduates, Applied Economics
34, 535-0548.
Bhuian, S., Al-Shammari, E., Jefri, O., 1996. Organizational commitment, job satisfaction
and job characteristics: An empirical study of expatriates in Saudi Arabia, International
Journal of Commerce & Management 6, 57-80.
Birdseys, M., Hill, J., 1995. Individual, organizational/work and environmental influence on
expatriate turnover tendencies: An empirical study, Journal of International Business,
Fourth Quarter, 787-813.
Blanchflower, D.,,Oswald, A., 2000. The rising well-being of the young, in Youth
Employment and Joblessness in Advanced Countries, edited by D. Blanchflower and R.
Freeman. Chicago: The University of Chicago Press.
Blanchflower, D., Oswald, A., 1999. Well-being, insecurity and the decline of American job
satisfaction, mimeo, Dartmouth College, USA.
Borjas, G., 1979. Job satisfaction, wages and union, Human Resources XIV, 21-40.
Booth, A., Francesconi, M., Frank, J., 2002. Temporary jobs: Stepping stones or dead ends?
Economic Journal 112, F189-F213.
Clark, A. E., 1997. Job satisfaction and gender: Why are women so happy at work? Labour
Economics 4, 341-372.
Clark A. E., 1999. Are wages habit-forming? Evidence from micro data, Journal of Economic
Behavior and Organization 39, 179-200.
Clark, A. E., 2001. What really matters in a job? Hedonic measurement using quit data,
Labour Economics 8, 223-242.
Clark A. E., Oswald, A. J., 1994. Unhappiness and unemployment, Economic Journal 104,
May, 648-659.
13
Clark A. E., Oswald A. J., 1996. Satisfaction and comparison income, Journal of Public
Economics 61, 359-381.
Clark, A., Oswald, A., Warr, P., 1996. Is job satisfaction U-shaped in age? Journal of
Occupational and Organizational Psychology 69, 57-81.
Clark, A., Geogellis, Y., Sanfey P., 1997. Job satisfaction, wage changes and quits: Evidence
from Germany, mimeo, University of Orleans.
Easterlin, R. A., 2001. Income and happiness: Towards a unified theory, Economic Journal
111, July, 465-484.
Freeman, R. N., 1978. Job satisfaction as an economic variable, American Economic Review
68, 2, 135-141.
Frey, B., Stutzer A., 2002a. Happiness and Economics. Princeton and Oxford: Oxford
University Press.
Frey, B., Stutzer A., 2002b. What can economists learn from happiness research? Journal of
Economic Literature, forthcoming.
Galdeano, A., 2001. Gender differences in job satisfaction and labour market participation:
UK evidence from propensity score, mimeo, European University Institute, Florence.
Groot, W., van den Brink, H., 1999. Job satisfaction and preference drift, Economic Letters
63, 363-367.
Hamermesh, D., 1977. Economic aspects of job satisfaction, in Essays in Labor Market
Analysis, edited by Ashenfelter O and Oates W, Toronto: John Wiley&Son.
Hamermesh, D., 1999. Changing inequality in markets for work place amenities. Quarterly
Journal of Economics CXIV, 1085-1123.
Hamermesh, D., 2001. The changing distribution of job satisfaction, Human Resources 36
(1), 1-30.
Iaffaldano, M. T., Muchinski, P. M., 1985. Job satisfaction and job performance: a meta-
analysis, Psychological Bulletin 97 (2), 251-273.
Iverson, R., 2000. An event history analysis of employee turnover: The case of Hospital
employees in Australia, Human Research Management Review 9, 397-418.
Landerweerd, J., Bousmans, N., 1994. The effect of work dimensions and need for autonomy
on nurses’ work satisfaction and health, Journal of Occupational and Organizational
Psychology 67, 207-217.
Layard, R., 2003. Happiness – Has social science a clue? Centre for Economic Performance,
Lionel Robbins Memorial Lectures 2002/3, London School of Economics.
14
Levy-Garboua, L., Montmarquette, C., 1997. Reported job satisfaction: What does it mean?,
Cirano: Canada.
Levy-Garboua, L., Montmarquette, C., Simonnet. V., 2001. Job satisfaction and quits: Theory
and evidence from the German socioeconomic panel, Cirano, Canada.
Lydon, R., Chevalier, A., 2002. Estimates of the effect of wages on job satisfaction, Centre
for Economic Performance, LSE.
McBride, M., 2001. Relative-income effects on subjective well-being in the cross-section,
Journal of Economic Behavior and Organization 45, 251-278.
Moguerou, P., 2002. Job satisfaction among US Ph.D. graduates: the effects of gender and
employment sector, mimeo, Universite de Bourgogne.
Neumark, D., Postlewaite, A., 1995. Relative income concerns and the rise in married
women’s employment, mimeo, Journal of Public Economics 70, 157-183.
Phelps, C., 2001. A clue to the paradox of happiness, Journal of Economic Behavior &
Organization 45, 293-300.
Schienman, S., (2002), Socio-economic status, job conditions, and well-being: Self-concept
explanations for gender-contingent effects, The Sociological Quarterly 43, 627-646.
Shields, M., Price, S., 2002. Racial harassment, job satisfaction and intentions to quit:
Evidence from the British nursing profession, Economica 69, 295-362.
Sloane, P., Ward, M., 2001. Cohort effects and job satisfaction of academics, Applied
Economics Letters 8, 787-791.
Sloane, P., William, H., 2000. Job satisfaction, comparison earnings and gender, Labour 14,
473-501.
Sousa-Poza, A., Sousa-Poza, A., 2000. Well-being at work: a cross-national analysis of the
levels and determinants of job satisfaction, Journal of Socio-economics 29, 517-538.
Stutzer, A., 2002. The role of income aspirations in individual happiness, mimeo, School of
Law, University of California at Berkeley.
Tsang, M. C., Rumberger, R. W., Levin, M., 1991. The impact of surplus schooling on
worker productivity, Industrial Relations, 209-228.
van Praag, B.M.S., Frijter, P., Ferrer-Carbonell, A., 2003. The anatomy of subjective well-
being, Journal of Economic Behavior and Organization, forthcoming 51, 29-49.
Ward, M., Sloane, P., 1999. Job satisfaction within the Scottish academic profession,
Discussion Paper No. 39, Institute for the Study of Labour (IZA), Bonn.
15
Weaver, C., 1977. Relationship among pay, race, sex, occupational prestige, supervison,
work autonomy and job satisfaction in a national sample, Personnel Psychology 30,
437-445.
16
Table 1. Correlation matrix
Satisfaction
with job as a
whole
Satisfaction
with pay
Satisfaction
with fringe
benefit
Satisfaction
with
promotion
Satisfaction
with job
security
Satisfaction with pay (yes=73%) 0.374
Satisfaction with fringe benefit (yes=81%) 0.347 0.277
Satisfaction with promotion (yes=74%) 0.457 0.308 0.281
Satisfaction with job security (yes=91%) 0.322 0.171 0.277 0.260
Satisfaction with importance of job (yes=86%) 0.507 0.230 0.222 0.364 0.210
Note: Respondents were asked to indicate whether they are/were satisfied with their current/most recent job
with respect to several aspects of the job, including pay, fringe benefits, importance and challenge of work,
opportunities for promotion and job security. They were then asked: “Overall, would you say you are satisfied
or dissatisfied with your job as a whole?” 88% answered 'yes' to this question.
Source: National Educational Longitudinal Study (NELS), 1988/2000.
17
Table 2. Logit estimates of overall job satisfaction
Explanatory variables Dependent variable = overall
job satisfaction
Marginal
effects
Standard
error
Constant -0.114 0.007
Satisfaction with pay 0.060*** 0.004
Satisfaction with fringe benefit 0.041*** 0.004
Satisfaction with promotion opportunity 0.067*** 0.004
Satisfaction with job security 0.042*** 0.005
Satisfaction with work importance and challenge 0.085*** 0.005
Log-likelihood
-1828
Chi-squared
495
Number of observations
9045
Note: ( ) = standard error; * = significant at 0.05; ** = significant at 0.01;
*** = significant at 0.001.
Source: National Educational Longitudinal Study (NELS), 1988/2000.
18
Table 3. Job satisfaction and job autonomy: % satisfied with particular aspects of their job
Job autonomy
Aspect of job No freedom
in job
Small amount
of freedom
Some
freedom
Much
freedom
Pay 55.0 70.0 73.7 80.8
Fringe benefits 65.3 76.6 81.0 78.4
Promotion opportunity 49.9 69.6 75.8 81.2
Job security 77.2 888.3 91.8 91.9
Importance and challenge of work 63.3 80.4 87.4 92.2
Job as a whole 60.0 84.5 90.3 94.2
Notes:
Respondents employed at any time between January 1994 and the survey date in 2000 were asked to
answer the following questions:
1. Considering your current/most recent job, would you say that you are/were satisfied or dissatisfied
with:
a. your pay?
b. fringe benefits?
c. opportunities for promotion and advancement?
d. importance and challenge of your work?
e. job security?
f. your job as a whole?
2. Which one of the following four statements best describes your job?:
a. someone else decides what you do and how you do it
b. someone else decides what you do, but you decide how to do it
c. you have some freedom in deciding what you do and how to do it
d. you are basically your own boss.
Source: National Educational Longitudinal Study (NELS), 1988/2000.
19
Table 4. Overall job satisfaction: estimated marginal effects for full-time workers
Explanatory variables Satisfied with job as a whole
Females Males Total
Job characteristics
Job autonomy: small freedom 0.128***
(0.017)
0.112***
(0.016)
0.120***
(0.012)
Job autonomy: some freedom 0.180***
(0.016)
0.146***
(0.0145)
0.165***
(0.011)
Job autonomy: basically own boss 0.244***
(0.024)
0.170***
(0.020)
0.207***
(0.015)
Current income (log) 0.036***
(0.011)
0.067***
(0.012)
0.049***
(0.008)
Hours worked (log) -0.025
(0.019)
-0.031
(0.020)
-0.027
(0.014)
Occupation: professional 0.106***
(0.024)
0.048**
(0.019)
0.076***
(0.015)
Occupation: managerial 0.050*
(0.023)
0.007
(0.019)
0.025
(0.014)
Occupation: non-manual 0.051**
(0.019)
0.013
(0.016)
0.031**
(0.012)
Occupation: skilled manual 0.048
(0.029)
-0.002
(0.015)
0.013
(0.014)
Personal characteristics
Male
0.012
(0.008)
Black -0.076***
(0.017)
-0.058***
(0.016)
-0.070***
(0.012)
Asian -0.032
(0.022)
-0.015
(0.021)
-0.024
(0.015)
Hispanic -0.018
(0.017)
-0.007
(0.015)
-0.013
(0.012)
American Indian -0.073
(0.044)
0.012
(0.043)
-0.035
(0.031)
Single -0.018
(0.012)
-0.026*
(0.013)
-0.024**
(0.009)
Divorced or widowed -0.027
(0.022)
-0.009
(0.023)
-0.020
(0.016)
Constant -0.335
(0.090)
-0.526
(0.101)
-0.329
(0.058)
Log-likelihood -1270 -1211 -2507
Chi-squared 326 296 580
Number of observations 3494 3576 7070
Note: The base group includes persons with the following characteristics: no freedom in job, unskilled
or semi-skilled, white, born in 1974, married, no children. Dummy variables were also included for year
of birth, number of children, highest level of educational attainment, industry worked in and region of
residence (results not reported here). ( ) = robust standard error; * = significant at 0.05; ** = significant
at 0.01; *** = significant at 0.001.
20
Table 5. Job satisfaction: estimated marginal effects for full-time workers
Explanatory variables Domains of job satisfaction
Pay Fringe
benefits
Promotion
prospects
Job security Importance /
challenge of
work
Job characteristics
Job autonomy: small freedom 0.110***
(0.020)
0.073***
(0.018)
0.150***
(0.020)
0.050***
(0.010)
0.098***
(0.014)
Job autonomy: some freedom 0.135***
(0.019)
0.095***
(0.017)
0.196***
(0.019)
0.072***
(0.010)
0.155***
(0.013)
Job autonomy: basically own boss 0.171***
(0.023)
0.109***
(0.021)
0.239***
(0.024)
0.082***
(0.013)
0.216***
(0.018)
Current income (log) 0.211***
(0.014)
0.098***
(0.011)
0.071***
(0.011)
0.023***
(0.007)
0.014
(0.008)
Hours worked (log) -0.200***
(0.022)
0.037
(0.020)
0.018
(0.021)
0.023*
(0.011)
0.046***
(0.013)
Occupation: professional -0.041
(0.022)
0.021
(0.019)
0.086***
(0.022)
0.018
(0.013)
0.125***
(0.017)
Occupation: managerial -0.008
(0.022)
0.039*
(0.019)
0.093***
(0.022)
0.040**
(0.013)
0.037*
(0.016)
Occupation: non-manual -0.011
(0.020)
0.046**
(0.016)
0.070***
(0.019)
0.010
(0.010)
0.016
(0.013)
Occupation: skilled manual -0.041
(0.022)
0.016
(0.018)
0.046
(0.021)
0.004
(0.012)
0.034*
(0.016)
Personal characteristics
Male 0.015
(0.012)
-0.015
(0.011)
0.017
(0.012)
0.001
(0.007)
-0.005
(0.009)
Black -0.131***
(0.019)
-0.075***
(0.017)
-0.074***
(0.019)
-0.056***
(0.010)
-0.041**
(0.014)
Asian -0.038
(0.023)
-0.017
(0.022)
0.012
(0.024)
-0.002
(0.014)
-0.040*
(0.016)
Hispanic -0.033
(0.017)
-0.040**
(0.015)
0.017
(0.017)
-0.036***
(0.009)
-0.013
(0.013)
American Indian -0.100*
(0.043)
-0.056
(0.040)
-0.018
(0.047)
-0.049*
(0.023)
0.018
(0.042)
Single -0.049***
(0.012)
-0.033**
(0.012)
-0.027*
(0.012)
-0.017*
(0.008)
-0.047***
(0.010)
Divorced or widowed -0.039
(0.023)
-0.053**
(0.020)
-0.028
(0.023)
-0.013
(0.014)
-0.016
(0.018)
Constant -1.271
(0.109)
-1.093
(0.096)
-0.834
(0.101)
-0.213
(0.053)
-0.223
(0.070)
Log-likelihood -3786 -3305 -3853 -2080 -2851
Chi-squared 575 554 404 319 607
Number of observations 7072 6953 6962 7055 7046
Note: The base group includes persons with the following characteristics: no freedom in job, unskilled
or semi-skilled, female, white, born in 1974, married, no children. Dummy variables were also included
for year of birth, number of children, highest level of educational attainment, industry worked in and
region of residence (results not reported here). ( ) = robust standard error; * = significant at 0.05; ** =
significant at 0.01; *** = significant at 0.001.
21
Table 6. Job satisfaction: estimated marginal effects for full-time female workers
Explanatory variables Domains of job satisfaction
Pay Fringe
benefits
Promotion
prospects
Job security Importance /
challenge of
work
Job characteristics
Job autonomy: small freedom 0.113***
(0.030)
0.069**
(0.026)
0.147***
(0.030)
0.053***
(0.015)
0.107***
(0.019)
Job autonomy: some freedom 0.142***
(0.028)
0.094***
(0.025)
0.222***
(0.028)
0.065***
(0.014)
0.167***
(0.018)
Job autonomy: basically own boss 0.183***
(0.035)
0.119***
(0.031)
0.251***
(0.035)
0.086***
(0.018)
0.263***
(0.026)
Current income (log) 0.215***
(0.020)
0.101***
(0.015)
0.077***
(0.016)
0.016
(0.009)
0.018
(0.011)
Hours worked (log) -0.211***
(0.032)
0.047
(0.026)
-0.009
(0.028)
0.029*
(0.014)
0.022
(0.018)
Occupation: professional -0.036
(0.038)
0.020
(0.029)
0.082*
(0.035)
0.042
(0.020)
0.141***
(0.027)
Occupation: managerial -0.013
(0.037)
0.027
(0.029)
0.116***
(0.035)
0.053**
(0.020)
0.046
(0.024)
Occupation: non-manual 0.001
(0.033)
0.048
(0.025)
0.085**
(0.030)
0.024
(0.016)
0.017
(0.020)
Occupation: skilled manual 0.035
(0.049)
0.082*
(0.040)
0.071
(0.046)
0.027
(0.025)
0.041
(0.032)
Personal characteristics
Black -0.131***
(0.029)
-0.086***
(0.024)
-0.109***
(0.027)
-0.049***
(0.014)
-0.048*
(0.019)
Asian -0.036
(0.034)
-0.009
(0.033)
-0.009
(0.034)
0.026
(0.024)
-0.030
(0.024)
Hispanic -0.059*
(0.026)
-0.048*
(0.023)
-0.006
(0.026)
-0.040**
(0.014)
-0.021
(0.018)
American Indian -0.167*
(0.065)
-0.072
(0.054)
-0.030
(0.069)
-0.068
(0.031)
-0.039
(0.055)
Single -0.052***
(0.018)
-0.018
(0.016)
-0.019
(0.018)
-0.015
(0.010)
-0.048***
(0.013)
Divorced or widowed -0.061*
(0.032)
-0.053*
(0.026)
-0.071*
(0.030)
-0.012
(0.018)
-0.044*
(0.023)
Constant -1.279
(0.160)
-1.040
(0.135)
-0.796
(0.144)
-0.192
(0.074)
-0.144
(0.093)
Log-likelihood -2094 -1629 -1965 -1049 -1402
Chi-squared 231 331 196 158 366
Number of observations 3721 3408 3436 3492 3481
Note: The base group includes persons with the following characteristics: no freedom in job, unskilled
or semi-skilled, white, born in 1974, married, no children. Dummy variables were also included for year
of birth, number of children, highest level of educational attainment, industry worked in and region of
residence (results not reported here). ( ) = robust standard error; * = significant at 0.05; ** = significant
at 0.01; *** = significant at 0.001.
22
Table 7. Job satisfaction: estimated marginal effects for full-time male workers
Explanatory variables Domains of job satisfaction
Pay Fringe
benefits
Promotion
prospects
Job security Importance /
challenge of
work
Job characteristics
Job autonomy: small freedom 0.117***
(0.027)
0.084***
(0.025)
0.160***
(0.028)
0.048***
(0.014)
0.091***
(0.020)
Job autonomy: some freedom 0.134***
(0.026)
0.104***
(0.023)
0.178***
(0.027)
0.076***
(0.013)
0.141***
(0.019)
Job autonomy: basically own boss 0.165***
(0.031)
0.107***
(0.028)
0.233***
(0.032)
0.077***
(0.017)
0.175***
(0.024)
Current income (log) 0.211***
(0.019)
0.098***
(0.016)
0.070***
(0.016)
0.033***
(0.010)
0.012
(0.012)
Hours worked (log) -0.190***
(0.032)
0.021
(0.031)
0.057
(0.032)
0.017
(0.017)
0.072***
(0.022)
Occupation: professional -0.051
(0.028)
0.023
(0.025)
0.083**
(0.028)
-0.000
(0.016)
0.112***
(0.022)
Occupation: managerial -0.004
(0.029)
0.048*
(0.026)
0.078**
(0.028)
0.031
(0.018)
0.037
(0.021)
Occupation: non-manual -0.032
(0.025)
0.040*
(0.023)
0.057*
(0.025)
-0.001
(0.014)
0.021
(0.018)
Occupation: skilled manual -0.010
(0.024)
-0.006
(0.021)
0.041
(0.033)
-0.005
(0.014)
-0.031
(0.018)
Personal characteristics
Black -0.126***
(0.026)
-0.063*
(0.025)
-0.032
(0.028)
-0.057***
(0.014)
-0.028*
(0.021)
Asian -0.039
(0.029)
-0.025
(0.030)
0.032
(0.035)
-0.018
(0.017)
-0.053
(0.022)
Hispanic -0.010
(0.022)
-0.035
(0.020)
0.044
(0.024)
-0.033**
(0.012)
-0.005
(0.018)
American Indian 0.009
(0.060)
-0.037
(0.059)
-0.001
(0.064)
-0.026
(0.035)
0.093
(0.066)
Single -0.046**
(0.018)
-0.051**
(0.017)
-0.031
(0.019)
-0.019
(0.011)
-0.036*
(0.014)
Divorced or widowed -0.016
(0.035)
-0.053
(0.030)
0.031
(0.036)
-0.011
(0.020)
0.034
(0.033)
Constant -1.281
(0.160)
-0.997
(0.149)
-0.954
(0.155)
-0.264
(0.077)
-0.341
(0.112)
Log-likelihood -1791 -1661 -1857 -1008 -1425
Chi-squared 309 250 247 207 272
Number of observations 3574 3545 3526 3563 3565
Note: The base group includes persons with the following characteristics: no freedom in job, unskilled
or semi-skilled, white, born in 1974, married, no children. Dummy variables were also included for year
of birth, number of children, highest level of educational attainment, industry worked in and region of
residence (results not reported here). ( ) = robust standard error; * = significant at 0.05; ** = significant
at 0.01; *** = significant at 0.001.
23
Figure 1. Two-tier model of job satisfaction
Satisfaction with pay
Satisfaction with promotion prospects
Satisfaction with fringe benefits
Satisfaction job security
Job satisfaction with importance of work
Personal
and job
characteristics
Overall job
satisfaction
24
