Applying the effort-reward imbalance model to household and family work: a population-based study of German mothers

The health-related impact of qualitative psychosocial work characteristics has been amply demonstrated, but research has predominantly focused on paid labour. In this context, a variety of studies have examined the health-related impact of effort-reward imbalance (ERI) [1]. This model postulates that work characterized by high effort and low reward reflects a state of failed reciprocity between 'costs' and 'gains'. In case of an imbalance between high effort and low reward strong emotions of anger and frustration may emerge resulting in a sense of being treated unfairly. In the long run sustained activation of the autonomic nervous system may contribute to the development of physical and mental disease [2, 3]. A specific personality trait characterized by excessive work-related overcommitment increase the risk of ERI and is the most important reason for maintaining a high cost/low gain condition. Although the ERI model has primarily been used as a theoretical framework for explaining work-related health risks, Siegrist [4] stresses that reciprocity of exchange is not confined to working life. It may be experienced in a similar way also in other domains, such as marital and parental relationships. Against this background, Knesebeck and Siegrist [5] extended the ERI model to social relationships beyond paid work, namely marital and parental roles and less specific civic roles. Li et al. [6] applied the model to school settings in order to analyse students' psychosocial stress and its association with self-rated health in adolescents. These studies provided first evidence that the ERI model may also be important outside the work-life.

As far as we know no studies exist where Siegrist's model had been applied to unpaid household and family work. However, there are analogies between paid and unpaid work justifying such an extension. Firstly, similarly to paid employment, household and family work are defining one's social identity and social status to a substantial degree. Secondly, unpaid work might be equally demanding, and some services such as caring, cooking and cleaning can also be purchased on the labour market. Thirdly and lastly, household and family work may also offer 'rewards' in terms of promoting self-esteem and may therefore provide the potential of experiencing a favourable self-concept.

However, there are also important differences between paid and unpaid work need to be taken into account. Constraints of demands in household and family work are often less pervasive and less obvious. Rewards received are more often of an emotional than of a material nature [5]. For that reason, rewards as assessed with the original ERI questionnaire in terms of 'financial and career-related rewards', 'esteem rewards' and 'gratification of job security' cannot be generalized to unpaid work. Thus, we developed a new questionnaire taking the specific efforts given and rewards received in household and family work into account. Initially, a review of the literature was carried out to identify the relevant aspects of effort and reward in this domain. This review revealed that some crucial efforts of the original ERI also hold for work at home, in particular 'time pressure', 'interruptions and disturbances' and 'pressure to work overtime'. A more specific expression of effort in household and family work was 'unappreciated work' and 'the feeling of being available to others'. The household and family work was characterised as 'diffuse' and 'never-ending' [7‐9]. With respect to 'rewards' the review yielded that this work foremost provides interpersonal rewards, particularly with respect to child care [8]. Baruch and Barnett [10] concluded that women found the most rewarding aspects of the mother-role in the love of children, liking the kind of people they are, and pleasure in their accomplishments. For the role as a wife they also identified aspects related to the partner as important, in particular having an emotionally supporting partnership. Another source of reward is the meaning of the role as a mother and wife in persons' self-conception. In this perception a high value of being a mother and taking care for the family provides self-rewards resulting in better psychological well-being [11‐13]. On the other hand, it was reported that household and family work has little institutional recognition and social prestige, and the recognition of being a housewife has significantly decreased over the last decades [8].

On the base of the literature review questionnaire items were generated and discussed with mothers as part of qualitative interviews. With their feedback in mind the questionnaire was developed and tested for reliability using clinical data of mothers with burnout [14].

This paper reports findings of this questionnaire using data of a cross-sectional population survey of German mothers. In more detail, the study addresses the following questions:

The age of women ranged between 17 and 60 years (mean age 39.1 ± 6.8), their youngest child ranged from 0 to 18 (mean age 9.4 ± 5.3). On average, the mothers had 1.8 children, around 40.5% of them had one child, 43.5% had two and about 16% had three and more children. Most of the women (71.6%) were married, 16.8% were single mothers. One third (32.6%) of the participants had secondary general school and about 36% got a per capita income below 926 Euro. Overall, about 78% of the mothers were employed, including mothers working up to 19 hours a week (21.8%), working half time (37.8%) and full time (17.9%). The majority of women (74.4%) reported that they are exclusively or mainly responsible for household and family work, whereas this holds only for 1% of the male counterpart. A total of 23.9% stated that their partners are equally involved in household and family work. Among women who working full time more than one in every two (55.7%) reported to bear the main burden of household and family work whereas this applied to only 4.8% of men. In total, 38.7% of full time working women stated that household and family obligations were equally shared with the partner (not displayed). Negative spillover from work to family applied to 62.7% of working women. Out of these, 41.0% perceived moderate and 21.7% marked work to family conflicts. The basic socio-demographic and family-related characteristics of the study sample are provided in Table 1.

The application of effort-reward imbalance (ERI) to household and family work provides an access to information on work at home similar to what was reported for paid work. Although there are crucial differences between paid and unpaid work, we postulate that nonreciprocal exchange in household and family work may lead to similar emotional distress caused by the fact of being treated unfairly in the role as a wife and mother. In a recent study, a questionnaire for assessing effort-reward imbalance in household and family work was developed and tested on clinical data of mothers with burnout. Analogous to the original version, the newly developed questionnaire contains an extrinsic (dysbalance of effort and reward) as well as an intrinsic component (overcommitment). For assessing overcommitment only moderate linguistic changes from the original questionnaire were needed. In this study the factorial validity of this questionnaire was evaluated using data of a population sample of mothers with minor children. The findings confirmed the factorial structure of the theoretical ERI model. The subscale 'reward' is devided into four dimensions: 'intrinsic value of household and family work', 'societal esteem', 'recognition from the partner', and 'affection from the child(ren)'. Though overcommitment is part of the ERI-model the theoretical position of this component is not completely clear so far. Originally, overcommitment was conceptualized as the intrinsic part of effort. In this view, highly overcommitted employees tend to invest too many efforts. Therefore, the amount of effort invested is dependent upon extrinsic demands as well as intrinsic efforts. Meanwhile overcommitment is seen as an independent concept which influences the perception of both, high efforts and low rewards [3]. According to this current approach overcommitment mediates the relationship between the effort-reward ratio and health outcomes. Therefore, we kept overcommitment out of confirmatory factor analysis and controlled for this personality characteristic when analysing the impact of the ratio on health. However, in a recent publication overcommitment is regarded as an integral part of ERI [20]. Thus, clarification about theoretical conceptualization of overcommitment is needed in order to realize a more coherent application of the ERI model.

Irrespectively of how overcommitment was conceptualized, our findings revealed that a significant proportion of mothers showed high rates of overcommitment in terms of 'inability withdraw from household and family work obligations'. We also found high rates of effort, expressed particularly by the feeling 'as never being off duty' and 'having to do a thousand things all at the same time'. Lack of reward was less relevant, but statements concerning less 'societal esteem' received considerable approval. This supports findings from earlier studies suggesting that the value of being a housewife and mother has declined over the last decades [8].

Our analyses have shown that ERI differs according to socio-demographic factors and family characteristics. As expected, ERI ratio scores increased with the number of children and decreased when children are getting older. Beside child-related risk factors also socioeconomic status seems to affect the mismatch between requested efforts and given rewards. Our findings suggest that in particular low income is associated with higher effort as well as with noticeably lower reward, resulting in higher ERI scores. A high mismatch was also found for single mothers. This indicates that ERI in household and family work may particularly capture the stress experience of socially disadvantaged mothers.

Through increased participation in the labour force, combining work and family demands are an increasingly common source of conflict for women [21]. Negative work to family spillover describes the fact that employment obligations adversely affect family obligations. Several studies have pointed out that negative work to family spillover hamper marital satisfaction and functioning [22]. Our analyses revealed that employment status per se was of minor relevance for women's perception of ERI. However, high levels of negative work to family spillover turned out to have a superior effect on ERI scores, in particular with respect to the 'effort' component. For future research it would be interesting to assess ERI at work and at home simultaneously. This approach may provide a more comprehensive picture of women's total daily stress perception and may give insights into the interplay between stress experiences in both domains.

In line with Bebbington [23] our findings suggest that women still perform the majority of the household and family work even when they are working fulltime. Perceived inequality in the division of household and family work was of major importance for high ERI-scores in women. Interestingly, different to the effect of negative spillover, men's lack of engagement did not primarily affect the perception of effort, but of reward. Thus, inequality in divison of domestic work seems to go along with women's perception of undervaluing their work. As Siegrist [24] pointed out, recurrent experience of reward deficiency in a core social role impairs successful self-regulation and in the long run increases illness susceptibility. Analysis of the health-related impact of ERI confirmed that mothers experiencing an imbalance between efforts and rewards at home are at higher risk of ill health. With the exception of hypertension we found a linear association between ERI and health outcomes such that health impairments continuously increased with ERI. This result provides first evidence that ERI in household and family work defines a state of emotional distress with negative health consequences, analogous to ERI at work. Given that disadvantaged mothers are at higher risk of ERI in household and family work, this concept may provide a complementary framework for explaining the social gradient in women's health. However, further analyses are needed in order to specify the role of ERI for explaining health inequalities among mothers.

Finally, some crucial limitations of the study need to be addressed. As our study is based exclusively on subjective data, the association between ERI and health outcomes should be interpreted cautiously, as there is some evidence that individual differences in personality traits may affect the reporting of stressors and health problems alike [25]. Additionally, with our cross-sectional study design we cannot draw conclusions about causality, but we assume in line with literature [1‐3] that ERI acts on health rather than the other way around. In order to minimize sources of error caused by self-reported measures, we adjusted for pessi-mism and overcommitment as personality traits that may influence mother's responses. However, further investigations with repeated measures over time as well as more objective measures of mental health are needed in order to further validate the findings. Finally, it has to be mentioned that the study was performed by means of an access panel sample survey. Access panel designs are composed of respondents giving their general consent to participate in surveys. Hence, besides usual response bias due to lack of controllability of target persons also selection bias due to the preselection of participants may reduce the representativeness of our study. On the other hand, using access panel data allows control over sample coverage with respect to socio-demographic and household characteristics. In addition, this panel facilitates access to target groups that are otherwise difficult to reach. In particular this applies to mothers with minor children. In light of the declining willingness to participate in surveys [26] access panel surveys may become more important in the future. This underlines the need for more basic research in order to assess the representativeness of such a sample type.

In sum, it can be stated that the newly developed questionnaire was shown to have high validity according to its factorial structure, and promising results for extending the ERI model to household and family work were obtained. Further analyses are necessary in order to clarify whether ERI may provide an appropriate theoretical approach to depict health-related stress experiences in household and family work.