Working Conditions and Time Use

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Working Conditions and Time Use

Author(s): Morten Blekesaune


Source: Acta Sociologica, Vol. 48, No. 4, Acta Sociologica 50 Years: Celebrating Scandinavian
Sociology 1955-2005/2006 (Dec., 2005), pp. 308-320
Published by: Sage Publications, Ltd.
Stable URL: http://www.jstor.org/stable/20059961
Accessed: 07-05-2018 19:00 UTC

REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/20059961?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Sage Publications, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to
Acta Sociologica

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
ACTA SOCIOL?GICA 2005

Working Conditions and Time Use


Morten Blekesaune
Institute for Social & Economic Research, University of Essex, UK

abstract: What we do when we get home is influenced by what we do at work. In


this article I analyse the relationship between market work and domestic activities
from the perspectives of economic theories of time use, i.e. the work-leisure model
(compares income work versus other activities) and the household production
model (compares income work and household work), and sociological theory
treating the issue of how leisure activities compensate for work strains. The
empirical analysis of time use is based on the Norwegian time-use survey from
2000-2001. Estimates on working conditions are based on the Norwegian level of
living surveys from 1996 and 2000 for 100 occupations found in both surveys. The
analyses indicate that both physical strains and low autonomy at the place of
employment lead to an allocation of time away from paid work towards unpaid
work and leisure, but with considerable differences for men and women. The
findings indicate that working conditions are part of the trade-off between paid and
unpaid work, and that leisure activities may compensate for human needs being
neglected in the workplace.

keywords: household work leisure time use work working conditions

Introduction
Within time-use studies, work is normally predefined according to what is labelled a third
person criterion. Work includes activities other people could have done for us. This third
person criterion implies that work is an instrumental activity, something we do for its results
rather than for the work itself. However, we may to some extent think about all activities as
something we do in order to achieve results as well as for the intrinsic value attached to
performing them. Both income and household work can involve pleasure and other direct
benefits, and leisure activities may be instrumental in their nature. Our occupational work can
be enjoyable and provide opportunities for freedom of choice and self-actualization. Leisure
activities yield opportunities for recreating our working capability, enhancing our productiv
ity or enhancing the benefits we may gain from various forms of consumption.
Time-use studies normally provide no measures of whether the various activities performed
are instrumental or whether they are done for their own sake. It is difficult to study the direct
effects of various activities on quality of life. One approach is to ask whether activities
performed can be termed 'leisure' or whether they are related to similar concepts (Shaw, 1985;
Siegenthaler and Vaughan, 1998). However, similar labelling concepts do not exist in all
languages, and even when they do they will be vague, and research indicates that their use
varies between groups of people (Shaw, 1985).
An alternative strategy is to measure or to provide estimates of the unpleasantness in
performing various activities. Survey data often provide information about strains within

Acta Sociol?gica December 2005 Voi 48(4): 308-320 DOI: 10.1177/0001699305059944


Copyright ? 2005 Nordic Sociological Association and SAGE (London, Thousand Oaks, CA and New Delhi)
www.sagepublications.com

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Blekesaune: Working Conditions and Time Use
lllllllll
occupational work; strains that may be seen as burdens we accept because they allow us to
realize benefits later on via, for example, income. Occupational burdens may be important
sources of inequality, since not everyone has to put up with them. In this article, I explore the
relationships between job-related strains and time allocated to income work, household work
and leisure activities, as well as the composites of household work and leisure activities.
Estimates of working conditions are based on Norwegian level of living surveys from 1996
and 2000, and time-use data are from the Norwegian time-use survey of 2000-2001.

Theory and hypotheses


The relationship between work and other activities has been studied from various perspec
tives. Economists tend to begin with theoretical reasoning, whereas sociologists more often
take empirical observations as the point of departure. However, the relevant theoretical
models have not been specified in such a way that unambiguous implication is apparent
between different types of job stress and time use. Furthermore, little empirical research has
been published on the relationship between working conditions and time use, Karasek's thesis
from 1976 being one exception. There are, however, other research traditions that can help
explain how working conditions may affect time use.
Labour supply models in economics distinguish between work and leisure (Killingsworth,
1983). Work is typically seen as an instrumental (non-enjoyable) activity only, but it provides
an income that can pay for enjoyment during leisure time. This work-leisure model assumes
that individuals make a choice between the time they allocate to income work and leisure in
such a way that they derive similar benefits from the last hour allocated to both sets of
activities.
The theory of the household production function (Becker, 1965,1991) assumes that a similar
choice is made between the time allocated to income work and household work (Peters, 1995).
Many of the products we consume comprise a combination of inputs from both income work
and household work. We may buy food in the market which is brought to the household and
prepared for final consumption. Shopping, travelling, preparation of the meal and cleaning
are all activities entailed within the production function of the meal along with occupational
work where income is realized, thus allowing us to buy market food inputs. Similar lines of
reasoning have been presented elsewhere within economics (Linder, 1970) and sociology
(Gershuny, 1988, 2002).
Similar empirical implications can also be made from sociological traditions. First, a time
available model posits a trade-off between household work and income work similar to that
described in the theory of household production function (Presser, 1994). The time available
model also emphasizes how changes in family composition affect the allocation of time
between household work and income work, particularly among women. Second, there is a
tradition which predicts time use in general, and household work in particular, as an outcome
of gender role attitudes (Berk, 1985). These gender-related attitudes may be a result of early
socialization (Cunningham, 2001), or they may develop through day to day marital interaction
(Hochschild, 1989). Either way, household work is not primarily seen as an alternative to
income work, but is predicted from other considerations. The empirical implications are thus
likely to be similar to the work-leisure model.
These models provide few indications as to who is expected to allocate more/less time to
the alternative time uses of income work. The theory of the household production function
posits that high-income consumers will buy a large part of what they consume ready made
in the market, while low-income consumers are expected to do a larger part of the production
process themselves as unpaid work (Becker, 1965, 1991). Lancaster (1966a, b) developed the
economic theory by arguing that income work can also possess characteristics that give rise
309

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Acta Sociol?gica 48(4)

to both satisfaction and dissatisfaction, i.e. from the work activity itself, in addition to the
benefits being realized via the goods produced or the income earned. One implication of this
is that if income work is unpleasant as a result of poor working conditions, this should lead
to an allocation of time away from income work towards other activities. The kind of unpleas
ant working conditions, and the other activities they should lead to, however, has not been
specified so far.
There is a tradition for studying the relationship between working conditions and retire
ment behaviour among elderly workers that is typically based on the reasoning of the
work-leisure model, i.e. early retirement is viewed as a choice of leisure in place of further
income work (Quinn, 1977; Filer and Petri, 1988). At least three kinds of working conditions
have been found to be predictive of early retirement: hard physical work (Quinn, 1978; Filer
and Petri, 1988), repetitive job tasks (Quinn, 1978; Filer and Petri, 1988; Chirikos and Nestel,
1991) and low job autonomy (Quinn, 1978). Furthermore, these three types of working
conditions have been found to be associated with job satisfaction (Fagan and Burchell, 2002)
and sick leave (Eshoj et al., 2001). In this article, the impacts of these three kinds of job strain
on everyday time use are studied. The hypotheses are: people with physically hard work, in
low autonomy jobs and with monotonous/repetitive job tasks will allocate less time to income
work and more time to household work and leisure. In so far as these correlations are found,
similar effects on the composition of household work and leisure activities are also studied.
Will the presence of these three job strains lead to more time being allocated to household
work or to leisure activities? Studies of retirement behaviour give no clear implications. We
can approach the issue by studying the extent to which household work and leisure activities
are themselves associated with the job strains in question: are these activities physically
strenuous, do they involve freedom of choice or monotonous/repetitive job tasks?
It is hypothesized that leisure activities comprise few involuntary strains. Thus, any
unpleasant job strain leads to a stronger preference for leisure. Household work, however,
comprises some, but not all, of the job strains studied. Research comparing home workers with
paid workers indicates that household work is not very satisfying, although it includes
autonomy or freedom of choice to a considerable extent (Bird and Ross, 1993; Ross and Wright,
1998). It is therefore hypothesized that low autonomy jobs lead to time being allocated from
income work to household work. Conversely, physically hard work and monotonous/repeti
tive job tasks lead to time being allocated from income work and primarily to leisure activi
ties, but some extra time may also be allocated to household work.
Where combinations of household work and leisure are concerned, theory provides even
fewer guidelines. But even here some guidelines can be established on the basis of empirical
research on the relationship between work and leisure. A traditional idea has been that in their
Tree time' people compensate for what they lack in their jobs, i.e. leisure activities compen
sate for strains or other drawbacks in occupational work (Parker, 1971; Staines, 1980). The
reasons for this compensating characteristic of leisure may be that leisure activities provide
opportunities for satisfaction that cannot be realized in occupational work, or that leisure may
help to restore working capability in the form of recovery (rest) or recreation (activity). The
latter explanation implies that job strains may require more time allocated to leisure activity.
Compensation theory has not generally been applied in studies of time use, but instead in
studies of participation in specific leisure activities. Leisure research has not found unambigu
ous support for these ideas, however. For example, workers who are allowed little freedom
of choice in their jobs make fewer decisions (Meissner, 1971) and participate less in civic
activity (Wilson and Musick, 1997) outside the workplace. Hence, an 'extension' or 'spillover'
model claims similarities between occupational work and leisure activity as an alternative to
the compensation model mentioned.
These positive correlations between work and leisure can have several explanations. One
310

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Blekesaune: Working Conditions and Time Use

is that skills and knowledge developed in occupational work can also be utilized in
household work and leisure activities. Similarly, leisure activities can promote skills and
knowledge that are beneficial in our occupational work. Another explanation is that individ
ual characteristics and preferences lead to similar choices in both income work and leisure
activity. This ambiguity between selection and causation effects in positive relationships
between work and leisure indicates that the 'extension' model is not well defined; it is
partially about empirical correlations and partially about why these variables correlate.
However, it should be possible to confine the influence of spurious (selection) effects on the
relationship between work and leisure through a careful choice of the variables studied, i.e.
dependent variables, explanatory variables and variables utilized for statistical control, as
well as through cautious interpretations.
The first hypothesis is that job strains lead to less time being allocated to income work. But
can people choose how many hours they work? Standardized working hours and requirements
from employers are likely to restrict individual choice with regard to the number of hours spent
at work. In Norway, standard working hours are not primarily defined by law but negotiated
between representatives for employers and employees, mostly at the national level. Specific
arrangements can be made for different occupations and businesses/trades. Negotiations are
also made at the level of firms/companies as well as between individual employees and
employers. Preferences for shorter working hours among workers with jobs strains may enter
the relevant negotiations at all these levels. Negotiated agreements (at collective levels) or job
contracts (for individual workers) may include adjustments in the number of hours worked or
in how these numbers should be specified (e.g. the number of job tasks to be carried out).

Data and methods


The time-use data were from the Norwegian time-use survey of 2000-2001, which employed
time diaries of 15-minute intervals.1 The dependent variables were use of time for income
work (including travel to and from work and rest periods during working hours), household
work (excluding care for children or other dependants) and leisure ('free') time, all measured
in minutes per day. Household work was also divided into housework (cooking, cleaning),
maintenance work and shopping, including travelling. Leisure time was divided into active
leisure (outdoor activities/entertainment, cafes/restaurants, organizations), social leisure
(being with friends/relatives) and resting leisure (home entertainment such as reading or
watching TV). Linear regression analyses of the time-use data were applied. Descriptive statis
tics of the variables are given in Appendix 1.
Eighty-three percent of the individuals studied were observed on two (typically consecu
tive) days. The number of days recorded for each individual varied from one to four, however.
In the data matrix, the observation units were days and not individuals. In statistical terms
this leads to a dependence of the daily observations 'within' each individual. This was
corrected by 'sandwich' (Huber/White) estimators of the standard errors, which is also robust
for heteroscedasticity (skewed distribution of the dependent variables). The analysis of the
combination of household work and leisure activities also included multivariate regression in
order to assess whether the job strains affected the three kinds of household work or the three
kinds of leisure variables differently. For this purpose, the data matrix was collapsed from
days to individuals, and the number of days recorded was applied as a weighting variable.
Estimates of working conditions were based on two Norwegian level of living surveys from
1996 and 2000 for 100 (ISC088, 3 digit codes) occupations found in both the level of living and
the time-use surveys for those between 22 and 60 years of age. (Four occupations were
estimated with two-digit codes because no three-digit estimates were available.) The working
condition estimates applied were simple means for each occupation for men and women
311

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Acta Sociol?gica 48(4)

jointly. Their construction is presented in Appendix 2. The empirical analysis was restricted
to the working population (working 15-99 hours per week), and separate models were
estimated for men and women.
Time use is influenced by family phases involving factors such as number of (young)
children. It is also possible that time-use preferences influence the decision on having children,
perhaps also living with a partner. Hence, family phases and child care were controlled for,
i.e. marital status (living alone), number of children living in the household (one, two/three),
the presence of both preschool (below seven years) and grown-up children in the household
(i.e. above 20 years, an age when children become economically independent) as well as time
allocated to care for own children. Control variables included age of the respondent and
weekend days (Saturdays/Sundays).
Both income and education levels are possible confounders in this analysis. If those with
nice jobs work longer hours than those with unpleasant jobs, the reason could be that their
jobs are more exciting and less strenuous, but it could also be because they are more highly
paid. Similarly, highly educated groups could gain more from income work than household
work compared with the less educated. An income per hour worked variable was calculated
by subtracting work-related income (within a limit of $60,000) on the number of hours worked.
Unfortunately, the income variable was measured (from administrative records) for the
previous year, hence the analysis was also restricted to those with a reasonable work-related
income in that year (at least $12,000). The regression results are presented with education but
without the somewhat inaccurate income variable. However, all results were double-checked
concerning potential confounding effects of both income and education by taking these
variables in and out of the analysis.
The empirical models presented in the tables include all three types of job strain. Hence, the
impact of each strain is estimated while controlling for other strains. The three job strains
correlate. Particularly monotonous work correlates with both physical strains and low
autonomy jobs. In the analysis of time use, however, there is surprisingly low collinearity
between the three job strains studied, i.e. their coefficients change little when other job strains
are included in the analysis. However, statistical accuracy sometimes decreases when job
strains are studied jointly (the standard errors become larger).

Regression results
Income work, household work and leisure
Among women, physically strenuous jobs covary with less time allocated to income work and
more time allocated to both household work and leisure. Among men, however, there are no
similar covariations. These are the expected correlations for women, but not for men. The
different results for men and women are statistically significant for income work and leisure.
Apparently, physical strain in the job has more impact on the daily activities of women than
on those of men, as they lead to an allocation of time away from hard income work towards
more time in both household work and leisure. These gender differences may reflect that men
are physically stronger and tolerate physical strains better than women before it affects their
allocation of time between income work and other activities. The difference may also reflect
gender roles, as men are expected to endure physical strain and traditionally are assumed to
carry a major responsibility as breadwinners, which may make them more reluctant to reduce
their income work in situations involving heavy or unpleasant tasks.
Among both men and women, low autonomy jobs covary with less time allocated to income
work and more time allocated to leisure (significant in a one-tailed test also among men).
Among men, but not among women, low autonomy jobs covary with more time allocated to
household work. (The gender difference is statistically significant.) By and large, these are
312

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Blekesaune: Working Conditions and Time Use

expected correlations, particularly in the case of men. Low autonomy jobs lead to an alloca
tion of time away from income work towards other activities. For women, this substitution is
primarily between income work and leisure, and for men it is between income work and
household work. The substitution between income work and leisure probably reflects how
neglect or suppression of individual freedom in the job can be expressed through leisure activi
ties. If income work is unsatisfying, leisure becomes comparatively more attractive.
Why is the impact of a low autonomy job different for men than it is for women with regard
to household work? One possible explanation could be found in a gendered distribution of
household responsibilities; for example, in that women's household work is defined by family
composition, whereas men's household work is more strongly influenced by job characteristics.
If family responsibilities can be defined by the presence of children, this hypothesis can be
tested by comparing women with and without children, but it finds no support in the data.
Another explanation could be how men and women satisfy needs that are neglected in low
autonomy jobs; for instance, needs for self-actualization and esteem. It appears that for men
these needs are accommodated in household work, while for women they are better met in
leisure activity The reason may be related to the kind of household work men and women do.
Among men, but not among women, monotonous/repetitive jobs covary with less time
allocated to household work (the gender difference is statistically significant). But there is no
opposite covariation with income work. Apparently, monotonous jobs lead to more time in
sleep and in restoring physical capability, the major types of time use not studied in the tables.
This extra time spent in sleep occurs at the expense of household work. The reason could be
that monotonous/repetitive jobs lead to exhaustion that drains one's efforts in the home from
productive activities (household work) to pure rest/sleep. If this is so, among men monoton
ous/repetitive jobs have a detrimental effect on life when they get home, too. For some reason,
similar effects do not appear among women.
There is a possibility that some of the covariation between job strains and time use reflect
other characteristics of the individuals as well. One prospect is that of how much income one
can earn in an occupation. High-income groups should normally allocate more time to income
work than household work. If some of the job strains studied correlate with low incomes, it
may be that covariations found between unpleasant jobs and time allocation away from income
work towards household work, perhaps also to leisure, reflect how their income work pays off
in monetary terms. This alternative hypothesis can be tested by including the income per hour
variable in the analysis. However, this income variable has surprisingly little effect on the time
use categories studied, and nowhere does it change the results concerning the impact of work
strains on these categories despite the fact that both physical strains and monotonous work are
associated with low income. The reason could be that unpleasant jobs are to some extent
compensated with better incomes (Filer, 1989). Only among men does income per hour correlate
negatively with one of the work strain variables, i.e. physical strain. Number of years of
education makes no difference for the covariations between job strains and time use reported
above. Altogether, the results found are surprisingly robust for individual resources.
In addition to the factors mentioned above, the family composition variables have less effect
on time use than could be expected. However, having children leads to more time spent on
household work, even when child care is excluded from the category. For men, this effect
already appears with a first child; for women, it appears with the arrival of a second or third
child. Grown-up children in the household means more household work for mum, but not for
dad. Time spent in caring for children usually reduces income work, and to a lesser extent
leisure activity, but goes well with household work because the two activities correlate posi
tively. Among women, household work increases with age, also when family composition is
controlled for. This is most likely a reflection of a historical reduction in women's household
work that to some extent manifests as a difference between young and older cohorts.
313

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Acta Sociol?gica 48(4)

Table 1 Men's time spent in income work, household work and leisure as a result of characteristics of the men
(22-59 years of age), their households and occupations. Linear regression with robust/sandwich standard errors
(in parentheses); 840 men, 1773 days (observations)
Income work Household work Leisure time
Living alone 24 (21) -11 (10) 14 (15)
One child -4 (21) 30* (12) -8 (14)
2/3 children -12 (25) 31* (12) -15 (17)
Child below 7 years 41 (25) -11 (12) -24 (17)
Child above 20 years 11 (28) -2 (15) -12 (21)
Care-time children -75* (11) 15* (5) -31* (6)
Age (+40/10) 2 (19) 7 (10) -7 (13)
AgeA2 -2 (9) -4 (4) 0 (6)
AgeA3 -2 (8) 3 (4) 4 (6)
Education (years) 0 (3) -1 (1) 0 (2)
Weekend day -375* (12) 54* (7) 205* (10)
Physical strains 4 (11) 4 (5) -1 (8)
Low autonomy -37* (14) 29* (6) 17 (10)
Monotonous work 3 (13) -20* (6) -1 (9)
Constant 524 87 275
RA2 0.392 0.091 0.255
* Coefficients with p < 0.05 in two-tailed

Table 2 Women's time spent in income w


women (22-59 years of age), their hous
errors (in parentheses); 760 women, 1
Income work Household work Leisure time
Living alone -6 (18) -3 (9) 7 (14)
One child -14 (21) 8 (11) 25 (15)
2/3 children 11 (23) 38* (13) -10 (17)
Child below 7 years 10 (22) 10 (12) -30 (17)
Child above 20 years -36 (22) 35* (13) 10 (18)
Care-time children -82* (8) 13* (4) -24* (5)
Age (+40/10) 1 (20) 26* (9) -28 (15)
AgeA2 6 (10) 4 (5) -2 (7)
AgeA3 -10 (10) -2 (5) 13 (7)
Education (years) -2 (3) -1 (2) 2 (2)
Weekend day -286* (13) 28* (8) 161* (10)
Physical strains -36* (13) 14* (7) 26* (10)
Low autonomy -40* (14) 5 (7) 28* (10)
Monotonous work 7 (14) 1 (7) -7 (10)
Constant 524 124 201
RA2 0.317 0.068 0.206
* Coefficients with p < 0.05 in two-tailed tests.

Combination of leisure and household work


Do job strains (in income work) also affect the combination of household work tasks or t
combination of leisure activities? Household work was divided into housework, maintena
work and shopping/travelling. Leisure time was divided into active leisure, social and res
314

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Blekesaune: Working Conditions and Time Use

leisure. These hypotheses can be tested assuming equal influence on the three kinds of
household work or leisure in minutes per day, disregarding that time is allocated differently
between them in the first place. Alternatively, the combination of household work and leisure
activity can be studied by calculating each category of household work and leisure activity as
the percentage of all household work and leisure time, respectively. Hence, we can talk about
the effects on the combination of household work and leisure both in absolute and in relative
terms.
For men, we found that two out of six hypotheses tested for household work and leisure
were statistically significant (Table 1): the impact of low autonomy jobs on both household
work and leisure, disregarding the effect of monotonous work on household work (wrong sign
and no opposite covariation with income work). For women, three out of six hypotheses tested
were significant (Table 2): the effect of physically hard work on both household work and
leisure, and the effect of low autonomy jobs. When testing these five associations on the combi
nation of household work and leisure (with multivariate regression) two tests are significant
(in 2 d.f. tests for the three dependent variables studied), both concerning the effect of low
autonomy jobs on leisure activity, for men in absolute, but not in relative terms, for women in
relative, but not in absolute terms.
Among men, low autonomy jobs covary with more time in resting leisure in particular and
less time (although non-significant) in active leisure. The increase in resting leisure (16
minutes) corresponds to the similar increase in total leisure (17 minutes). We can therefore
conclude that low autonomy jobs are associated with more resting leisure only. Why do low
autonomy jobs lead to more time in resting leisure such as reading and watching TV? One
reason could be that this kind of leisure time satisfies human needs that are neglected or
suppressed in low autonomy jobs; for example, needs for self-actualization and esteem. If so,
it appears that when these needs are not accommodated in the job (among men) this does not
lead to more ambitious leisure activities like active or social leisure, but rather to some kind
of escape into other realms through media at home. Alternatively, there could also be a
selection effect in that men with a preference for watching TV, for whatever reasons, end up
in low autonomy jobs. However, all effects mentioned are virtually unchanged when
controlling for education level and hourly income. Hence, the alternative selection effect is not
supported by the data.
Among women, low autonomy jobs are associated with a leisure time combination of more

Table 3 A combination of three kinds of men's leisure activities, in minutes and in percentages, as a result of low
autonomy in the job. Models similar to Table 1 (other coefficients are not reported). Coefficients presented with
standard errors in parentheses

Low autonomy Active leisure Social leisure Resting leisure


In minutes -7 (6) 6 (5) 16 (6)
In percentages -2 (1) 1 (1) 1 (2)

Table 4 A combination of three kinds of women's leisure activities, in minutes and in percentages, as
low autonomy in the job. Models similar to Table 1 (other coefficients are not reported). Coefficients p
standard errors in parentheses

Low autonomy Active leisure Social leisure Resting leisure


In minutes 15 (8) 5 (6) 9 (6)
In percentages 3 (1) -2 (2) -1 (2)
315

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Acta Sociol?gica 48(4)

active leisure and relatively less social leisure. Women's active leisure increases by 15 minutes
per day, which otherwise makes up less than a quarter of women's leisure time (Appendix 1),
whereas social leisure increases by a mere 5 minutes per day. Taken together, low autonomy
jobs appear to have a different effect on leisure activities for men and women. Among women,
low autonomy jobs are associated with more active leisure, among men with less active leisure.
(The gender difference is statistically significant.) The reason could be that men and women
make use of different strategies when human needs for self-actualization and esteem are
neglected in their jobs. For men, the strategy is one of escape into other realms, whereas
women choose a more proactive strategy in being more active in alternative arenas.

Discussion
This article has as its point of departure the idea that the allocation of time between occu
pational work and alternative activities depends on how unpleasant the occupational work is,
measured as three kinds of job strain: physically hard work, low autonomy jobs and mon
otonous/repetitive job tasks. These estimates were based on external data (other surveys) for
100 occupations. Hence, generalizations are made about occupations and not about indi
viduals. This methodology has advantages and disadvantages. What people report in response
to questions about their job strains may reflect strains related to their occupations, but also
personal attributes such as fitness, ways of coping and preferences for work and leisure.
External estimates of job strains for occupations are likely to have smaller correlations with
personal, rather than occupational, attributes that may affect how time is allocated between
income work and other activities. As a result, some ambiguity concerning causal relationships
between job characteristics and leisure activities is avoided. Nonetheless, one argument could
be made that those with strong preferences for leisure or household work may choose certain
kinds of occupations, which could contribute to some of the observed results. However, I find
no (indirect) indication of this (selection hypothesis) in the contribution of individual level
resource variables such as education and hourly income. Conversely, if major differences in
occupational strains are found within the occupations studied (e.g. if they vary between
sectors and workplaces), the estimated effects of job strains on time use will be inaccurate and
probably underestimated.
The empirical analysis supported the hypothesis about off-work compensation for occu
pational strains for two of the three occupational strains studied. Low autonomy jobs are
associated with shorter working hours and more time at leisure, and for men also more time
doing household work. For women, physically hard jobs are also associated with shorter
working hours and more time in both household work and leisure activity. Seemingly, only
one of the occupational work characteristics affect time use at home for men and women
similarly i.e. that of low job autonomy on leisure time. But even this similarity vanished when
I distinguished between the three main types of leisure activity involved. A major finding,
therefore, is that the job strains studied affect the time use of men and women differently, and
a major puzzle is why these gender differences occur.
Apparently, men and women experience seemingly similar occupational strains differently,
and in terms of time use they respond to them in different ways. Women, but not men, work
shorter hours when the work is physically strenuous, and the working hours avoided are spent
in both household work and leisure activity. This finding was expected for women, but not
for men. The reasons may be that men endure physically hard work better than women, or
that they are expected to endure this kind of hardship in their jobs. Both men and women
work shorter hours in low autonomy jobs. For men, the working hours avoided are primarily
spent in household work, but also in leisure. For women, the working hours avoided are
mainly spent in leisure activity It is speculated that the difference may be related to the kinds
316

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Blekesaune: Working Conditions and Time Use

of household work men and women do. However, the extra housework for men in low
autonomy jobs is mostly spent in ordinary housework (cooking, cleaning) and in
shopping/travelling (both effects are statistically significant), which is otherwise typical for
women's household work. The explanation cannot therefore be found in a gendered distri
bution of household responsibilities. The effect of low autonomy jobs on added leisure time
probably reflects how the need for individual freedom neglected at work can be realized
through leisure activities. If income work is unsatisfying, leisure becomes more attractive.
However, even this effect is expressed differently for men and women ? for men, mostly
through media use at home; for women, primarily through leisure activities out of the home.
The stronger trade-off between income and household work among men compared to
women in cases of low autonomy jobs may be surprising, since most household work, particu
larly ordinary housework, is traditionally in the female domain. Hence, one could assume that
income work and household work are more likely substitutes for women than they are for
men. One interpretation is that men, more so than women, perceive their income work and
household work as being similar types of activity done mainly for their results. Alternatively,
men could express more individual freedom in cooking, the main type of time use involved,
compared with women, whereas women may have a more instrumental attitude toward
income work compared to men. A study of non-monetary rewards from both income work
and household work by Ross and Mirowski (1996) found stronger associations between
earnings from income work and psychological well-being for women than for men. In another
study (Bird and Ross, 1993), men and women evaluated housework as having similar charac
teristics and leading to similar levels of individual control. Both findings provide some
support for the latter interpretation (i.e. income work is more instrumental for women than
for men), but not to the first interpretation (i.e. men find both income work and household
work more instrumental than women do).
This study was based on developments of economic theories of time use (i.e. the work-leisure
model and the household production model) and a variation of a sociological theory about
work-leisure relationships (i.e. a hypothesis of leisure compensation for job strains). The theor
etical development was the inclusion of job strains in standard theories of time use, and the
variation employed was the application of compensation theory on time allocated to leisure
activity. This modification of the household production function model, or its sociological
variation labelled 'chains of provision' (Gershuny, 1988, 2002), explains the impact of low
autonomy jobs among men. A similar modification of the work-leisure model explains the
development of low autonomy jobs and physically demanding jobs among women. The applied
variation of compensation theory of work-leisure relationships also explains the impact of low
autonomy jobs and physically demanding jobs among women, i.e. more time in leisure activity
and less time working when the job entails certain strains. In fact, the choice between these two
models of explanation is probably one about linkage to different theoretical and disciplinary
traditions only. A deviating and unexpected finding is that people in monotonous jobs tend to
spend more time in sleep or pure rest and less time in household work. This finding falls within
the extension pattern proposed by Parker (1971), but I would hesitate to speculate why.
There is little previous research into the effect of working conditions on time use. The three
job strain variables utilized in this study (physically hard work, low autonomy jobs and
monotonous/repetitive job tasks) have previously been utilized in studies of retirement
behaviour, where they have been found to have explanatory power. The findings in this study
indicate that they affect different time uses differently Furthermore, there is little collinearity
between them, likewise with individual resource variables such as education and income.
Thus, the findings indicate that physically strenuous work, low autonomy jobs and monoton
ous/repetitive job tasks represent different forms of stress that cannot be conceptualized
simply as merely relating to good and bad jobs.
317

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Acta Sociol?gica 48(4)

Appendix 1 Descriptive statistics of the data. N = 840 men and 760 women. Days with equal weights
Men Women
Mean St. dev. Mean St. dev. Low High
Income work 349 243 272 223 0 1190
Household work 135 101 182 98 0 625
Housework 55 50 109 68 0 430
Maintenance work 34 62 16 37 0 470
Shopping / travelling 37 46 44 46 0 320
Leisure time 341 158 332 150 0 930
Active leisure 81 104 75 93 0 730
Social leisure 91 89 109 92 0 600
Resting leisure 153 98 132 78 0 630
Living alone 19% 17% 0 1
One child 19% 26% 0 1
2/3 children 36% 39% 0 1
Child below 7 years 33% 40% 0 1
Child above 20 years 8% 10% 0 1
Care-time children 0.3 0.5 0.5 0.8 0 6.8
Age (?40/10) 0.1 1.0 0.0 0.9 -1.8 1.9
Education (years) 12.7 2.7 12.8 2.7 9 19
Income per hour 19.4 6.3 16.9 5.2 3.9 51.8
Weekend day 30% 29% 0 1
Physical strains 1.3 0.8 1.2 0.6 0 4
Low autonomy 1.6 0.6 1.7 0.5 0 4
Monotonous work 1.1 0.7 1.0 0.6 0 4

Appendix 2 Three kinds of work strain were estimated

Physical strain:
(i) Do you daily have to lift more than 20 kg in your work, and, if so, how often (more
20 times, 5-19 times, 1-4 times, no)?
(ii) How do you rate the risk of being injured from heavy loads/work (large, medium, s

Both items were first given values from 0 (no strains) to 2 (high strains) with intermed
values of 1.5 and 1 for item (i) and 1 for item (ii). However, as no occupations average m
than 1.33 in the first item, this item was re-scaled upwards with a factor of 1.5 in the ph
strains index in order to make an index from 0 (no strains) to 4 (high strains).

Low autonomy:
(i) To what extent can you decide on the planning of your work?
(ii) To what extent can you decide on deadlines and the speed of your work?

Alternatives for both items: 'Decide to a high degree' (1), 'to some degree' (2), 'to a
degree' (3). Since no occupations averaged 'to a high degree' on both items, the sum
was re-scaled with a range from 0 (originally 2.5) to 4 (originally 6), with high values fo
autonomy.

318

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Blekesaune: Working Conditions and Time Use

Monotonous work (one item):


(i) Does your work consist of repetitive motions, so that you have to do the same things hour
by hour?
Alternatives: 'Nearly all the time' (4), 'about % of the time' (3), 'about lA the time' (2), 'about Va
of the time' (1), 'rarely or never' (0).

Note
1. The data were made available by Norwegian Social Science Data Services in Oslo. NSD is not respon
sible for the use made of the data material in this article.

References
Becker, G. S. (1965) 'A Theory of the Allocation of Time7, The Economic Journal 75: 493-517.
Becker, G. S. (1991) A Treatise on the Family. Cambridge, MA: Harvard University Press.
Berk, S. F. (1985) The Gender Factory: The Apportionment of Work in American Households. New York: Plenum
Press.
Bird, CE. and Ross, CE. (1993) 'House worker s and Paid Workers: Qualities of Work and Effects of
Personal Control', Journal of Marriage and the Family 55: 913-25.
Chirikos, T. N. and Nestel, G. (1991) 'Occupational Differences in the Ability of Men to Delay Retirement7,
Journal of Human Resources 26: 1-26.
Cunningham, M. (2001) 'Parental Influences on the Gendered Division of Housework7, American Socio
logical Review 66: 184-203.
Eshoj, P., Jepsen, J. R. and Nielsen, C. V. (2001) 'Long-Term Sickness Absence - Risk Indicators among
Occupationally Active Residents of a Danish County', Occupational Medicine-Oxford 51: 347-53.
Fagan, C. and Burchell, B. J. (2002) Gender, Jobs and Working Conditions in the European Union. European
Foundation for the Improvement of Living and Working Conditions. Luxembourg: Office for Official Publi
cations of the European Communities. Available at: www.eurofound.eu.int/publications/files/
EF0249EN.pdf
Filer, R. K. (1989) 'Occupational Segregation, Compensating Differentials, and Comparative Worth', in
R. T. Michael, H. I. Hartmann and B. O'Farrell (eds) Pay Equity: Empirical Inquiries, pp. 153-71.
Washington: National Academy Press.
Filer, R. K. and Petri, P. A. (1988) 'A Job-Characteristics Theory of Retirement', Review of Economics and
Statistics 70: 123-8.
Gershuny, J. (1988) Changing Times: The Social Economics of Post-Industrial Societies: A Report to the Joseph
Rowntree Memorial Trust. University of Bath.
Gershuny, J. (2002) Changing Times: Work and Leisure in Post-Industrial Society. Oxford: Oxford University
Press.
Hochschild, A. (with Anne Machung) (1989) The Second Shift: Working Parents and the Revolution at Home.
London: Piatkus.
Karasek, R. A. Jr. (1976) The Impact of the Work Environment on Life Outside the lob. A Study of the Relation
ship Between lob Content and Leisure Activity: Political Participation and Mental Strain Using National
Survey Data. Cambridge, MA: Massachusetts Institute of Technology.
Killingsworth, M. R. (1983) Labor Supply. Cambridge Surveys of Economic Literature. Cambridge: Cambridge
University Press.
Lancaster, K. (1966a) 'A New Approach to Consumer Theory', Journal of Political Economy 72: 132-57.
Lancaster, K. (1966b) 'Change and Innovation in the Technology of Consumption', American Economic
Review 56: 14-23.
Linder, S. B. (1970) Den Rastl?sa V?lf?rdsm?nniskan: Tidsbrist i ?verfl?d - en Ekonomisk Studie. Stockholm:
Bonnier s.
Meissner, M. (1971) 'The Long Arm of the Job: A Study of Work and Leisure', Industrial Relations 10:
239-60.

319

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms
Acta Sociol?gica 48(4)

Parker, S. (1971) The Future of Work and Leisure. London: MacGibbon & Kee. Later revised as: Parker, S.
(1983) Leisure and Work.
Peters, H. E. (1995) 'An Economic Approach to the Study of Child Well-Being: Gary Becker on Altruism
and Household Production', Journal of Family Issues 16: 587-608.
Presser, H. B. (1994) 'Employment Schedules Among Dual-Earner Spouses and the Division of
Household Labor by Gender', American Sociological Review 59: 348-64.
Quinn, J. F. (1977) 'Microeconomic Determinants of Early Retirement: A Cross-Sectional View of White
Married Men', Journal of Human Resources 12: 329-46.
Quinn, J. F. (1978) 'Job Characteristics and Early Retirement', Industrial Relations 17: 315-23.
Ross, C. E. and Mirowsky, J. (1996) 'Economic and Interpersonal Work Rewards: Subjective Utilities of
Men's and Women's Compensation', Social Forces 75: 223-45.
Ross, C. E. and Wright, M. P. (1998) 'Women's Work, Men's Work, and the Sense of Control', Work and
Occupations 25: 333-55.
Shaw, S. M. (1985) 'The Meaning of Leisure in Everyday Life', Leisure Sciences 7: 1-24.
Siegenthaler, K. L. and Vaughan, J. (1998) 'Older Women in Retirement Communities: Perceptions of
Recreation and Leisure', Leisure Sciences 20: 53-66.
Staines, G. L. (1980) 'Spillover Versus Compensation: Review of the Literature on the Relationship
between Work and Non-Work', Human Relations 33: 111-29.
Wilson, J. and Musick, M. A. (1997) 'Work and Volunteering: The Long Arm of the Job', Social Forces 76:
251-72.

Biographical Note: Morten Blekesaune is Chief Research Officer at Institute for Social & Economic
Research, University of Essex, having joined the institute from Norwegian Social Research (NOVA)
where this paper was written and where he also wrote his PhD on family transitions and time use. He
has also studied retirement and ageing, and take-up of social insurance/security benefits.
Address: Institute for Social & Economic Research, University of Essex, Colchester, C04 3SQ, UK.
[email: mblek@essex.ac.uk]

320

This content downloaded from 177.205.236.153 on Mon, 07 May 2018 19:00:56 UTC
All use subject to http://about.jstor.org/terms

You might also like