- Email: [email protected]
The emotional costs of caring incurred by men and women in the British labour market
ARTICLE IN PRESS
Social Science & Medicine 65 (2007) 2160–2171 www.elsevier.com/locate/socscimed
The emotional costs of caring incurred by men and women in the British labour market$ David Nicholas Barrona,, Elizabeth Westb a
Said Buisness School, University of Oxford, Park End Street, Oxford OX1 1HP, UK School of Health and Social Care, University of Greenwich, Avery Hill Road, London SE9 2UG, UK
b
Available online 9 August 2007
Abstract This study investigates whether men and women in caring occupations experience more negative job-related feelings at the end of the day compared to the rest of the working population. The data are from Wave Nine of the British Household Panel Survey (1999) where respondents were asked whether, at the end of the working day, they tended to keep worrying or have trouble unwinding, and the extent to which work left them feeling exhausted or ‘‘used up.’’ Their responses to these questions were used to develop ordinal dependent variables. Control variables in the models include: number of children, age, hours worked per week, managerial responsibilities and job satisfaction, all of which have been shown in previous research to be significantly related to ‘‘job burnout.’’ The results are that those in caring occupations are more likely to feel worried, tense, drained and exhausted at the end of the working day. Women in particular appear to pay a high emotional cost for working in caring occupations. Men do not emerge unscathed, but report significantly lower levels of worry and exhaustion at the end of the day than do women. r 2007 Elsevier Ltd. All rights reserved. Keywords: UK; Emotional labour; Burnout; Stress; Caring professions; Gender
Introduction Caring work involves providing another person with a service in return for pay. Workers in caring occupations are likely to be in close, often face-to$ The data used in this article were made available through the UK Data Archive. The data were originally collected by the ESRC Research Centre on Micro-social Change at the University of Essex, now incorporated within the Institute for Social and Economic Research. Neither the original collector of the data nor the Archive bears any responsibility for the analyses or interpretations presented here. Corresponding author. Tel.: +44 1865 288906; fax: +44 1865 288805. E-mail addresses: [email protected] (D.N. Barron), [email protected] (E. West).
0277-9536/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.socscimed.2007.06.027
face contact with the recipients of their services for long periods of their working day. They are required to care about the person who comes to them for help or support, and to act in their best interests. Caring work demands a variety of skills but is often performed in the context of a relationship that may be brief or enduring. Hence the importance of social skills in managing and maintaining the interpersonal relationship within which the work is conducted. The concept of emotional labour, where employees are required to manage their own emotions and the feelings of others (England & Farkas, 1986; Hochschild, 1983), is relevant here. It is perhaps unsurprising, then, that the bulk of the research on the phenomenon of work-related burnout, which
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
describes a descending spiral of exhaustion, detachment and underachievement, has been conducted on workers in caring occupations (Maslach, Schaulfeli, & Leiter, 2001). Typical occupations with high caring components include nursing, teaching and social work (Cancian & Oliker, 2000). In the past, caring for others was often performed at home, but such work has increasingly shifted into the public sphere. The identification of caring with women may account, at least in part, for the finding that jobs with a high caring component offer low pay relative to the qualifications and skills of incumbents (Kilbourne, Farkas, Beron, Weir, & England, 1994). England and Folbre (1999) argue that the apparent financial penalties incurred by workers in caring occupations may be due to the fact that carers are often women, which raises cultural barriers to recognising the value, particularly the financial value, of this work. There may also be intrinsic rewards to caring work, which mean that workers who value relationships and find meaning in giving to others still want to do the work even though they could earn more in other sectors. This study focuses on the potential emotional, as opposed to financial costs that workers might incur as a result of working in a job where caring for others is the central concern. In this paper, we explore whether employees whose jobs are predominantly designed to serve the needs of others report higher levels of negative job related feelings at the end of the working day than do workers whose jobs demand primarily manual or intellectual labour. We also ask whether gender plays a role in shaping negative feelings that arise as a consequence of working in a caring occupation in contemporary Britain. Theoretical Framework Occupations and health Work is one of the most important determinants of social status in our society and higher social status and better physical and mental health generally go together (Mirowsky, Ross, & Reynolds, 2000). Large differences in the health status of employed and unemployed people have been detected, and it has long been known that being employed improves the likelihood of good physical and mental health for both men and women (Gove & Tudor, 1973; Kessler, House, & Turner, 1987).
2161
A recent report by the European Union showed that while the principal cause of the decline in mortality since the Second World War is the growth in per capita wealth, unemployment rates exert an independent and damaging effect on national health. In particular, the upward trend in the number of working women is related to decreased mortality across Europe (EU, 2002). Too much work, however, can have a detrimental effect on health. Overwork, as well as unemployment has been linked to psychological distress (Losocco & Spitze, 1990) and to higher rates of morbidity and mortality (House, Strecher, Metzner, & Robbins, 1986). Some job characteristics have also been associated with negative outcomes for workers. For example, simple routine jobs that are closely controlled by management are more likely to cause psychological distress than jobs that are more challenging and autonomous (Reynolds, 1997). Link, Lennon, and Dohrenwend (1993) demonstrated that workers in occupations where direction, control and planning are restricted are more prone to depression. Cheng, Kawachi, Coakley, Schwartz, and Colditz (2000) examined the relation between psychosocial work characteristics and changes in health-related quality of life (as measured by a variant of the SF36) over 4 years in a cohort of nurses working in the United States. The study tested the job strain model (Karasek & Theorell, 1990), which claims high psychological demands combined with low job control at work leads to mental and physical illness. The findings confirmed the importance of low control and high demands. In addition, low levels of support at work were also implicated. This triad of variables was associated with poorer health at the start of the study as well as greater functional decline over its 4-year course, even when a range of factors such as smoking and weight that are known to pose a significant risk to health were controlled. Another influential epidemiological model of the relationship between work and stress focuses on the balance between efforts and rewards, drawing on sociological arguments about the importance of reciprocity, fairness and distributive justice (Siegrist, 2001). In this model, job rewards are encapsulated in money, esteem and career opportunities. An imbalance in costs and rewards leads individuals to experience a loss of self-esteem, whereas balance leads to feelings of satisfaction and enhanced selfworth. A recent study (de Jonge, Bosma, Peter, & Siegrist, 2000) examined the effects of the demand
ARTICLE IN PRESS 2162
D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
control model (described above) and the effort reward imbalance models on employee well-being in the Netherlands. As predicted, these variables were associated with increased risk of experiencing emotional exhaustion, psychosomatic and physical health complaints and job dissatisfaction. High psychological and physical effort and low rewards (poor salary, job insecurity and low work support) had particularly strong adverse effects on employee well-being. This study also found that people who were overcommitted to work (defined as excessive striving for success combined with a deep need for approval (Siegrist, 2001)) had higher risks of poor health when efforts and rewards were unbalanced. The results of this study were not significantly different for men or women. However, gender has been an important variable in other studies of the relationship between work and stress, as we will describe below. Gender and work-related health outcomes Several studies indicate that being in the labour market has a beneficial effect on women’s psychological well-being (Aneshensel, Frerichs, & Clark, 1981; Gore & Mangione, 1983; Weich, Sloggett, & Lewis, 1998), but may cause additional burdens as well, particularly for parents (Rosenfield, 1989; Sogaard, Kritz-Silverstein, & Wingard, 1994). The effects of women’s labour force participation may vary by social class (Waldron & Jacobs, 1988) or by quality of their home environment (Warr & Parry, 1982). Recently Matthews et al. (1998) showed that even after taking into account women’s multiple roles (employment, marital status, domestic responsibility and parental status), work characteristics (psychological job strain, insecurity, unsociable working hours) and home characteristics (age of youngest child, number of children, childcare responsibilities and an older adult living in the family), there is still a social class gradient in women’s experience of psychological distress. Women would seem to be particularly vulnerable to the adverse effects of work. The gender gap in wages is well documented (England, 1992; Kilbourne et al., 1994). Women occupy lower status jobs than men and are more likely to have careers that are punctuated by career breaks. This means that they are less likely to have savings, investments or an adequate pension when they retire (Warren, Rowlinson, & Whyley, 2001). Despite their increasing numbers in the work place, women still occupy
less prestigious jobs than men. Women are less likely to be managers or to occupy positions of authority (Rosenfeld, Van Buren, & Kalleberg, 1998). Karasek and Theorell (1990) found that women experienced lower levels of decision latitude or control than men, while experiencing the same psychological demands. Women reported more negative work characteristics than men, such as fewer learning opportunities and monotony, in a recent study by Matthews et al. (1998). Parts of the labour market are still highly segregated by gender and some studies have investigated whether working in a sex atypical occupation has negative psychological consequences. Evans and Steptoe (2002), for example, found that female accountants were more likely to report anxiety than other groups of workers, and male nurses had higher rates of sickness absence. The authors interpret the results as suggesting that when men and women occupy jobs in which they are the cultural and numerical minority, they may experience adverse health effects that are gender specific. This work is relevant to our study because so many of the caring professions are predominantly female. Extrapolating from these results might suggest that men in caring occupations would be more susceptible to negative feelings than women. It has been argued that even when men and women perform the same jobs they are likely to have very different experiences of work. Williams (1995) conducted a qualitative study of men in traditionally female occupations: nurses, social workers, librarians and primary school teachers. She found that men tended to receive preferential treatment in hiring, were channelled into certain male identified specialties, and pressured to perform specific tasks that are associated with men. These tasks and specialties tended to be highly valued, so that men appear to benefit from being in the minority in terms of advancement in femaledominated careers. Because gender assumptions about work are seldom challenged, women’s lack of career success can be attributed to individual failings, to their collective inability to be competitive, or to their competing loyalties to home and family. This work might suggest that while men are often sex atypical in caring jobs, the fact that they are on a ‘‘glass escalator’’ would militate against any negative job-related feelings (Williams, 1995). In general, the literature linking work, gender and negative psychosocial outcomes supports the
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
hypothesis that women will experience more jobrelated negative feelings, with some evidence to the contrary. The literature also suggests, however, that in order to detect the effects of working in caring jobs we need to control for a large number of other variables that have been shown to be important contributors to the negative effects of work. These will include individual sociodemographic characteristics, job characteristics, home responsibilities and the amount of time devoted to work. Emotional labour Caring work involves many different skills. Nursing, for example, involves manual or physical work (lifting patients), managerial responsibility (co-ordinating the work of a team), technical skill (giving an injection) and intellectual work (assessing patients’ needs and planning care). However, the skills that are held to be most important to nursing are social and relational, including communication, co-ordination of people and activities, and the management of relationships with patients and other professionals. Nursing theory has focused on concepts such as empathy (Peplau, 1988), intimacy (Savage, 1995) and caring (Benner & Wrubel, 1989). More recently, the concept of emotional labour has been investigated in relation to nursing (James, 1989; Smith, 1992; Steinberg & Walter, 1992) as one example of a caring occupation. In a recent review, Steinberg and Figart (1999) describe two main approaches for defining emotional labour. Hochschild’s (1983) original statement emphasised the work that is involved in performing emotional labour, particularly the requirement that the worker display or suppress emotions, whereas others have focused on the effect that the emotional labour is designed to have on others. England and Farkas (1986) define emotional labour in terms of ‘‘efforts made to understand others, to have empathy with their situation, to feel their feelings as part of one’s own’’ (p. 91). The latter definition shifts the focus from the emotion displayed by the employee to the feelings of the client or recipient of services. This distinction is, of course, somewhat arbitrary because in order to manage the feelings of others, one must first be able to manage one’s own feelings. However, in this context it is a useful distinction because the latter definition is more consonant with the theory and philosophy of caring occupations such as medicine, education and social work, where the explicit focus
2163
is on the welfare and development of the recipient. Workers in these occupations are socialised to show genuine feelings of concern for the recipients of their service, and to display authentic, but controlled, rather than engineered emotion. Another way of delineating different kinds of occupations is by their relationship to the market. For many jobs that involve emotional labour, the commercial element of the transaction is obvious. Waiters and waitresses (Hall, 1993), hairdressers (Parkinson, 1996) and supermarket cashiers (Sutton & Rafaeli, 1988) are involved in an immediate exchange of money for goods and services. In some cases, the underlying economic nature of the relationship is often underscored by the tradition of tipping. Hochschild’s (1983) study of flight attendants showed that their training emphasises the commercial consequences of their behaviour. In caring relationships, market forces may be seen as part of the environment but not immediately relevant to the business in hand, particularly in systems set up to provide services that are ‘‘free at the point of delivery.’’ In sum, caring jobs as defined in this paper differ from many other jobs that involve emotional labour on a number of dimensions. Employees in caring occupations are required to focus on the feelings of their client, and to experience caring and compassion for them. The connection with the marketplace, while not entirely absent, is less pervasive than in many occupations. Finally, workers in caring occupations, which often require several years of education and training, are more likely to be of higher social class than the majority of the workforce. Studies that have investigated the effects of emotional labour on workers have focused on two main dependent variables: job satisfaction and pay. Wharton (1993) for example, gathered data from employees in a bank and a hospital and found that those in jobs that were considered to be high in emotional labour were no more prone to emotional exhaustion than other workers. Characteristics of the jobs themselves such as the level of autonomy and involvement were much more important in determining whether or not the worker experienced emotional exhaustion. More recently, Wharton (1999) has argued, using evidence from studies of the economic and non-economic costs and rewards both at work and at home, that although there are negative psychological consequences under certain circumstances, jobs requiring emotional labour can also have a positive effect on workers’ well-being.
ARTICLE IN PRESS 2164
D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
This work would seem to predict that we would find little clear evidence that working in caring occupations would have uniformly negative emotional consequences for workers. Burnout Burnout is a response to chronic interpersonal stress on the job (Freudenberger, 1975; Maslach, 1976). Since it was first described, three components have been identified: exhaustion, cynicism (or emotional withdrawal) and a sense of underachievement at work. Maslach et al. (2001) provide a critical review of what has been learnt about burnout over the last 25 years. Although there do appear to be three distinct components, exhaustion seems to be of primary importance with the other two features of the syndrome following almost in stages. A study of hospital nurses, for example, described the following sequence of events (Leiter & Maslach, 1998): (a) Stressful interactions with supervisors increase workers’ feelings of exhaustion. (b) High levels of exhaustion lead to cynicism, especially in the absence of support from coworkers. (c) Persistent cynicism leads to diminished feelings of efficacy, which can be ameliorated by support from co-workers. This study, in addition to describing the possible sequential relationship of the three components of burnout, reveals how important supervisors and coworkers can be in ameliorating the stresses and strains of work and in interrupting the process that leads to burnout. Burnout has been linked to a number of important job-related outcomes, such as absenteeism and turnover, and has been shown to have a detrimental effect on an individual’s social as well as work life. Some socio-demographic characteristics seem to be predisposing factors for the development of burnout, including age, gender, marital status and level of education. Younger people appear to be at greater risk but this finding may be an artefact of selection bias. If young people who become burned out change their jobs or leave the labour market, then long-serving employees may appear to be protected from burnout. But, long-serving employees may differ from those who leave in some important ways. They may, for example, have
invested less in their jobs from the beginning of their careers and so are less liable to disillusionment, or they may develop better coping mechanisms over time that protect them from burnout. Gender is also relevant, with men tending to score more highly on the second factor (cynicism) and women being slightly more likely to report higher levels of exhaustion. Previous research shows that being single (even more than being divorced), and having more education are also predisposing factors. Perhaps people who do not have an immediate family focus more on work and are less likely to have competing worries or compensating sources of emotional satisfaction. Those with higher levels of education may simply occupy more demanding and stressful jobs or it could be that having invested so much time and effort in their education, they expect more from their jobs, and are therefore more vulnerable to disillusionment. As well as individual characteristics, some job characteristics have been associated with burnout. These include; heavy workloads, time pressures, role conflict, lack of social support (particularly from supervisors and co-workers) and a lack of information or feedback. Low levels of participation in decision making and autonomy have also been implicated in burnout. Currently, Maslach et al. (2001) are focusing on six areas distilled from their previous research: workload, control (especially over resources needed to do the job) rewards (finance, status, appreciation), community (a sense of belonging at work, or conflict), fairness (inequity of workload or pay), and values (conflict between own and organisation’s values, or conflict between stated values and those that are enacted). Some recent research speaks directly to the relationship between emotional labour and burnout, which is close to some of the research questions we hope to address. Zapf, Seifert, Schmutte, Mertini, and Holz (2001) found that emotional labour explains at least part of the variance in burnout scores over and above job stress. Schaulfeli and Enzmann (1998) compared workers in the US and the Netherlands and found that while the basic patterns of burnout were similar across the five occupational groups (teaching, social services, medicine, mental health and law enforcement) that they studied, there were some interesting national differences. This suggests that characteristics of the setting in which work is performed, such as the way the work is organised, may also make employees
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
more prone to burnout. Increasing our understanding of these contextual variables will enable us to devise changes that will improve working conditions. In the meantime there are some precedents for research on the link between emotional labour, caring work and burnout. Research questions This study investigates the impact of caring work and gender on four outcomes: having difficulty unwinding and ceasing to worry after work and the extent to which workers feel exhausted or ‘‘used up’’ at the end of their shift. We ask: 1. Are workers in caring jobs more likely to report negative job-related feelings at the end of the working day than workers in other jobs, controlling for other factors that might contribute to feelings of stress and exhaustion? 2. Are women more likely to report negative jobrelated feelings at the end of the working day than men, controlling for all the other factors that might contribute to feelings of stress and exhaustion? Data The data for these analyses are taken from the British Household Panel Survey (BHPS) (Taylor, Brice, Buck, & Prentice-Lane, 2006). This survey has been undertaken annually since 1991 by the Institute for Social and Economic Research at the University of Essex. The BHPS involves an annual survey of each adult member of a nationally representative sample of more than 5000 households. This gives over 10,000 individual interviews each year. In each wave, interviewees are asked a wide range of questions about their employment status, including their occupation, pay, hours, and job satisfaction. In addition, questions about particular areas of sociological interest are added each year. For the purposes of this study, we restrict our analysis to those members of the BHPS sample that are either employed or self-employed, giving a sample of over 6500 people. Dependent variables In one single year, 1999, respondents to the annual BHPS survey were asked a series of supplementary questions about their health and
2165
well-being, including: ‘‘Thinking of the past few weeks, how much time has your job made you feel each of the following?’’ After I leave my work I keep worrying about job problems. I find it difficult to unwind at the end of the work day. I feel used up at the end of the work day. My job makes me feel quite exhausted by the end of the work day. Response categories ran from 1 (never) to 6 (all of the time), with categories for ‘‘missing,’’ ‘‘not applicable,’’ ‘‘proxy respondent’’ and ‘‘don’t know,’’ all of which are excluded in our analyses. Frequency distributions of these variables are shown in Table 1. These are the main outcome variables in this study. We describe in the Methods section below how they were used in the statistical analysis. Independent variables Caring occupations We classified all the occupations in the BHPS into two categories based on our knowledge of each occupation, taking into consideration the amount of interpersonal interaction involved every day and the extent to which their interactions are conducted in the context of a relationship and might be emotionally laden. We tried to capture the demands that the worker in one of these jobs will feel to care about the recipients of their services and to focus on their feelings, well-being or development, even at times at the expense of their own. The list of caring jobs is in Appendix A, which also gives the number of people in each occupational category in the data set. The majority of people in caring occupations work in health-related occupations, education and social work. The list of caring jobs that we used in this analysis turned out to be very similar to the list of jobs coded by England (1992) as scoring high on ‘‘nurturance.’’ Control variables These are included in the model to control for other aspects of the job or the individual that might explain negative job-related feelings irrespective of the type of job (caring/non-caring) or the gender of the worker. They include: overwork, caring for a sick or disabled person at home, sociodemographic
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
2166
Table 1 Tables showing cross-tabulations of being in a caring occupation and sex with the four outcome variables (column percentages in parentheses) Caring Yes Men
No Women
Men
Women
Worry Never Rarely Some of the time Much of the time Most of the time All of the time
53 53 36 21 14 6
(29.0) 337 (36.3) 1883 (46.0) 1529 (48.5) (29.0) 273 (29.4) 1077 (26.3) 808 (25.6) (19.7) 200 (21.5) 682 (16.6) 515 (16.3) (11.5) 63 (6.8) 229 (5.6) 148 (4.7) (7.6) 35 (3.8) 150 (3.7) 110 (3.5) (3.3) 21 (2.3) 74 (1.8) 40 (1.3)
Used up Never Rarely Some of the time Much of the time Most of the time All of the time
24 53 55 19 26 6
(13.1) (29.0) (30.0) (10.4) (14.2) (3.3)
165 236 256 111 121 39
(17.8) 893 (21.8) (25.4) 1335 (32.6) (27.6) 1029 (25.1) (12.0) 357 (8.7) (13.0) 356 (8.7) (4.2) 122 (3.0)
Exhausted Never Rarely Some of the time Much of the time Most of the time All of the time
28 59 41 26 24 5
(15.3) (32.2) (22.4) (14.2) (13.1) (2.7)
148 238 250 129 118 45
(15.9) 857 (20.9) 720 (22.9) (25.6) 1398 (34.2) 1032 (32.8) (26.9) 982 (24.0) 743 (23.6) (13.9) 363 (8.9) 269 (8.5) (12.7) 376 (9.2) 270 (8.6) (4.8) 117 (2.9) 116 (3.7)
Unwind Never Rarely Some of the time Much of the time Most of the time All of the time
51 56 37 20 16 3
(27.9) 302 (32.5) 1709 (41.7) 1288 (40.9) (30.6) 251 (27.0) 1210 (29.5) 893 (28.4) (20.2) 233 (25.1) 715 (17.5) 627 (19.9) (10.9) 67 (7.2) 213 (5.2) 150 (4.8) (8.7) 58 (6.2) 168 (4.1) 135 (4.3) (1.6) 18 (1.9) 81 (2.0) 55 (1.7)
724 987 790 270 268 107
Methods The first step in analysing these data was to investigate the structure of the four outcome variables. We first calculated all the pairwise polychoric correlations among these variable using the polycor package in R (Fox, 2006a; R Development Core Team, 2006). These correlations were then used as the basis for confirmatory factor analysis using the sem package in R (Fox, 2006b). First, we tested a model with a single latent factor; the results are shown in Fig. 1. As can be seen, this model does not provide a good fit to the data (w2 ¼ 2654.2, degrees of freedom ¼ 2, GFI ¼ .87). However, a two factor model shown in Fig. 2, fits
(23.0) (31.4) (25.1) (8.6) (8.5) (3.4)
characteristics (age, having a degree, marital status) and job characteristics (hours worked, overtime, working rotating shifts, liking the hours one works, job satisfaction, hourly pay, management responsibilities). Based on the finding of a social class gradient in the literature on gender, work and stress we used the Registrar General’s classification (where 1 ¼ professional and 5 ¼ unskilled worker) as a set of dummy variables. The inclusion of these variables as controls is supported theoretically by the literature reviewed on work and stress, emotional labour and burnout. We also included two variables, not having support in a crisis and not feeling appreciated which could be relevant to respondents’ emotional responses to work and a measure of general health.
Fig. 1. Confirmatory factor analysis with a single latent factor.
Fig. 2. Confirmatory factor analysis with two latent factors.
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
the data very well (w2 ¼ .425, degrees of freedom ¼ 1, GFI 4.999). One of the two factors loads on the two variables that relate to tiredness (‘‘I feel used up at the end of the work day’’ and ‘‘My job makes me feel quite exhausted by the end of the work day’’), while the other loads on the variables that relate to worry and stress. As a result of this, we combined these two pairs of variables into two indicators by summing the respective observed variables. The frequency distributions of these variables are shown in Fig. 3. Higher scores mean that a person feels more exhausted or more stressed. Findings In Table 2 we show linear regression estimates of effects on exhaustion and stress, the two composite outcome variables described above. The main explanatory variables of interest are working in a caring occupation and gender. The positive and significant coefficient for caring work suggests that nurses, doctors, teachers and social workers are more likely to feel exhausted and stressed, or to feel exhausted and stressed more frequently, than people in other jobs. Women are also much more likely to
2167
be exhausted and stressed than are men. These key variables are statistically significant after taking account of a number of important controls. It is worth noting that these findings relate to how much the respondents feel stressed when they return home from work, over and above the amount of worry they might experience ‘‘on the job.’’ Some individual characteristics, such as being older, seem to decrease the propensity to exhaustion and stress, while having a degree increases the chances of being stressed. Poor general health also increases the tendency to be tired and stressed after work. Some characteristics of the job seem to promote exhaustion and stress, including number of hours worked per week, overtime and higher pay. Work related factors that decrease problematic outcomes, including not having managerial responsibilities, having high job satisfaction, being able to work the hours that one prefers and not having promotion opportunities, are all negative and significant in this model. Taken together they suggest that people who are higher in the hierarchy might suffer more problems than people who are in non-managerial positions. Management also involves a degree of emotional labour because
Fig. 3. Frequency distributions of the two composite dependent variables.
ARTICLE IN PRESS 2168
D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
Table 2 Linear regression estimates of effects on worrying about work and difficulty unwinding after work Tired
Stressed
p-value ¼ .403), but there is between class and stress (F-statistic ¼ 30.5, p-value 5.05). The results show that people in lower skilled and manual occupations tend to have lower levels of stress than those in professional and managerial occupations.
Estimate t-value Estimate t-value Constant Caring occupation Female Single Number of children Age Carer Has a degree Hours worked per week Log monthly pay Not a manager No one to turn to in crisis Not appreciated Job satisfaction Hours spent on housework per week General health Works shifts Overtime hours per week Likes hours worked No promotion opportunities RG Class 2 RG Class 3a RG Class 3b RG Class 4 RG Class 5 Number of cases R-square
4.38* 0.638* 0.441* 0.031 0.037 0.011* 0.201* 0.032 0.034* 0.036 0.211* 0.103 0.139 0.234* 0.0004
8.39 6.64 5.86 0.464 1.12 4.05 2.30 0.373 7.53 .480 3.01 .812 1.31 9.77 .102
2.93* 0.421* 0.268* 0.062 0.008 0.011* 0.099 0.300* 0.020* 0.121* 0.46* 0.166 0.090 0.161* 0.002
6.57 5.29 4.35 1.16 0.303 4.81 1.39 3.99 5.10 1.80 7.93 1.49 0.989 8.05 0.492
0.133* 19.3 0.123 1.16 0.058* 11.2 0.579* 8.74 0.008 0.136
0.125* 0.003 0.046* 0.357* 0.112*
20.9 0.034 10.2 6.56 2.28
0.081 0.181 0.705* 0.770* 0.585*
0.723 1.49 5.74 5.96 3.76
0.188 0.010 0.118 0.052 0.084 6539 0.200
1.47 0.072 0.825 0.334 0.416
6542 0.27
t-Values are calculated from White-corrected estimated standard errors. Note: *po0.05.
managers are responsible for many aspects of workers’ welfare as well as their productivity, and managers in caring occupations may have a double burden, responsible as they are for both clients and junior staff. A few variables appear not to contribute much to the risk of being exhausted or stressed: being single, having no one to turn to in a crisis, not feeling appreciated and hours spent in housework each week are all insignificant. The last variables in the model are the dummy variables for social class, included here to control for the known association between social class and psychological distress. There is no statistically significant association between class and exhaustion (F-statistic ¼ 1.02,
Discussion and conclusions Our results suggest that if you ask women in caring occupations how they feel at the end of the day, they will often respond ‘‘worried, tense, drained and exhausted.’’ Men in caring jobs also bear emotional costs, but the penalties do not appear to be as great as those experienced by female employees. Women working in caring jobs are more likely than men to experience all four of the negative feelings that we studied. Other variables that increased the likelihood of negative job-related feelings include being young, being a manager, experiencing low levels of job satisfaction and not liking the hours that one works. The number of hours worked and the number of overtime hours worked had separate, independent effects on the experience of negative affect after work. Workers who have poor general health are also much more vulnerable to experiencing negative work-related feelings. We argued that worrying about work or having trouble unwinding seem to be much more likely if the employee were in a managerial position. The positive effect of having a degree and higher pay, in addition to the negative effect of not being a manager contributed to this interpretation, as did the fact that having no promotion prospects seemed to reduce the tendency to worry after work. Interestingly, working shifts also appeared to reduce the tendency to worry, but increased the tendency to have trouble unwinding. Although ‘‘not being a manager’’ was negatively related to feeling exhausted and ‘‘used up’’ after work, there was no association with pay or having a degree. The only variable that is significant in the model of exhaustion while insignificant in the other is ‘‘being a carer.’’ Caring for another adult at home can lead to exhaustion, but does not, according to these estimates, lead to increased feelings of stress. This study has a number of limitations: the data are cross-sectional, which restricts our ability to make causal arguments, and the coding of occupational categories is into binary categories. Most jobs have an element of contact with other human beings, so it would be preferable to construct a
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
variable to measure the amount of interpersonal interaction or emotional labour typically associated with each occupation. However, the study does show that caring work has a negative impact on workers in a large national sample of over 6500 workers across all economic sectors in contemporary Britain. There is increasing sociological interest in caring work as it affects employees’ experience of work, career trajectories, the domestic division of labour and labour market segregation. Caring work is interesting theoretically because it challenges traditional notions of a division between public and private spheres that have prevailed since the Industrial Revolution. It is also of increased concern to policy makers, as the number of people employed in caring occupations is increasing and is likely to continue to increase with the demands of an aging population on health and social services. The goals of future work will be to refine the classification of occupations, and to extend the study over time so that we can find out whether negative work related feelings have a cumulative impact on workers health and job tenure. What are the long-term consequences of feeling worried, stressed, drained and exhausted on a regular basis? We would also like to find out whether there is something inherently stressful in caring for others, or whether caring work is organised in such a way that workers are exposed to some of the known causes of burnout. Do nurses, for example, have less autonomy and control over their work than people working in jobs that are similar in other respects? Are teachers less able to regulate the amount of face-to-face contact they have in a working day than people in other occupations? The organisational context of caring work may turn out to be more important than the content of the jobs themselves. Important questions remain to be answered about how managerial responsibilities affect employees, particularly those who are also responsible for clients. How does the emotional labour of managers differ, if at all from that of clinicians or teachers? Are the additional demands of managerial jobs mitigated by the fact that managers are rewarded with higher pay than other workers? We also need to find out why gender is so salient in these analyses. Do men and women in caring jobs behave differently, or are they treated differently? Or does the answer simply lie in the fact that they respond differently to questions about their feelings? There may still be some gender differences in the propensity to report negative emotions. Finally,
2169
we have focused in this paper on the emotional costs of caring, ignoring for the time being the possibility that caring for others might also have rewards. Future work on caring occupations will need to consider what the rewards might be, whether these can be measured, and whether the balance between the rewards and costs is consistent with ideals of social justice. Appendix A. List of caring occupations with number of individuals in each category in the data set
Code 220 222 223 224 230 233 234 235 292 293 340 341 342 343 344 345 347 363 370 371 640 641 644 650 651 652 659
Job title Medical practitioners Ophthalmic opticians Dental practitioners Veterinarians University teachers Secondary education teachers Primary & nursery education teachers Special education teachers Clergy Social workers & probation officers Nurses Midwives Radiographers Physiotherapists Chiropodists Dispensing opticians OT, speech therapy, etc. Personnel & IR officers Matrons & houseparents Welfare & community work Asst. nurses & nursing auxs. Hospital ward assistants Care assistants & attendants Nursery nurses Playgroup leaders Educational assts Other childcare occs.
N % 34 2.9 13 1.1 11 0.9 3 0.3 38 3.3 104 8.9 101 8.7 21 1.8 24 2.1 62 5.3 177 15.2 14 1.2 9 0.8 8 0.7 2 0.2 1 0.1 29 2.5 17 1.5 21 1.8 77 6.6 45 3.9 17 1.5 142 12.2 43 3.7 15 1.3 50 4.3 84 7.2
References Aneshensel, C. S., Frerichs, R. R., & Clark, A. V. (1981). Family roles and sex differences in depression. Journal of Health and Social Behaviour, 22(4), 379–393. Benner, P., & Wrubel, J. (1989). The primacy of caring: Stress and coping with health and illness. Menlo Park, CA: Addison-Wesley.
ARTICLE IN PRESS 2170
D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
Cancian, F. M., & Oliker, S. J. (2000). Caring and gender. Thousand Oaks: Pine Forge Press. Cheng, Y., Kawachi, I., Coakley, E. H., Schwartz, J., & Colditz, G. (2000). Association between psychosocial work characteristics and health functioning in American women: Prospective study. British Medical Journal, 320, 1432–1436. de Jonge, J., Bosma, H., Peter, R., & Siegrist, J. (2000). Job strain, effort reward imbalance and employee well-being: A large scale cross-sectional study. Social Science & Medicine, 50(9), 1317–1327. England, P. (1992). Comparable worth: Theories and evidence. New York: Aldine de Gruyter. England, P., & Farkas, G. (1986). Households, employment and gender: A social, economic and demographic view. New York: Aldine. England, P., & Folbre, N. (1999). The costs of caring. The Annals of the American Academy of Political and Social Science, 561, 39–51. EU. (2002). Unemployment, employment and public health. Brussels: European Union. Evans, O., & Steptoe, A. (2002). The contribution of gender-role orientation, work factors and home stressors to psychological well-being and sickness absence in male and female dominated occupational groups. Social Science & Medicine, 54, 481–492. Fox, J. (2006a). Polycor: Polychoric and polyserial correlations (R package, version 0.7-2). Fox, J. (2006b). Sem: Structural equation models (R package, version 0.9-6). Freudenberger, H. J. (1975). The staff burnout syndrome in alternative institutions. Psychotherapy: Theory, Research and Practice, 12, 72–83. Gore, S., & Mangione, T. W. (1983). Social roles, sex roles and psychological distress: Additive and interactive models of sex differences. Journal of Health and Social Behaviour, 24, 300–312. Gove, W. R., & Tudor, J. F. (1973). Adult sex roles and mental illness. American Journal of Sociology, 78(4), 812–835. Hall, E. J. (1993). Smiling, deferring, and flirting: Doing gender by giving ‘‘good service’’. Work and Occupations, 20(4), 452–471. Hochschild, A. R. (1983). The managed heart: The commercialisation of human feeling. Berkely: University of California Press. House, J. S., Strecher, V., Metzner, H., & Robbins, C. A. (1986). Occupational stress and health among men and women in the Tecumseh community health study. Journal of Health and Social Behaviour, 27, 62–77. James, N. (1989). Emotional labour: Skill and work in the social regulation of feelings. Sociological Review, 37, 15–42. Karasek, R., & Theorell, T. (1990). Healthy work: Stress, productivity and the reconstruction of working life. New York: Basic Books. Kessler, R. C., House, J. S., & Turner, J. B. (1987). Unemployment and health in a community sample. Journal of Health and Social Behaviour, 28, 51–59. Kilbourne, B. S., Farkas, G., Beron, K., Weir, B., & England, P. (1994). Returns to skill, compensating differentials, and gender bias: Effects of occupational characteristics on the wages of white men and women. American Journal of Sociology, 100(3), 689–719.
Leiter, M. P., & Maslach, C. (1998). The impact of interpersonal environment on burnout and organisational committment. Journal of Organisational Behaviour, 9, 297–308. Link, B. G., Lennon, M. C., & Dohrenwend, B. P. (1993). Socioeconomic status and depression: The role of occupations involving direction, control and planning. American Journal of Sociology, 98, 1351–1387. Losocco, K. A., & Spitze, G. (1990). Working conditions, social support, and well-being of female and male factory workers. Journal of Health and Social Behaviour, 31, 313–327. Maslach, C. (1976). Burned-out. Human Behaviour, 5, 16–22. Maslach, C., Schaulfeli, W. B., & Leiter, M. P. (2001). Job burnout. Annual Review of Psychology, 52, 397–422. Matthews, S., Hertzman, C., Ostry, A., & Power, C. (1998). Gender, work roles and psychosocial work characteristics as determinants of health. Social Science and Medicine, 46, 1417–1424. Mirowsky, J., Ross, C., & Reynolds, J. (2000). Links between social status and health status. In C. E. Bird, P. Conrad, & A. M. Fremont (Eds.), Handbook of Medical Sociology. New Jersey: Prentice-Hall. Parkinson, B. (1996). Changing moods: The psychology of mood and mood regulation. New York: Adison Wesley Longman. Peplau, H. E. (1988). Interpersonal relations in nursing. London: Macmillan Press Ltd. R Development Core Team. (2006). R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. Reynolds, J. R. (1997). The effects of industrial employment conditions on job-related distress. Journal of Health and Social Behaviour, 38, 105–116. Rosenfeld, R. A., Van Buren, M. E., & Kalleberg, A. L. (1998). Gender differences in supervisory authority: Variation among advanced industrialised democracies. Social Science & Medicine, 27, 23–49. Rosenfield, S. (1989). The effects of womens’ employment: Personal control and sex differences in mental health. Journal of Health and Social Behaviour, 30, 77–91. Savage, J. (1995). Nursing intimacy: An ethnographic approach to nurse patient interaction. London: Scutari Press. Schaulfeli, W. B., & Enzmann, D. (1998). The burnout companion to study and practice: A critical analysis. Philadelphia: Taylor & Francis. Siegrist, J. (2001). Work stress and health. In W. Cockerham (Ed.), The Blackwell companion to medical sociology (pp. 114–125). Oxford: Blackwell Publishers Limited. Smith, P. (1992). The emotional labour of nursing. London: Macmillan. Sogaard, A. J., Kritz-Silverstein, D., & Wingard, D. L. (1994). Finnmark heart study: Employment status and parenthood as predictors of psychological health in women, 20–49 years. International Journal of Epidemiology, 23, 82–90. Steinberg, R. J., & Figart, D. M. (1999). Emotional labor since the managed heart. Annals of the American Academy of Political and Social Science, 561, 8–26. Steinberg, R., & Walter, W. L. (1992). Making women’s work visible: The case of nursing—First steps in the design of a gender-neutral job comparison system. Washington, DC: Institute for Women’s Policy Research. Sutton, R. I., & Rafaeli, A. (1988). Untangling the relationship between displayed emotion and organisational sales: The case of convenience stores. Academy of Management Journal, 31(3), 461–487.
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171 Taylor, M. F., Brice, J., Buck, N., Prentice-Lane, E. (2006). British household panel survey user manual volume A: Introduction, technical report and appendices. Colchester: University of Essex. Waldron, I., & Jacobs, J. A. (1988). Effects of labour force participation on women’s health: New evidence from a longitudinal study. Journal of Occupational Medicine, 30(12), 977–983. Warr, P., & Parry, G. (1982). Paid employment and women’s psychological well-being. Psychological Bulletin, 91(3), 493–498. Warren, T., Rowlinson, K., & Whyley, C. (2001). Female finances: Gender wage gaps and gender assets gaps. Work, Employment and Society, 15(3), 465–488.
2171
Weich, S., Sloggett, A., & Lewis, G. (1998). Social roles and gender difference in the prevalence of common mental disorders. British Journal of Psychiatry, 173, 489–493. Wharton, A. (1993). The affective consequences of service work: Managing emotions on the job. Work and Occupations, 20(2), 205–232. Wharton, A. (1999). Emotional labour. Annals of the American Academy of Political and Social Science, 561. Williams, C. L. (1995). Still a man’s world: Men who do women’s work. Berkeley: University of California Press. Zapf, D., Seifert, C., Schmutte, B., Mertini, H., & Holz, M. (2001). Emotion work and job stressors and their effects on burnout. Psychology and Health, 16(5), 527–545.
Social Science & Medicine 65 (2007) 2160–2171 www.elsevier.com/locate/socscimed
The emotional costs of caring incurred by men and women in the British labour market$ David Nicholas Barrona,, Elizabeth Westb a
Said Buisness School, University of Oxford, Park End Street, Oxford OX1 1HP, UK School of Health and Social Care, University of Greenwich, Avery Hill Road, London SE9 2UG, UK
b
Available online 9 August 2007
Abstract This study investigates whether men and women in caring occupations experience more negative job-related feelings at the end of the day compared to the rest of the working population. The data are from Wave Nine of the British Household Panel Survey (1999) where respondents were asked whether, at the end of the working day, they tended to keep worrying or have trouble unwinding, and the extent to which work left them feeling exhausted or ‘‘used up.’’ Their responses to these questions were used to develop ordinal dependent variables. Control variables in the models include: number of children, age, hours worked per week, managerial responsibilities and job satisfaction, all of which have been shown in previous research to be significantly related to ‘‘job burnout.’’ The results are that those in caring occupations are more likely to feel worried, tense, drained and exhausted at the end of the working day. Women in particular appear to pay a high emotional cost for working in caring occupations. Men do not emerge unscathed, but report significantly lower levels of worry and exhaustion at the end of the day than do women. r 2007 Elsevier Ltd. All rights reserved. Keywords: UK; Emotional labour; Burnout; Stress; Caring professions; Gender
Introduction Caring work involves providing another person with a service in return for pay. Workers in caring occupations are likely to be in close, often face-to$ The data used in this article were made available through the UK Data Archive. The data were originally collected by the ESRC Research Centre on Micro-social Change at the University of Essex, now incorporated within the Institute for Social and Economic Research. Neither the original collector of the data nor the Archive bears any responsibility for the analyses or interpretations presented here. Corresponding author. Tel.: +44 1865 288906; fax: +44 1865 288805. E-mail addresses: [email protected] (D.N. Barron), [email protected] (E. West).
0277-9536/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.socscimed.2007.06.027
face contact with the recipients of their services for long periods of their working day. They are required to care about the person who comes to them for help or support, and to act in their best interests. Caring work demands a variety of skills but is often performed in the context of a relationship that may be brief or enduring. Hence the importance of social skills in managing and maintaining the interpersonal relationship within which the work is conducted. The concept of emotional labour, where employees are required to manage their own emotions and the feelings of others (England & Farkas, 1986; Hochschild, 1983), is relevant here. It is perhaps unsurprising, then, that the bulk of the research on the phenomenon of work-related burnout, which
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
describes a descending spiral of exhaustion, detachment and underachievement, has been conducted on workers in caring occupations (Maslach, Schaulfeli, & Leiter, 2001). Typical occupations with high caring components include nursing, teaching and social work (Cancian & Oliker, 2000). In the past, caring for others was often performed at home, but such work has increasingly shifted into the public sphere. The identification of caring with women may account, at least in part, for the finding that jobs with a high caring component offer low pay relative to the qualifications and skills of incumbents (Kilbourne, Farkas, Beron, Weir, & England, 1994). England and Folbre (1999) argue that the apparent financial penalties incurred by workers in caring occupations may be due to the fact that carers are often women, which raises cultural barriers to recognising the value, particularly the financial value, of this work. There may also be intrinsic rewards to caring work, which mean that workers who value relationships and find meaning in giving to others still want to do the work even though they could earn more in other sectors. This study focuses on the potential emotional, as opposed to financial costs that workers might incur as a result of working in a job where caring for others is the central concern. In this paper, we explore whether employees whose jobs are predominantly designed to serve the needs of others report higher levels of negative job related feelings at the end of the working day than do workers whose jobs demand primarily manual or intellectual labour. We also ask whether gender plays a role in shaping negative feelings that arise as a consequence of working in a caring occupation in contemporary Britain. Theoretical Framework Occupations and health Work is one of the most important determinants of social status in our society and higher social status and better physical and mental health generally go together (Mirowsky, Ross, & Reynolds, 2000). Large differences in the health status of employed and unemployed people have been detected, and it has long been known that being employed improves the likelihood of good physical and mental health for both men and women (Gove & Tudor, 1973; Kessler, House, & Turner, 1987).
2161
A recent report by the European Union showed that while the principal cause of the decline in mortality since the Second World War is the growth in per capita wealth, unemployment rates exert an independent and damaging effect on national health. In particular, the upward trend in the number of working women is related to decreased mortality across Europe (EU, 2002). Too much work, however, can have a detrimental effect on health. Overwork, as well as unemployment has been linked to psychological distress (Losocco & Spitze, 1990) and to higher rates of morbidity and mortality (House, Strecher, Metzner, & Robbins, 1986). Some job characteristics have also been associated with negative outcomes for workers. For example, simple routine jobs that are closely controlled by management are more likely to cause psychological distress than jobs that are more challenging and autonomous (Reynolds, 1997). Link, Lennon, and Dohrenwend (1993) demonstrated that workers in occupations where direction, control and planning are restricted are more prone to depression. Cheng, Kawachi, Coakley, Schwartz, and Colditz (2000) examined the relation between psychosocial work characteristics and changes in health-related quality of life (as measured by a variant of the SF36) over 4 years in a cohort of nurses working in the United States. The study tested the job strain model (Karasek & Theorell, 1990), which claims high psychological demands combined with low job control at work leads to mental and physical illness. The findings confirmed the importance of low control and high demands. In addition, low levels of support at work were also implicated. This triad of variables was associated with poorer health at the start of the study as well as greater functional decline over its 4-year course, even when a range of factors such as smoking and weight that are known to pose a significant risk to health were controlled. Another influential epidemiological model of the relationship between work and stress focuses on the balance between efforts and rewards, drawing on sociological arguments about the importance of reciprocity, fairness and distributive justice (Siegrist, 2001). In this model, job rewards are encapsulated in money, esteem and career opportunities. An imbalance in costs and rewards leads individuals to experience a loss of self-esteem, whereas balance leads to feelings of satisfaction and enhanced selfworth. A recent study (de Jonge, Bosma, Peter, & Siegrist, 2000) examined the effects of the demand
ARTICLE IN PRESS 2162
D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
control model (described above) and the effort reward imbalance models on employee well-being in the Netherlands. As predicted, these variables were associated with increased risk of experiencing emotional exhaustion, psychosomatic and physical health complaints and job dissatisfaction. High psychological and physical effort and low rewards (poor salary, job insecurity and low work support) had particularly strong adverse effects on employee well-being. This study also found that people who were overcommitted to work (defined as excessive striving for success combined with a deep need for approval (Siegrist, 2001)) had higher risks of poor health when efforts and rewards were unbalanced. The results of this study were not significantly different for men or women. However, gender has been an important variable in other studies of the relationship between work and stress, as we will describe below. Gender and work-related health outcomes Several studies indicate that being in the labour market has a beneficial effect on women’s psychological well-being (Aneshensel, Frerichs, & Clark, 1981; Gore & Mangione, 1983; Weich, Sloggett, & Lewis, 1998), but may cause additional burdens as well, particularly for parents (Rosenfield, 1989; Sogaard, Kritz-Silverstein, & Wingard, 1994). The effects of women’s labour force participation may vary by social class (Waldron & Jacobs, 1988) or by quality of their home environment (Warr & Parry, 1982). Recently Matthews et al. (1998) showed that even after taking into account women’s multiple roles (employment, marital status, domestic responsibility and parental status), work characteristics (psychological job strain, insecurity, unsociable working hours) and home characteristics (age of youngest child, number of children, childcare responsibilities and an older adult living in the family), there is still a social class gradient in women’s experience of psychological distress. Women would seem to be particularly vulnerable to the adverse effects of work. The gender gap in wages is well documented (England, 1992; Kilbourne et al., 1994). Women occupy lower status jobs than men and are more likely to have careers that are punctuated by career breaks. This means that they are less likely to have savings, investments or an adequate pension when they retire (Warren, Rowlinson, & Whyley, 2001). Despite their increasing numbers in the work place, women still occupy
less prestigious jobs than men. Women are less likely to be managers or to occupy positions of authority (Rosenfeld, Van Buren, & Kalleberg, 1998). Karasek and Theorell (1990) found that women experienced lower levels of decision latitude or control than men, while experiencing the same psychological demands. Women reported more negative work characteristics than men, such as fewer learning opportunities and monotony, in a recent study by Matthews et al. (1998). Parts of the labour market are still highly segregated by gender and some studies have investigated whether working in a sex atypical occupation has negative psychological consequences. Evans and Steptoe (2002), for example, found that female accountants were more likely to report anxiety than other groups of workers, and male nurses had higher rates of sickness absence. The authors interpret the results as suggesting that when men and women occupy jobs in which they are the cultural and numerical minority, they may experience adverse health effects that are gender specific. This work is relevant to our study because so many of the caring professions are predominantly female. Extrapolating from these results might suggest that men in caring occupations would be more susceptible to negative feelings than women. It has been argued that even when men and women perform the same jobs they are likely to have very different experiences of work. Williams (1995) conducted a qualitative study of men in traditionally female occupations: nurses, social workers, librarians and primary school teachers. She found that men tended to receive preferential treatment in hiring, were channelled into certain male identified specialties, and pressured to perform specific tasks that are associated with men. These tasks and specialties tended to be highly valued, so that men appear to benefit from being in the minority in terms of advancement in femaledominated careers. Because gender assumptions about work are seldom challenged, women’s lack of career success can be attributed to individual failings, to their collective inability to be competitive, or to their competing loyalties to home and family. This work might suggest that while men are often sex atypical in caring jobs, the fact that they are on a ‘‘glass escalator’’ would militate against any negative job-related feelings (Williams, 1995). In general, the literature linking work, gender and negative psychosocial outcomes supports the
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
hypothesis that women will experience more jobrelated negative feelings, with some evidence to the contrary. The literature also suggests, however, that in order to detect the effects of working in caring jobs we need to control for a large number of other variables that have been shown to be important contributors to the negative effects of work. These will include individual sociodemographic characteristics, job characteristics, home responsibilities and the amount of time devoted to work. Emotional labour Caring work involves many different skills. Nursing, for example, involves manual or physical work (lifting patients), managerial responsibility (co-ordinating the work of a team), technical skill (giving an injection) and intellectual work (assessing patients’ needs and planning care). However, the skills that are held to be most important to nursing are social and relational, including communication, co-ordination of people and activities, and the management of relationships with patients and other professionals. Nursing theory has focused on concepts such as empathy (Peplau, 1988), intimacy (Savage, 1995) and caring (Benner & Wrubel, 1989). More recently, the concept of emotional labour has been investigated in relation to nursing (James, 1989; Smith, 1992; Steinberg & Walter, 1992) as one example of a caring occupation. In a recent review, Steinberg and Figart (1999) describe two main approaches for defining emotional labour. Hochschild’s (1983) original statement emphasised the work that is involved in performing emotional labour, particularly the requirement that the worker display or suppress emotions, whereas others have focused on the effect that the emotional labour is designed to have on others. England and Farkas (1986) define emotional labour in terms of ‘‘efforts made to understand others, to have empathy with their situation, to feel their feelings as part of one’s own’’ (p. 91). The latter definition shifts the focus from the emotion displayed by the employee to the feelings of the client or recipient of services. This distinction is, of course, somewhat arbitrary because in order to manage the feelings of others, one must first be able to manage one’s own feelings. However, in this context it is a useful distinction because the latter definition is more consonant with the theory and philosophy of caring occupations such as medicine, education and social work, where the explicit focus
2163
is on the welfare and development of the recipient. Workers in these occupations are socialised to show genuine feelings of concern for the recipients of their service, and to display authentic, but controlled, rather than engineered emotion. Another way of delineating different kinds of occupations is by their relationship to the market. For many jobs that involve emotional labour, the commercial element of the transaction is obvious. Waiters and waitresses (Hall, 1993), hairdressers (Parkinson, 1996) and supermarket cashiers (Sutton & Rafaeli, 1988) are involved in an immediate exchange of money for goods and services. In some cases, the underlying economic nature of the relationship is often underscored by the tradition of tipping. Hochschild’s (1983) study of flight attendants showed that their training emphasises the commercial consequences of their behaviour. In caring relationships, market forces may be seen as part of the environment but not immediately relevant to the business in hand, particularly in systems set up to provide services that are ‘‘free at the point of delivery.’’ In sum, caring jobs as defined in this paper differ from many other jobs that involve emotional labour on a number of dimensions. Employees in caring occupations are required to focus on the feelings of their client, and to experience caring and compassion for them. The connection with the marketplace, while not entirely absent, is less pervasive than in many occupations. Finally, workers in caring occupations, which often require several years of education and training, are more likely to be of higher social class than the majority of the workforce. Studies that have investigated the effects of emotional labour on workers have focused on two main dependent variables: job satisfaction and pay. Wharton (1993) for example, gathered data from employees in a bank and a hospital and found that those in jobs that were considered to be high in emotional labour were no more prone to emotional exhaustion than other workers. Characteristics of the jobs themselves such as the level of autonomy and involvement were much more important in determining whether or not the worker experienced emotional exhaustion. More recently, Wharton (1999) has argued, using evidence from studies of the economic and non-economic costs and rewards both at work and at home, that although there are negative psychological consequences under certain circumstances, jobs requiring emotional labour can also have a positive effect on workers’ well-being.
ARTICLE IN PRESS 2164
D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
This work would seem to predict that we would find little clear evidence that working in caring occupations would have uniformly negative emotional consequences for workers. Burnout Burnout is a response to chronic interpersonal stress on the job (Freudenberger, 1975; Maslach, 1976). Since it was first described, three components have been identified: exhaustion, cynicism (or emotional withdrawal) and a sense of underachievement at work. Maslach et al. (2001) provide a critical review of what has been learnt about burnout over the last 25 years. Although there do appear to be three distinct components, exhaustion seems to be of primary importance with the other two features of the syndrome following almost in stages. A study of hospital nurses, for example, described the following sequence of events (Leiter & Maslach, 1998): (a) Stressful interactions with supervisors increase workers’ feelings of exhaustion. (b) High levels of exhaustion lead to cynicism, especially in the absence of support from coworkers. (c) Persistent cynicism leads to diminished feelings of efficacy, which can be ameliorated by support from co-workers. This study, in addition to describing the possible sequential relationship of the three components of burnout, reveals how important supervisors and coworkers can be in ameliorating the stresses and strains of work and in interrupting the process that leads to burnout. Burnout has been linked to a number of important job-related outcomes, such as absenteeism and turnover, and has been shown to have a detrimental effect on an individual’s social as well as work life. Some socio-demographic characteristics seem to be predisposing factors for the development of burnout, including age, gender, marital status and level of education. Younger people appear to be at greater risk but this finding may be an artefact of selection bias. If young people who become burned out change their jobs or leave the labour market, then long-serving employees may appear to be protected from burnout. But, long-serving employees may differ from those who leave in some important ways. They may, for example, have
invested less in their jobs from the beginning of their careers and so are less liable to disillusionment, or they may develop better coping mechanisms over time that protect them from burnout. Gender is also relevant, with men tending to score more highly on the second factor (cynicism) and women being slightly more likely to report higher levels of exhaustion. Previous research shows that being single (even more than being divorced), and having more education are also predisposing factors. Perhaps people who do not have an immediate family focus more on work and are less likely to have competing worries or compensating sources of emotional satisfaction. Those with higher levels of education may simply occupy more demanding and stressful jobs or it could be that having invested so much time and effort in their education, they expect more from their jobs, and are therefore more vulnerable to disillusionment. As well as individual characteristics, some job characteristics have been associated with burnout. These include; heavy workloads, time pressures, role conflict, lack of social support (particularly from supervisors and co-workers) and a lack of information or feedback. Low levels of participation in decision making and autonomy have also been implicated in burnout. Currently, Maslach et al. (2001) are focusing on six areas distilled from their previous research: workload, control (especially over resources needed to do the job) rewards (finance, status, appreciation), community (a sense of belonging at work, or conflict), fairness (inequity of workload or pay), and values (conflict between own and organisation’s values, or conflict between stated values and those that are enacted). Some recent research speaks directly to the relationship between emotional labour and burnout, which is close to some of the research questions we hope to address. Zapf, Seifert, Schmutte, Mertini, and Holz (2001) found that emotional labour explains at least part of the variance in burnout scores over and above job stress. Schaulfeli and Enzmann (1998) compared workers in the US and the Netherlands and found that while the basic patterns of burnout were similar across the five occupational groups (teaching, social services, medicine, mental health and law enforcement) that they studied, there were some interesting national differences. This suggests that characteristics of the setting in which work is performed, such as the way the work is organised, may also make employees
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
more prone to burnout. Increasing our understanding of these contextual variables will enable us to devise changes that will improve working conditions. In the meantime there are some precedents for research on the link between emotional labour, caring work and burnout. Research questions This study investigates the impact of caring work and gender on four outcomes: having difficulty unwinding and ceasing to worry after work and the extent to which workers feel exhausted or ‘‘used up’’ at the end of their shift. We ask: 1. Are workers in caring jobs more likely to report negative job-related feelings at the end of the working day than workers in other jobs, controlling for other factors that might contribute to feelings of stress and exhaustion? 2. Are women more likely to report negative jobrelated feelings at the end of the working day than men, controlling for all the other factors that might contribute to feelings of stress and exhaustion? Data The data for these analyses are taken from the British Household Panel Survey (BHPS) (Taylor, Brice, Buck, & Prentice-Lane, 2006). This survey has been undertaken annually since 1991 by the Institute for Social and Economic Research at the University of Essex. The BHPS involves an annual survey of each adult member of a nationally representative sample of more than 5000 households. This gives over 10,000 individual interviews each year. In each wave, interviewees are asked a wide range of questions about their employment status, including their occupation, pay, hours, and job satisfaction. In addition, questions about particular areas of sociological interest are added each year. For the purposes of this study, we restrict our analysis to those members of the BHPS sample that are either employed or self-employed, giving a sample of over 6500 people. Dependent variables In one single year, 1999, respondents to the annual BHPS survey were asked a series of supplementary questions about their health and
2165
well-being, including: ‘‘Thinking of the past few weeks, how much time has your job made you feel each of the following?’’ After I leave my work I keep worrying about job problems. I find it difficult to unwind at the end of the work day. I feel used up at the end of the work day. My job makes me feel quite exhausted by the end of the work day. Response categories ran from 1 (never) to 6 (all of the time), with categories for ‘‘missing,’’ ‘‘not applicable,’’ ‘‘proxy respondent’’ and ‘‘don’t know,’’ all of which are excluded in our analyses. Frequency distributions of these variables are shown in Table 1. These are the main outcome variables in this study. We describe in the Methods section below how they were used in the statistical analysis. Independent variables Caring occupations We classified all the occupations in the BHPS into two categories based on our knowledge of each occupation, taking into consideration the amount of interpersonal interaction involved every day and the extent to which their interactions are conducted in the context of a relationship and might be emotionally laden. We tried to capture the demands that the worker in one of these jobs will feel to care about the recipients of their services and to focus on their feelings, well-being or development, even at times at the expense of their own. The list of caring jobs is in Appendix A, which also gives the number of people in each occupational category in the data set. The majority of people in caring occupations work in health-related occupations, education and social work. The list of caring jobs that we used in this analysis turned out to be very similar to the list of jobs coded by England (1992) as scoring high on ‘‘nurturance.’’ Control variables These are included in the model to control for other aspects of the job or the individual that might explain negative job-related feelings irrespective of the type of job (caring/non-caring) or the gender of the worker. They include: overwork, caring for a sick or disabled person at home, sociodemographic
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
2166
Table 1 Tables showing cross-tabulations of being in a caring occupation and sex with the four outcome variables (column percentages in parentheses) Caring Yes Men
No Women
Men
Women
Worry Never Rarely Some of the time Much of the time Most of the time All of the time
53 53 36 21 14 6
(29.0) 337 (36.3) 1883 (46.0) 1529 (48.5) (29.0) 273 (29.4) 1077 (26.3) 808 (25.6) (19.7) 200 (21.5) 682 (16.6) 515 (16.3) (11.5) 63 (6.8) 229 (5.6) 148 (4.7) (7.6) 35 (3.8) 150 (3.7) 110 (3.5) (3.3) 21 (2.3) 74 (1.8) 40 (1.3)
Used up Never Rarely Some of the time Much of the time Most of the time All of the time
24 53 55 19 26 6
(13.1) (29.0) (30.0) (10.4) (14.2) (3.3)
165 236 256 111 121 39
(17.8) 893 (21.8) (25.4) 1335 (32.6) (27.6) 1029 (25.1) (12.0) 357 (8.7) (13.0) 356 (8.7) (4.2) 122 (3.0)
Exhausted Never Rarely Some of the time Much of the time Most of the time All of the time
28 59 41 26 24 5
(15.3) (32.2) (22.4) (14.2) (13.1) (2.7)
148 238 250 129 118 45
(15.9) 857 (20.9) 720 (22.9) (25.6) 1398 (34.2) 1032 (32.8) (26.9) 982 (24.0) 743 (23.6) (13.9) 363 (8.9) 269 (8.5) (12.7) 376 (9.2) 270 (8.6) (4.8) 117 (2.9) 116 (3.7)
Unwind Never Rarely Some of the time Much of the time Most of the time All of the time
51 56 37 20 16 3
(27.9) 302 (32.5) 1709 (41.7) 1288 (40.9) (30.6) 251 (27.0) 1210 (29.5) 893 (28.4) (20.2) 233 (25.1) 715 (17.5) 627 (19.9) (10.9) 67 (7.2) 213 (5.2) 150 (4.8) (8.7) 58 (6.2) 168 (4.1) 135 (4.3) (1.6) 18 (1.9) 81 (2.0) 55 (1.7)
724 987 790 270 268 107
Methods The first step in analysing these data was to investigate the structure of the four outcome variables. We first calculated all the pairwise polychoric correlations among these variable using the polycor package in R (Fox, 2006a; R Development Core Team, 2006). These correlations were then used as the basis for confirmatory factor analysis using the sem package in R (Fox, 2006b). First, we tested a model with a single latent factor; the results are shown in Fig. 1. As can be seen, this model does not provide a good fit to the data (w2 ¼ 2654.2, degrees of freedom ¼ 2, GFI ¼ .87). However, a two factor model shown in Fig. 2, fits
(23.0) (31.4) (25.1) (8.6) (8.5) (3.4)
characteristics (age, having a degree, marital status) and job characteristics (hours worked, overtime, working rotating shifts, liking the hours one works, job satisfaction, hourly pay, management responsibilities). Based on the finding of a social class gradient in the literature on gender, work and stress we used the Registrar General’s classification (where 1 ¼ professional and 5 ¼ unskilled worker) as a set of dummy variables. The inclusion of these variables as controls is supported theoretically by the literature reviewed on work and stress, emotional labour and burnout. We also included two variables, not having support in a crisis and not feeling appreciated which could be relevant to respondents’ emotional responses to work and a measure of general health.
Fig. 1. Confirmatory factor analysis with a single latent factor.
Fig. 2. Confirmatory factor analysis with two latent factors.
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
the data very well (w2 ¼ .425, degrees of freedom ¼ 1, GFI 4.999). One of the two factors loads on the two variables that relate to tiredness (‘‘I feel used up at the end of the work day’’ and ‘‘My job makes me feel quite exhausted by the end of the work day’’), while the other loads on the variables that relate to worry and stress. As a result of this, we combined these two pairs of variables into two indicators by summing the respective observed variables. The frequency distributions of these variables are shown in Fig. 3. Higher scores mean that a person feels more exhausted or more stressed. Findings In Table 2 we show linear regression estimates of effects on exhaustion and stress, the two composite outcome variables described above. The main explanatory variables of interest are working in a caring occupation and gender. The positive and significant coefficient for caring work suggests that nurses, doctors, teachers and social workers are more likely to feel exhausted and stressed, or to feel exhausted and stressed more frequently, than people in other jobs. Women are also much more likely to
2167
be exhausted and stressed than are men. These key variables are statistically significant after taking account of a number of important controls. It is worth noting that these findings relate to how much the respondents feel stressed when they return home from work, over and above the amount of worry they might experience ‘‘on the job.’’ Some individual characteristics, such as being older, seem to decrease the propensity to exhaustion and stress, while having a degree increases the chances of being stressed. Poor general health also increases the tendency to be tired and stressed after work. Some characteristics of the job seem to promote exhaustion and stress, including number of hours worked per week, overtime and higher pay. Work related factors that decrease problematic outcomes, including not having managerial responsibilities, having high job satisfaction, being able to work the hours that one prefers and not having promotion opportunities, are all negative and significant in this model. Taken together they suggest that people who are higher in the hierarchy might suffer more problems than people who are in non-managerial positions. Management also involves a degree of emotional labour because
Fig. 3. Frequency distributions of the two composite dependent variables.
ARTICLE IN PRESS 2168
D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
Table 2 Linear regression estimates of effects on worrying about work and difficulty unwinding after work Tired
Stressed
p-value ¼ .403), but there is between class and stress (F-statistic ¼ 30.5, p-value 5.05). The results show that people in lower skilled and manual occupations tend to have lower levels of stress than those in professional and managerial occupations.
Estimate t-value Estimate t-value Constant Caring occupation Female Single Number of children Age Carer Has a degree Hours worked per week Log monthly pay Not a manager No one to turn to in crisis Not appreciated Job satisfaction Hours spent on housework per week General health Works shifts Overtime hours per week Likes hours worked No promotion opportunities RG Class 2 RG Class 3a RG Class 3b RG Class 4 RG Class 5 Number of cases R-square
4.38* 0.638* 0.441* 0.031 0.037 0.011* 0.201* 0.032 0.034* 0.036 0.211* 0.103 0.139 0.234* 0.0004
8.39 6.64 5.86 0.464 1.12 4.05 2.30 0.373 7.53 .480 3.01 .812 1.31 9.77 .102
2.93* 0.421* 0.268* 0.062 0.008 0.011* 0.099 0.300* 0.020* 0.121* 0.46* 0.166 0.090 0.161* 0.002
6.57 5.29 4.35 1.16 0.303 4.81 1.39 3.99 5.10 1.80 7.93 1.49 0.989 8.05 0.492
0.133* 19.3 0.123 1.16 0.058* 11.2 0.579* 8.74 0.008 0.136
0.125* 0.003 0.046* 0.357* 0.112*
20.9 0.034 10.2 6.56 2.28
0.081 0.181 0.705* 0.770* 0.585*
0.723 1.49 5.74 5.96 3.76
0.188 0.010 0.118 0.052 0.084 6539 0.200
1.47 0.072 0.825 0.334 0.416
6542 0.27
t-Values are calculated from White-corrected estimated standard errors. Note: *po0.05.
managers are responsible for many aspects of workers’ welfare as well as their productivity, and managers in caring occupations may have a double burden, responsible as they are for both clients and junior staff. A few variables appear not to contribute much to the risk of being exhausted or stressed: being single, having no one to turn to in a crisis, not feeling appreciated and hours spent in housework each week are all insignificant. The last variables in the model are the dummy variables for social class, included here to control for the known association between social class and psychological distress. There is no statistically significant association between class and exhaustion (F-statistic ¼ 1.02,
Discussion and conclusions Our results suggest that if you ask women in caring occupations how they feel at the end of the day, they will often respond ‘‘worried, tense, drained and exhausted.’’ Men in caring jobs also bear emotional costs, but the penalties do not appear to be as great as those experienced by female employees. Women working in caring jobs are more likely than men to experience all four of the negative feelings that we studied. Other variables that increased the likelihood of negative job-related feelings include being young, being a manager, experiencing low levels of job satisfaction and not liking the hours that one works. The number of hours worked and the number of overtime hours worked had separate, independent effects on the experience of negative affect after work. Workers who have poor general health are also much more vulnerable to experiencing negative work-related feelings. We argued that worrying about work or having trouble unwinding seem to be much more likely if the employee were in a managerial position. The positive effect of having a degree and higher pay, in addition to the negative effect of not being a manager contributed to this interpretation, as did the fact that having no promotion prospects seemed to reduce the tendency to worry after work. Interestingly, working shifts also appeared to reduce the tendency to worry, but increased the tendency to have trouble unwinding. Although ‘‘not being a manager’’ was negatively related to feeling exhausted and ‘‘used up’’ after work, there was no association with pay or having a degree. The only variable that is significant in the model of exhaustion while insignificant in the other is ‘‘being a carer.’’ Caring for another adult at home can lead to exhaustion, but does not, according to these estimates, lead to increased feelings of stress. This study has a number of limitations: the data are cross-sectional, which restricts our ability to make causal arguments, and the coding of occupational categories is into binary categories. Most jobs have an element of contact with other human beings, so it would be preferable to construct a
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
variable to measure the amount of interpersonal interaction or emotional labour typically associated with each occupation. However, the study does show that caring work has a negative impact on workers in a large national sample of over 6500 workers across all economic sectors in contemporary Britain. There is increasing sociological interest in caring work as it affects employees’ experience of work, career trajectories, the domestic division of labour and labour market segregation. Caring work is interesting theoretically because it challenges traditional notions of a division between public and private spheres that have prevailed since the Industrial Revolution. It is also of increased concern to policy makers, as the number of people employed in caring occupations is increasing and is likely to continue to increase with the demands of an aging population on health and social services. The goals of future work will be to refine the classification of occupations, and to extend the study over time so that we can find out whether negative work related feelings have a cumulative impact on workers health and job tenure. What are the long-term consequences of feeling worried, stressed, drained and exhausted on a regular basis? We would also like to find out whether there is something inherently stressful in caring for others, or whether caring work is organised in such a way that workers are exposed to some of the known causes of burnout. Do nurses, for example, have less autonomy and control over their work than people working in jobs that are similar in other respects? Are teachers less able to regulate the amount of face-to-face contact they have in a working day than people in other occupations? The organisational context of caring work may turn out to be more important than the content of the jobs themselves. Important questions remain to be answered about how managerial responsibilities affect employees, particularly those who are also responsible for clients. How does the emotional labour of managers differ, if at all from that of clinicians or teachers? Are the additional demands of managerial jobs mitigated by the fact that managers are rewarded with higher pay than other workers? We also need to find out why gender is so salient in these analyses. Do men and women in caring jobs behave differently, or are they treated differently? Or does the answer simply lie in the fact that they respond differently to questions about their feelings? There may still be some gender differences in the propensity to report negative emotions. Finally,
2169
we have focused in this paper on the emotional costs of caring, ignoring for the time being the possibility that caring for others might also have rewards. Future work on caring occupations will need to consider what the rewards might be, whether these can be measured, and whether the balance between the rewards and costs is consistent with ideals of social justice. Appendix A. List of caring occupations with number of individuals in each category in the data set
Code 220 222 223 224 230 233 234 235 292 293 340 341 342 343 344 345 347 363 370 371 640 641 644 650 651 652 659
Job title Medical practitioners Ophthalmic opticians Dental practitioners Veterinarians University teachers Secondary education teachers Primary & nursery education teachers Special education teachers Clergy Social workers & probation officers Nurses Midwives Radiographers Physiotherapists Chiropodists Dispensing opticians OT, speech therapy, etc. Personnel & IR officers Matrons & houseparents Welfare & community work Asst. nurses & nursing auxs. Hospital ward assistants Care assistants & attendants Nursery nurses Playgroup leaders Educational assts Other childcare occs.
N % 34 2.9 13 1.1 11 0.9 3 0.3 38 3.3 104 8.9 101 8.7 21 1.8 24 2.1 62 5.3 177 15.2 14 1.2 9 0.8 8 0.7 2 0.2 1 0.1 29 2.5 17 1.5 21 1.8 77 6.6 45 3.9 17 1.5 142 12.2 43 3.7 15 1.3 50 4.3 84 7.2
References Aneshensel, C. S., Frerichs, R. R., & Clark, A. V. (1981). Family roles and sex differences in depression. Journal of Health and Social Behaviour, 22(4), 379–393. Benner, P., & Wrubel, J. (1989). The primacy of caring: Stress and coping with health and illness. Menlo Park, CA: Addison-Wesley.
ARTICLE IN PRESS 2170
D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171
Cancian, F. M., & Oliker, S. J. (2000). Caring and gender. Thousand Oaks: Pine Forge Press. Cheng, Y., Kawachi, I., Coakley, E. H., Schwartz, J., & Colditz, G. (2000). Association between psychosocial work characteristics and health functioning in American women: Prospective study. British Medical Journal, 320, 1432–1436. de Jonge, J., Bosma, H., Peter, R., & Siegrist, J. (2000). Job strain, effort reward imbalance and employee well-being: A large scale cross-sectional study. Social Science & Medicine, 50(9), 1317–1327. England, P. (1992). Comparable worth: Theories and evidence. New York: Aldine de Gruyter. England, P., & Farkas, G. (1986). Households, employment and gender: A social, economic and demographic view. New York: Aldine. England, P., & Folbre, N. (1999). The costs of caring. The Annals of the American Academy of Political and Social Science, 561, 39–51. EU. (2002). Unemployment, employment and public health. Brussels: European Union. Evans, O., & Steptoe, A. (2002). The contribution of gender-role orientation, work factors and home stressors to psychological well-being and sickness absence in male and female dominated occupational groups. Social Science & Medicine, 54, 481–492. Fox, J. (2006a). Polycor: Polychoric and polyserial correlations (R package, version 0.7-2). Fox, J. (2006b). Sem: Structural equation models (R package, version 0.9-6). Freudenberger, H. J. (1975). The staff burnout syndrome in alternative institutions. Psychotherapy: Theory, Research and Practice, 12, 72–83. Gore, S., & Mangione, T. W. (1983). Social roles, sex roles and psychological distress: Additive and interactive models of sex differences. Journal of Health and Social Behaviour, 24, 300–312. Gove, W. R., & Tudor, J. F. (1973). Adult sex roles and mental illness. American Journal of Sociology, 78(4), 812–835. Hall, E. J. (1993). Smiling, deferring, and flirting: Doing gender by giving ‘‘good service’’. Work and Occupations, 20(4), 452–471. Hochschild, A. R. (1983). The managed heart: The commercialisation of human feeling. Berkely: University of California Press. House, J. S., Strecher, V., Metzner, H., & Robbins, C. A. (1986). Occupational stress and health among men and women in the Tecumseh community health study. Journal of Health and Social Behaviour, 27, 62–77. James, N. (1989). Emotional labour: Skill and work in the social regulation of feelings. Sociological Review, 37, 15–42. Karasek, R., & Theorell, T. (1990). Healthy work: Stress, productivity and the reconstruction of working life. New York: Basic Books. Kessler, R. C., House, J. S., & Turner, J. B. (1987). Unemployment and health in a community sample. Journal of Health and Social Behaviour, 28, 51–59. Kilbourne, B. S., Farkas, G., Beron, K., Weir, B., & England, P. (1994). Returns to skill, compensating differentials, and gender bias: Effects of occupational characteristics on the wages of white men and women. American Journal of Sociology, 100(3), 689–719.
Leiter, M. P., & Maslach, C. (1998). The impact of interpersonal environment on burnout and organisational committment. Journal of Organisational Behaviour, 9, 297–308. Link, B. G., Lennon, M. C., & Dohrenwend, B. P. (1993). Socioeconomic status and depression: The role of occupations involving direction, control and planning. American Journal of Sociology, 98, 1351–1387. Losocco, K. A., & Spitze, G. (1990). Working conditions, social support, and well-being of female and male factory workers. Journal of Health and Social Behaviour, 31, 313–327. Maslach, C. (1976). Burned-out. Human Behaviour, 5, 16–22. Maslach, C., Schaulfeli, W. B., & Leiter, M. P. (2001). Job burnout. Annual Review of Psychology, 52, 397–422. Matthews, S., Hertzman, C., Ostry, A., & Power, C. (1998). Gender, work roles and psychosocial work characteristics as determinants of health. Social Science and Medicine, 46, 1417–1424. Mirowsky, J., Ross, C., & Reynolds, J. (2000). Links between social status and health status. In C. E. Bird, P. Conrad, & A. M. Fremont (Eds.), Handbook of Medical Sociology. New Jersey: Prentice-Hall. Parkinson, B. (1996). Changing moods: The psychology of mood and mood regulation. New York: Adison Wesley Longman. Peplau, H. E. (1988). Interpersonal relations in nursing. London: Macmillan Press Ltd. R Development Core Team. (2006). R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. Reynolds, J. R. (1997). The effects of industrial employment conditions on job-related distress. Journal of Health and Social Behaviour, 38, 105–116. Rosenfeld, R. A., Van Buren, M. E., & Kalleberg, A. L. (1998). Gender differences in supervisory authority: Variation among advanced industrialised democracies. Social Science & Medicine, 27, 23–49. Rosenfield, S. (1989). The effects of womens’ employment: Personal control and sex differences in mental health. Journal of Health and Social Behaviour, 30, 77–91. Savage, J. (1995). Nursing intimacy: An ethnographic approach to nurse patient interaction. London: Scutari Press. Schaulfeli, W. B., & Enzmann, D. (1998). The burnout companion to study and practice: A critical analysis. Philadelphia: Taylor & Francis. Siegrist, J. (2001). Work stress and health. In W. Cockerham (Ed.), The Blackwell companion to medical sociology (pp. 114–125). Oxford: Blackwell Publishers Limited. Smith, P. (1992). The emotional labour of nursing. London: Macmillan. Sogaard, A. J., Kritz-Silverstein, D., & Wingard, D. L. (1994). Finnmark heart study: Employment status and parenthood as predictors of psychological health in women, 20–49 years. International Journal of Epidemiology, 23, 82–90. Steinberg, R. J., & Figart, D. M. (1999). Emotional labor since the managed heart. Annals of the American Academy of Political and Social Science, 561, 8–26. Steinberg, R., & Walter, W. L. (1992). Making women’s work visible: The case of nursing—First steps in the design of a gender-neutral job comparison system. Washington, DC: Institute for Women’s Policy Research. Sutton, R. I., & Rafaeli, A. (1988). Untangling the relationship between displayed emotion and organisational sales: The case of convenience stores. Academy of Management Journal, 31(3), 461–487.
ARTICLE IN PRESS D.N. Barron, E. West / Social Science & Medicine 65 (2007) 2160–2171 Taylor, M. F., Brice, J., Buck, N., Prentice-Lane, E. (2006). British household panel survey user manual volume A: Introduction, technical report and appendices. Colchester: University of Essex. Waldron, I., & Jacobs, J. A. (1988). Effects of labour force participation on women’s health: New evidence from a longitudinal study. Journal of Occupational Medicine, 30(12), 977–983. Warr, P., & Parry, G. (1982). Paid employment and women’s psychological well-being. Psychological Bulletin, 91(3), 493–498. Warren, T., Rowlinson, K., & Whyley, C. (2001). Female finances: Gender wage gaps and gender assets gaps. Work, Employment and Society, 15(3), 465–488.
2171
Weich, S., Sloggett, A., & Lewis, G. (1998). Social roles and gender difference in the prevalence of common mental disorders. British Journal of Psychiatry, 173, 489–493. Wharton, A. (1993). The affective consequences of service work: Managing emotions on the job. Work and Occupations, 20(2), 205–232. Wharton, A. (1999). Emotional labour. Annals of the American Academy of Political and Social Science, 561. Williams, C. L. (1995). Still a man’s world: Men who do women’s work. Berkeley: University of California Press. Zapf, D., Seifert, C., Schmutte, B., Mertini, H., & Holz, M. (2001). Emotion work and job stressors and their effects on burnout. Psychology and Health, 16(5), 527–545.