ERC Working Papers in Economics 03/03

January 2002

Job Satisfaction in Britain: Individual and Job Related Factors

Şaziye Gazioğlu
Department of Economics, University of Aberdeen, Aberdeen AB24 3HQ,UK
and
Middle East Technical University, 06531, Ankara, Turkey
e mail: [email protected]

Aysıt Tansel
Department of Economics, Middle East Technical University, 06531 Ankara, Turkey
e mail: [email protected]

Economic Research Center

Middle East Technical University
Ankara 06531 Turkey
www.erc.metu.edu.tr
 Abstract
Recently there is a resurgence of interest in the analysis of job satisfaction variables. Job satisfaction is correlated with
labor market behavior such as productivity, quits and absenteeism. Recent work examined job satisfaction in relation to
various factors. In this paper four different measures of job satisfaction are related to a variety of personal and job
characteristics. We use a unique data of 28 240 British employees Workplace Employee Relations Survey (WERS97). Our
data set is larger and more recent than in the previous studies. The four measures of job satisfaction considered are
satisfaction with influence over job, satisfaction with amount of pay, satisfaction with sense of achievement and satisfaction
with respect from supervisors. Although the job satisfaction measures we use are somewhat different than those that are
previously used in the literature, a number of results that are commonly obtained with international data are found to hold
in our data set as well.

Key Words: Job Satisfaction, Individual Characteristics, Job Related Factors, Britain

JEL Classifications: L20, L29, J50

2
1. Introduction:

Many economists consider self-reported job satisfaction as a fascinating subjective variable

(Levy-Garboua and Montmarquette, 2002:1). This attraction recently resulted in a number of studies
in the empirical analysis of well-being, specifically in job satisfaction. Job satisfaction has been
investigated in several disciplines such as psychology (Argyle 1989), sociology (Hodson,1985;
Kalleberg and Loscocco ,1983), economics (Hamermesh 1977, 2001; Freeman, 1978), and
management sciences (Hunt and Saul ,1975). Employers prefer that their employees be satisfied, since
employees satisfaction is closely related to their labor market behavior such as productivity, quits and
absenteeism. Several findings indicate that job satisfaction is as good predictor of quits as wages
(Freeman, 1978; Akerlof et al., 1988; Clark et al. (1998). For this reason it is important to study the
determinants of job satisfaction. Different aspects of job satisfaction are studied in the literature.
These include job satisfaction with gender (Clark ,1997), wage growth (Clark ,1999), age (Hunt and
Saul (1975), Clark et al., 1996), comparison income and unemployment (Clark and Oswald,
1994,1996) work environment (Idson, 1990), work environment and relations with managers
(Gazioglu and Tansel, 2002). Locke (1976) defines the job satisfaction as the individual’s subjective
valuation of different aspects of their job. Higher job satisfaction may be due to improvements in the
objective aspects of the job either from reduced expectations or dissatisfing aspect of job is
downplayed while pleasing aspects are given greater weight. Hamermesh (1977) is one of the first
studies that used job satisfaction data to investigate a model of occupational choice.

This paper analyses four different measures of job satisfaction, using British data and
investigates their relationship to individual and job characteristics. Individual characteristics include
age, sex, education and gender. Job characteristic include income, establishments size, hours of work
and industrial composition. Clark (1996)used British Household Panel Data (BHPD91) from 1991 to
investigate the determinants of job satisfaction. We use a more recent (1997) and larger data set in
order to see whether the previously obtained results are upheld. Although the job satisfaction measures
we use are somewhat different than those of Clark and others, a number of results that are commonly
obtained with international data are found to hold in our data set as well. One special difference from
the previous results in that the married individuals have lower job satisfaction levels than the
unmarried in our data set. Further, it was noteworthy that individuals in education and health sectors
are less satisfied with their pay but more satisfied with their sense of achievement.

3
 Section 2 presents a brief discussion of our data and the various measures of job satisfaction
used in this study. Section 3 analyses the ordered probit estimation results of the relationship between
job satisfaction measures and a number of individual and job characteristics. Section 4 includes
concluding remarks.

2. Data
We use a unique data of 28 240 British employees Workplace Employee Relations Survey from 1997
(WERS97). This is a matched employer-employee survey. It involves interviews with employees and
managers in over 3000 establishments. Survey represents employers of 15,8 million workers. This is
three-fourths of all employees in Britain. Clark 1996 also uses British data from BHPS 1991 survey of
5000 employees. Our data set is much larger and more recent than Clark’s and involves larger
establishments. Table 1 gives the distribution of reported job satisfaction measures. The job
satisfaction measures are recorded in the survey as five category ordered measure. In this ordered
measure the value of one corresponds to ‘very dissatisfied’ and five corresponds to ‘very satisfied’.
These four measures are satisfaction with influence over job, with amount of pay, with sense of
achievement and with respect of supervisors. The most frequent (mode) response for all measures of
satisfaction is ‘satisfied’ category. ‘Very dissatisfied’ and ‘dissatisfied’ categories are substantial,
especially for satisfaction with the amount of pay. Nearly 41 percent of employees reported to be
dissatisfied or very dissatisfied with their pay. While respective figures for satisfaction with influence
over job and with sense of achievements is each about 15 percent and satisfaction with the respect
from supervisors is 21 percent. Conversely, at the other tail of the distribution, those employees who
are very satisfied with the amount of their pay is only 3.5 percent while the same figure is between 11
and 15 percent for all other measures of job satisfaction. Thus, British workers seem less satisfied
with their pay but more satisfied with their job by other measures of job satisfaction. Table 2 cross
tabulates the four satisfaction measures with various individual and firm characteristics. It reports the
means of the variables for the categories of ‘satisfied’ and ‘very satisfied’ for each of the job
satisfaction measures. Findings in this table will be discussed along with the ordered probit
estimation results in the next section.

3. Estimation Results
Table 3 reports the maximum likelihood ordered probit estimates of job satisfaction. In each
regression the dependent variable is five category ordered measure of job satisfaction. In such ordinal

4
measures of dependent variable it is proper to use ordered probit techniques (Greene, 2002; Maddala,
1983). Below we provide a discussion of each factor separately.

Gender:
Table 2 indicates that women are more satisfied (or very satisfied) with various aspects of their jobs
compared to men. This result is confirmed in Table 3 when other control factors are introduced into the
regressions. The coefficient estimate of the male dummy variable is negative and statistically
significant throughout. The largest coefficient estimate for this variable is for satisfaction with the
amount of pay indicating that man are most dissatisfied with this aspect of their job compared to
women. This unexpected result has been confirmed in a number of studies (Clark ,1996;1997; Meng,
1990 and others). Some plausible explanations for this finding may be the following: Men and
women have different expectations from their jobs. Further more, their comparison groups may be
different. These may have a reflection on the way they answer the job satisfaction questions. Another
reason for the finding that women are more satisfied with their jobs might be that the types of jobs that
men and women do are different as well as their qualifications. There may also be a participation
effect. In the situations where women is a secondary bread-winner they may find it easier to leave the
labor market. Thus more of the women who are satisfied at work would be working compared to
dissatisfied women. This would create a sample selection problem (Heckman, 1979). The extent of
the gender differencial in job satisfaction is investigated in detail by Clark (1997). He found this
differential can not be explained by different jobs that man and women do or by sample selection. He
found that for groups for which the gender differential in job expectations is less likely, the gender
differential in job satisfaction disappears. Such groups included the young, the higher educated the
profesionals, those in male-dominated work places and those whose mothers had a professional job.
Clark also found some evidence that women have lower expectations.

Age
Clark (1996) and Clark et al. (1996) report a statistically significant U-shaped pattern in age for
several job satisfaction measures. Clark et al.(1996) also found that the U-shape in age is particularly
strong for full-time employees and stronger for men than for women. The same pattern is reported by
Warr (1992) for job-related well-being. However, there are some contradictory evidence on this issue
in the literature, such as that provided by O’Brien and Dowling (1981). We observe a non-linear
relationship between age and four measures of job satisfaction in Table 2. In all cases non-linearity
shows a U-shaped relationship, with those in the very young and old age groups being the most
satisfied. In the regressions in Table 3 where other variables are controlled for, age-square captures

5
the non-linear relationship between age and job satisfaction observed in Table 2. In all four measures
of job satisfaction equations, age and age-squared are statistically significant and carry negative and
positive signs respectively indicating a U- shaped relationship between age and job satisfaction.

The various measures of satisfaction reach a minimum at different ages. Satisfaction with
influence over job, satisfaction with the amount of pay, satisfaction with the sense of achievements,
satisfaction with respect from supervisors reach a minimum at the ages of 33, 36, 22 and 28
respectively. It is worthy to note that workers reach a minimum for satisfaction with the amount of
pay at an older age than in the case of other measures of satisfaction. Higher minimum age for
satisfaction with pay is also reported by Clark et al. (1996:67). Clark et al. (1996) provided
explanation for the U shape of job satisfaction with respect to age. They suggest that young workers
may feel satisfied because they have little experience about the labor market against which to judge
their own work. As they learn about the labor market with some years of experience they are able to
better judge the conditions of their work. With this experience, satisfaction drops in the middle ages.
One factor is the effect of the reduced aspirations with age. Older workers may have reduced
aspirations as they realize that they face limited alternative choices as they get older. It may also be
true that they may attach less importance to such ambitions. The second factor that may be
responsible for older workers’ higher levels of satisfaction could be a self-selection effect. Since
dissatisfied older workers may find it easier to leave the labor market or find an alternative more
satisfying job, the ones we observe in the labor market would the satisfied ones. The details of this
argument and a discussion of several other factors can be found in Clark et al and Kalleberg and
Loscocco (1983).

Education
In Table 2 we observe that the percentage of those who are satisfied or very satisfied with their
jobs is largest for the ‘other’ category, which is the lowest level of education. This result is confirmed
in Table 3, where we observe that ‘degree and postgraduate degree’ and ‘A-level-O-level’ holders
have lower levels of satisfaction compared to the individuals with lower level of education. The
differential dissatisfaction is highest in the case of degree and post-graduate degree holders. Degree
and postgraduate degree holders show the largest differential dissatisfaction with their sense of
achievement with their job and next largest dissatisfaction is shown with the influence over their job.
Association of higher levels of education with less satisfaction is a surprising but a well establish result
(Clark, 1996; Clark, Oswald and Warr, 1996; Clark and Oswald, 1996). Clark and Oswald (1996)
suggested that due to expectation differential between different levels of education, causal relationship

6
between education and job satisfaction is ambiguous. Blanch flower and Oswald (1999) and Clark and
Oswald (1996) initially found a positive effect of education. But, this positive effect disappeared once
a control for income is used in the regressions.

Marital Status
We separated the marital status into two categories as married and single. The married
category includes those living with spouse or partner and the single category includes those who are
single, widowed, divorced or separated. The cross-tabulation in Table 2 shows that the married
employees are less satisfied with their job than the single ones by all four measures of satisfaction.
The regressions in Table 3 confirm this results. The findings in the literature on the job satisfaction
and marital status have been mixed. Clark (1996) reports that the married employees are more
satisfied. It is well known that married individuals are happier in general. However, our resuts
indicate that they are less-satisfied with their jobs compared to the single individuals.

Health Problems
The survey question asks if the employee has any long-standing health problems or disabilities,
which limit what they can do at work, at home or in their leisure time. An affirmative answer indicates
health problems. The cross tabulation in Table 2 shows that a high proportion of employees with
health problems report high levels of satisfaction with all of the four measures of job satisfaction
except for satisfaction with amount of pay. Whereas, when other variables are controlled for, in the
regressions in Table 3 we observe a statistically significant negative relationship between health
problems and all for measures of job satisfaction, indicating that health problems lead to lower levels
of job satisfaction. This is similar to the results obtained by Clark (1996).

Race
The ethnic group that employee belongs to are categorized into White, Black and Asian. We
observe in the cross-tabulation in Table 2 that the proportion of Blacks who are satisfied with the
amount of their pay is very low. In concordance with this observation, in the regressions in the Table
3, Blacks have lower levels of satisfaction with the amount of their pay in comparison to the Asians.
Whites have higher levels of satisfaction with their pay in comparison to Asians. For the other
measures of job satisfaction Blacks and Whites are not significantly different from Asians. Clark
(1996) also finds that the Blacks are relatively dissatisfied with their pay.

7
Job Characteristics
The cross-tabulation in Table 2 show that higher weekly level of income is associated with
higher levels of job satisfaction. This table also shows a nonlinear relationship between weekly
income and different measures of job satisfaction. Therefore the regressions in Table 3 include
logarithm of weekly income. The coefficient estimates are positive and statistically significant
indicating that higher pay is associated with higher job satisfaction, except in the case of satisfaction
with the sense of achievement and respects from supervisors. The strongest relationship is found
between weekly income and satisfaction with the amount of pay. Clark and Oswald (1996) and others
found a strong negative relationship between job satisfaction and a comparison income where the latter
is measured in various ways.

Table 2 shows the proportion of employees who are satisfied with different hours of work per
week. As it is expected longer hours of work are associated with lower levels of satisfaction in all four
measures of job satisfaction. Accordingly, Table 3 includes logarithm of weekly hours of work, to
take the non-linearity into account. Hours of work are strongly and negatively related to the
satisfaction with the amount of pay consistent with economic theory. The coefficient estimate is
negative and statistically insignificant in the case of satisfaction with the sense of achievement. The
coefficient estimates are negative and statistically significant in the cases of satisfaction with influence
over job and satisfaction with respect from supervisors.

Establishment Size
The cross-tabulation in Table 2 indicate a non-linear relationship between establishment size
and the four satisfaction measures. Accordingly in Table 3 logarithm of establishment size is
introduced. It enters with a statistically significant negative coefficient throughout indicating lower
levels of satisfaction in larger establishments. The same result is found by Idson (1990) with the U.S.
data, and by Clark with British data. Gazioglu and Tansel (2002) investigated the nature of this
relationship and its connection to structure of work environment and employee-manager relationship.
The effect of establishment size on satisfaction with pay is weaker than in the other satisfaction
measures since it is well known that larger firms pay higher wages.

In Table 3 we observe a very strong negative relationship between union membership and the
four measures of job satisfaction. The results indicate that the union members are less satisfied with
their jobs. However there might be an issue of endogeneity since dissatisfied workers are more likely

8
to join the unions. The relationship between job satisfaction and union membership has been
investigated by several researchers such as Borjas (1979), Freeman (1978), Meng (1990) and Miller
(1990).

Job Security
With regards to job security the survey question asked if the respondent agrees with the
following statement: ‘I feel that that my job is secure in this work place’. In this study, those who
strongly agree or agree with this statement are assigned a value of one and zero otherwise. Inclusion
of this dummy variable in the regressions in Table 3 indicates that a secure job leads to highly
significant, higher satisfaction levels for all measures of satisfacion considered. The issue of job
security and job satisfaction is investigated by Blanchflower and Oswald (1999). They also found that
job satisfaction is higher among those with secure jobs. Blanch flower and Oswald further asked the
question if US job satisfaction falling because of increasing job insecurity or because of the decline of
trade unions. They found that the answer was negative to both of the querries.

Occupation and Industrial Composition

Three occupational categories are included. The cross-tabulation in Table 2 indicates that the
proportion of those employees in managerial and professional occupations who are satisfied or very
satisfied with their jobs are larger than the proportion of those who are in Clerical or Sales
occupations. Table 3 shows this pattern. Managers, professionals and clerical employees are more
satisfied with the influence over their job (insignificant in the case of clerical), with the sense of
achievements and with the respect they get from their supervisors, as compapered to the sales
employees. However, they are less satisfied with the amount of their pay as compared to the sales
employees. Clark (1996) also found that those at the higher end of the occupational scale report higher
satisfaction with various aspects of their work but are less satisfied with their pay.

With regards to the industrial composition Table 3 indicates that manufacturing sector is not
significantly different from the whole-sale and retail sector for all measures of the job satisfaction.
The coefficient estimates in the electricity, gas and water sector are all statistically significant and
positive throughout indicating higher levels of satisfaction in this sector compared to whole-sale and
retail trade. Similarly in the construction sector higher satisfaction levels are obtained relative to the
whole-sale and trade sector. In the transportation sector negative and statistically significant
coefficient estimates indicate that employees in this sector are less satisfied with respect to all
satisfaction measures as compared to whole sale and retail trade sector employees. The coefficient

9
estimates for financial services sector are all insignificant except in the case of satisfaction with the
respect from supervisors. In the education sector the coefficient estimate on satisfaction with the
influence over job is insignificant. However, education sector employees seem unsatisfied with their
pay but highly satisfied with their sense of achievement and with respect from supervisors. The health
sector employees do significantly differ in their satisfaction with their influence over their job and with
respect from their supervisors from the whole-sale and retail trade employees. However they are less
satisfied with their pay and more satisfied with sense of acheivement as compared to wholesale and
retail trade sector employees.

Training Opportunities
With respect to training opportunities during the past year we observe that all of the
coefficients are statistically significant and positive indicating that training avaliability during the past
year leads to higher levels of job satisfaction as compared to no training. This is consistent with the
hypothesis that job satisfaction is an increasing function of training opportunities (Hamermesh, 1977)
4. Conclusions
This paper provides empirical analysis of the determinants of job satisfaction in Britain
considering a variety of personal and job related characteristics. This paper attempts to further the
analysis on employee job satisfaction. In many ways it updates the analysis of Clark (1996) with a
larger and more recent data set and somewhat different measures of job satisfaction. We used data
from the 1997 Workplace Employees Relationship Survey (WERS97) for 28,240 employees. A
number of results that are commonly obtained with international data are found to hold in our data set
as well, with a few exceptions. These can be summarized as follows: Man are less satisfied than
women various aspects of their jobs; job satisfaction is U-shaped in relation to age; higher educated
are less satisfied relative to those with no or few qualifications; married or spouse present employees
are less satisfied than non-married. This result is one of those, which is different from the other
studies. Health problems significantly reduce satisfaction; higher income produces higher levels of
satisfaction; long working hours reduce satisfaction; satisfaction is lower in larger establishments.
Union members are less satisfied than non-members; employees who feel that their job is secure
exhibit higher levels of job satisfaction; those who are in managerial professional and clerical
occupations are more satisfied than sales employees. Those employees who had job training were
more satisfied than those who had no training opportunities. We also reported results on job
satisfaction and the industrial composition and occupation which are not reported in the previous work.
With regards to the industrial composition satisfaction of those who work in the manufacturing sector
and the financial services sector is not significantly different from those in the whole-sale and retail

10
trade sector. Satisfaction of those who are in the electricity gas-water and construction sector is
significantly more than those in the whole sale and retail trade sector. It was note-worthy that those in
the education and health sector are less satisfied with their pay but more satisfied with their sense of
achievement. All in all, our results confirm the robustness of the findings in the literature on job
satisfaction both over time and with different measures of job satisfaction.

11
TABLES
Table 1: Distribution of Reported Job Satisfaction Measures
Satisfaction with Satisfaction with Satisfaction with Satisfaction with
Influence Over Job Amount of Pay Sense of Achievement Respect from Supervisors
Number Percent Number Percent Number Percent Number Percent

Levels:
Very Dissatisfied 887 3.2 3 478 12.5 1 275 4.6 2 358 8.5
Dissatisfied 3 308 12.0 7 896 28.2 2 872 10.3 3 515 12.7
Neither 7 162 26.0 6 568 23.5 5 941 21.4 5 636 20.3
Satisfied 13 075 47.4 8 911 32.3 13 519 48.6 12 250 44.2
Very Satisfied 3 156 11.4 1 028 3.5 4 180 15.0 3 973 14.3
Total 27 589 100.0 27 880 100.0 27 785 100.0 28 240 100.0

Table 2: Means of the Variables (%)

Satisfaction with Satisfaction with Satisfaction with Satisfaction with

Influence Over Job Amount of Pay Sense of Achievement Respect from Supervisors
Very Very Very Very
Satisfied Satisfied Satisfied Satisfied Satisfied Satisfied Satisfied Satisfied
Individual Characteristics:
Male 45.7 12.3 29.5 3.3 46.8 14.1 42.1 12.3
Female 48.9 11.2 35.0 3.6 51.2 16.0 46.3 16.3
Age:
20 or less 45.3 7.7 38.9 3.7 42.3 9.1 44.4 16.1
20-24 49.8 8.2 29.4 2.8 47.3 12.9 44.3 13.3
25-29 47.3 10.0 27.1 3.7 46.9 11.8 42.4 12.2
30-39 45.0 11.1 32.7 3.6 47.7 13.4 42.4 11.3
40-49 47.1 11.2 31.5 3.6 50.9 14.6 44.4 14.0
50-59 49.5 13.1 32.7 3.7 49.7 19.8 46.5 17.7
60 and over 54.6 21.7 37.8 6.2 52.8 27.3 49.1 26.3

Level of Education:
Degree+Postgraduate 47.8 12.8 34.0 3.9 49.6 16.2 45.8 13.9
A Level+0 level 46.4 10.3 32.1 2.9 48.2 13.0 43.6 13.0
Other 49.2 13.8 30.9 4.2 50.3 18.7 44.2 17.9
Marital Status:
Married+Spouse Present 45.9 9.9 30.0 2.7 46.8 13.1 43.3 14.0
Single 47.9 12.5 33.3 3.7 50.0 15.9 44.7 14.5

Health Problems 55.2 20.1 38.9 6.6 52.2 25.8 50.7 25.4

Race:
White 47.4 11.7 32.7 3.5 9.1 15.2 44.4 14.3
Black 42.9 11.8 19.1 1.3 49.0 10.6 40.1 12.1
Asian 47.4 12.0 25.9 3.1 46.8 13.0 43.6 14.5

Job Characteristics:
Weekly Income:
Less than £140 49.0 10.6 36.0 3.7 49.7 17.0 47.0 19.1
£141-£260 44.7 10.1 33.4 1.9 47.6 12.8 41.4 12.9
£261-£430 46.6 11.1 31.5 2.6 48.6 14.1 43.0 11.7
£431 or more 50.9 16.8 45.0 7.2 52.0 18.1 48.2 14.8

Hours of Work per Week:

Less than 16 48.5 11.6 39.2 4.5 49.0 17.9 47.1 18.9
16-29 49.2 10.1 39.1 3.2 50.6 17.2 48.1 17.5
30-39 46.6 10.4 28.6 3.1 48.2 12.4 42.7 12.8
40 hours or more 47.0 13.2 31.4 3.6 49.2 15.8 43.7 13.4

Union Member 44.2 9.5 31.0 2.9 57.3 13.6 41.7 11.5

12
Gender Concentration of the Work Place
Mostly Men 46.4 12.0 31.1 3.5 48.1 13.8 42.7 12.1
Mostly Women 48.0 11.6 33.3 3.4 49.9 16.2 45.7 16.3

Establishment Size:
Less than 25 employees 49.9 14.4 33.1 3.3 51.9 18.1 44.4 19.0
25-99 employees 47.9 11.8 31.9 3.4 49.9 16.0 44.5 15.5
100-199 employees 46.1 11.9 31.1 3.7 47.9 14.0 43.6 12.6
200-499 employees 46.5 11.2 33.4 3.3 47.7 13.7 43.7 12.6
500 or more employees 47.0 10.4 32.7 3.7 48.3 14.1 43.3 13.3

Occupation:
Managerial/Professional 49.9 14.5 35.7 4.4 51.5 18.0 46.6 15.2
Clerical 47.1 9.8 29.7 2.5 49.1 14.2 43.6 13.7
Sales 43.7 9.9 31.1 3.3 45.2 11.3 42.1 13.3

The Satisfaction Means for Those Who Agree or Strongly Agree with the Following Statements:
I Feel that My Job is Secure in this Work Place:
52.7 15.3 37.3 4.7 52.9 18.8 49.4 19.4

Table 3: Maximum Likelihood Ordered Probit Estimates of Job Satisfaction

Satisfaction with Satisfaction with Satisfaction with Satisfaction with

Influence Over Job Amount of Pay Sense of Achievement Respect from Superv.
Coefficient t-Ratio* Coefficient t-Ratio* Coefficient t-Ratio* Coefficient t-Ratio*

Male -0.066 3.95 -0.244 14.64 -0.919 5.44 -0.113 6.86

Age -0.188 4.10 -0.027 6.20 -0.010 2.11 -0.021 5.00

Age Square (x10-3) 0.305 5.48 0.374 6.99 0.231 4.20 0.374 7.18

Level of Education
Degree + Postgraduate -0.275 10.41 -0.205 7.74 -0.295 11.09 -0.200 7.65
A level + 0 level -0.155 7.52 -0.063 3.08 -0.178 8.56 -0.127 6.46

Married -0.077 4.65 -0.404 2.45 -0.065 3.96 -0.020 1.20

Health Problems -0.210 6.95 -0.142 4.66 -0.164 5.61 -0.143 4.88

Race:
White -0.054 1.24 0.135 3.09 0.033 0.76 -0.063 1.47
Black -0.045 0.64 -0.143 1.88 -0.044 0.60 -0.107 1.50

Job Characteristics:
Log Weekly Income 0.133 6.81 0.621 37.05 0.021 1.08 0.025 1.30
Log Hours of Work -0.158 6.45 -0.856 45.63 -0.038 1.54 -0.174 7.50
Log Estab. Size -0.026 4.48 -0.016 2.84 -0.038 6.41 -0.025 4.27
Union Member -0.239 15.48 -0.143 9.25 -0.161 10.35 -0.207 13.62
Secure Job 0.521 36.29 0.341 24.02 0.455 31.72 0.532 37.89

Occupation:
Managerial/Professional 0.272 10.75 -0.060 2.43 0.351 13.79 0.256 10.38
Clerical 0.064 3.14 -0.170 8.55 0.166 8.22 0.101 5.24

Gender Concentration:
Mostly Men 0.011 0.58 -0.032 1.67 0.073 3.81 -0.267 1.42

Industrial Composition:
Manufacturing 0.031 1.30 0.037 1.56 0.004 0.16 -0.020 0.78
Electricity+Gas+Water 0.142 3.86 0.378 10.51 0.063 1.70 0.152 4.01
Construction 0.112 3.10 0.067 1.85 0.083 2.36 0.087 2.48
Transportation -0.138 4.46 -0.060 1.89 -0.108 3.37 -0.089 2.86
Financial Services 0.026 1.13 0.022 0.94 0.361 0.56 0.089 3.86
Education Sector -0.014 0.54 -0.153 5.82 0.269 10.27 0.129 5.03
Health Sector -0.024 0.91 -0.124 4.90 0.220 8.45 0.012 0.50

13
Training:
Less Than 5 Days 0.074 4.51 0.104 6.38 0.124 7.58 0.169 10.02
5 Days or More 0.222 10.61 0.189 9.15 0.336 10.10 0.356 17.16

Constant
Threshold Parameters: 2.035 17.72 1.43 13.44 1.614 14.30 2.058 19.31
μ (1) 0.904 57.50 0.986 92.98 0.709 53.59 0.618 59.71
μ (2) 1.724 99.68 1.630 133.97 1.431 94.74 1.264 102.36
μ (3) 3.236 164.00 3.190 167.08 2.922 165.68 2.646 172.44

- Log Likelihood 30 420 32 879 30 762 33 750

Chi-Squared (26) 2 465 2 834 2 590 3 022

Number of Observations 23 948 23 895 23 948 24 670

References

Akerlof, G.A. A.K. Rose, and J.L. Yellen, (1988): ‘Job Switching and Job Satisfaction in the US
Labour Market’ Brooking Papers on Economic Activity, 2: 495-582.

Argyle, M. (1989): The social Psychology of work, 2nd edn. Harmondsworth: Penguin.

Blanchflower, D.G. and A.J. Oswald (1999) ‘Well-Being, Insecurity and the Decline of American Job
Satisfaction’, mimeograph, Dartmouth College, USA and University of Warwick, UK.

Borjas, G, (1979) ‘Job Satisfaction, Wages and Unions’ Journal of Human Resources, 14: 21-40.

Clark, A.E. (1996): ‘Job Satisfaction in Britain’, British Journal of Industrial Relations, 34: 2 June,
189-217.

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 Habit-Forming? Evidence from Micro Data’ Journal of Economic Behavior
and Organisation,19, 79-200.

Clark, A. E. and A. J. Oswald (1994): ‘Unhappiness and Unemployment’ Economic Journal 104,
648-659.

Clark, A. E. and A. J. Oswald (1996): ‘ Satisfaction and Comparison Income’ Journal of Public
Economics 61, 359-381.

Clark, A.E., A. Oswald and P. Warr (1996) ‘Is Job Satisfaction U-Shaped in Age?’ Journal of
Occupational and Organizational Psychology, 69: 57-81.

Clark, A.E., Y. Georgellis and P. Sanfey (1998): ‘Job Satisfaction, Wage Changes, and Quits’
Research in Labour Economics, 17, 95-121.

Freeman, R. (1978) ‘Job Satisfaction as an Economic Variable’ American Economic Review, 68:
135-141.

14
Gazioglu, S. and A. Tansel (2002) ‘Job Satisfaction, Work Environment and Relations with Managers
in Britain’, miniograph, Department of Economics, Middle East Technical University, Ankara.

Heckman, J. (1979) ‘Sample Selection Bias as a Specification Error’ Econometrica, 47: 153-161.

Green W.H. (2002): Econometric Analysis, Fourth Edition, Prentice Hall, term inc., NY.

Hamermesh, D. S. (1977): ‘Economic Aspects of Job Satisfaction’ in Ashenfelter, O. C. Oates, W. E.

(Eds.); Essays in Labor Market Analysis, New York: John Wiley.

Hamermesh, D. S. (2001): ‘The Changing Distribution of Job Satisfaction’ Journal of Human

Resources, 36: 1-30

Hodson, R. (1985): ‘ Workers Comparisons’ Social Science Quarterly, 66: 266-280.

Hunt, J. W. and P.N. Saul (1975) ‘The Relationship of Age, Tenure, and Job Satisfaction in Males and
Females’ Academy of Management Journal, 18: 690-702.

Kalleberg, A.L. and K.A., Loscocco, (1983): “Aging, Values and Rewards Explaining Age Differences
in Job Satisfaction”, American Sociological Review, 48, 78-90.

Levy-Garboua, L. And C. Montmarquette (2002): ‘Reported Job Satisfaction: What Does It Mean?’
mimeograph, Department of Economics,University of Montreal.

Locke, E. A. (1976) ‘The Nature and Causes of Job Satisfaction’ in M. D. Dunnette (eds) Handbook
of Industrial and Organizational Psychology. Chicago: Rand McNally.

Maddala, G.S. (1983) Limited-Dependent and Qualitative Variables in Econometrics, Cambridge

University Press.

Meng, R. (1990): ‘The Relationship Between Unions and Job Satisfaction’ Applied Economics,
22:1635-1648.

Miller, J. (1980): ‘Individual and Occupational Determinants of Job Satisfaction in Britain’ Sociology
of Work and Occupations 7: 337-66.

Miller, P. (1990): ‘Trade Unions and Job Satisfaction’ Australian Economic Papers, 29: 226-248.

O’Brien, G. E. and P. Dowling (1981): ‘Age and Job Satisfaction’ Australian Psychologist, 16:49-61.

Warr, P.B. (1992): ‘Age and Occupational Well-Being’, Psychology and Aging, 7:37-45.

15