The main aim of this study was to examine the effect of parental employment status on the health of their adolescent children in Slovakia and the Netherlands. Four main results emerged from it. Firstly, having an unemployed father negatively influences the health of Slovak male adolescents but has no effect on the health of Dutch adolescents. Secondly, having a disabled father has an effect on some aspects of the health of Dutch males and females but does not influence the health of Slovak adolescents. Thirdly, having a mother disabled, unemployed or a housewife has a negative effect on the self-esteem of Slovak adolescents. Fourthly, Dutch males whose mother was a housewife had worse long-term well being in comparison with those with an unemployed mother, whereas Dutch females whose mother was a housewife reported better psychological well-being than those with an employed mother.
One of the assumptions of the present study was that parental unemployment in Slovakia would have greater negative effects on the adolescents' health than in the Netherlands. Not all analyses supported this hypothesis, but a number of them did, mainly with regard to the fathers. One of the explanations could be the worse income circumstances in Slovakia. Men in Slovakia, whose incomes are usually higher than women's, are mostly perceived as family breadwinners. Therefore, when the father is unemployed, the family income decreases more significantly. As mentioned above, the financial support during unemployment is worse in Slovakia than in the Netherlands. Lack of financial resources has been associated with worse health in adolescence . This could be the reason why father's unemployment seems to affect the health of Slovak adolescents more than Dutch ones in our research. On the other hand, our previous study on the consequences of parental unemployment  showed that the negative effect of father's unemployment on the health of adolescents remains even after controlling for perceived financial strain and educational level of the father. This raises the question of whether it is only finances or also a psychological effect which negatively influences adolescents' health and persists regardless of the financial situation. Stress caused by unemployment may decrease the father's support given to his children, and his frequent presence at home may increase the potential for conflicts with adolescents. Both frequent conflicts and low father's support alike have been negatively related to various aspects of health [23, 24]. The present study does not allow us to answer this question. Further work should therefore be directed towards deeper understanding of the economic and psychological consequences of unemployment and their effect on the family.
Our results suggest that adolescents' health reactions to father's unemployment or disability vary not only between boys and girls within each country, but also between the studied countries. In contrast to the results from Slovakia, in the Dutch sample solely having a disabled father was negatively associated with some health outcomes. As yet, though, we have no ready explanation for this difference.
Among Dutch adolescents, two interesting associations between mother's employment status and adolescents' health were found. Having an unemployed or disabled mother had no negative effect on Dutch adolescents' health. However, females whose mothers were housewives had better psychological well-being than those with employed mothers. On the other hand, males whose mothers were housewives had worse long-term well-being than males with employed mothers. These contrasting findings need further clarification. Is it beneficial for 16-year-old adolescents to have their mother as a housewife? Do girls appreciate it, in contrast to boys? Aughinbaugh and Gittleman  suggest that for adolescents not having parents at home the whole day implies greater responsibility. Some of them are able to benefit from this situation and others not. Although being a housewife is very frequent in many countries, studies concerning the effect of this type of mother's daytime activity on adolescents are rare in the literature. Aughinbaugh and Gittleman  report no correlations between maternal employment during adolescence and increased involvement in risky activities. The daily experience of time spent with parents has been measured among early adolescents . The results showed no differences in daily experience based on mothers' employment status. We are not aware of any similar study focusing on health and well-being of adolescents. Such studies would be beneficial for explaining our contrasting findings.
Despite our expectations, relationships between parental employment status and adolescents' health were not found in all health indicators and were not very strong. Previous research showed the importance of the duration of unemployment. Sleskova et al.  reported that parental long-term unemployment (longer than one year) has negative consequences on adolescents' health, but short-term unemployment is not important. Harland et al.  suggested that long-term parental unemployment has a negative effect on somatic complaints and attention problems while short-term unemployment affects behavioural and emotional problems. This indicates that the length of unemployment should also be considered in future research.
In this work we studied the effect of parental employment status on the adolescents' health, but we did not control this effect for possible confounding factors such as education, income or parental occupational status. Unemployment is often considered as an indicator of low socio-economic status [28, 29]. This is due to several factors. Firstly, less-educated people and those from lower occupational classes run a greater risk of failing to find a full-time job . Secondly, unemployment very often entails considerable financial loss  and subsequent decrease in socio-economic status. It could therefore be possible that it is not the fact that a parent is unemployed but simply low income or low parental education that affects the health of adolescents. With regard to disablement the situation may be similar. However, as mentioned before, our previous study conducted among Slovak adolescents  showed that low socio-economic status, represented by financial strain and low education, does not mediate the relationships between parental unemployment and health of their children. It is highly probable that such result would also be obtained in the present study, but then similar analyses would need to be made among Dutch adolescents.
Some strengths and limitations of the study should be mentioned. A key strength is that the study contains identical health and employment status indicators used in two countries and very similar samples so that it enables cross-cultural comparisons. Furthermore, the data contain closer information about the reasons for not being employed. As mentioned above, studies focusing on the trans-generational impact of parents' employment status on the health of their children are rare. Of those which have been published in recent years, we are not aware of any which has considered the reasons for not being gainfully employed. Those parents who were not employed because of retirement, disability or maternity leave, or who were housewives/househusbands, were either considered together with the unemployed as "non-employed" , or no information about the reasons for being without a paid job were available . This work therefore adds to the knowledge about the effects of parental employment status by distinguishing between the reasons for parents not being gainfully employed.
The first limitation of this study is the categorization of mothers' employment status. The category 'housewife' is rather problematic for Slovak adolescents. As mentioned above, women become housewives in Slovakia mainly because of their long-term inability to find a job. So it may happen that adolescents whose mother is actually housewife and is not looking for the job anymore indicated her to be unemployed or vice versa. However, according to our results, it seems to be that with regard to health of children, being unemployed or being a housewife has the same effect. Therefore we believe it is unlikely that this limitation will bias our study.
The second limitation is the difference between the mean ages of the studied cohorts, which is somewhat more than one year. Even though we tried to make our data sets as comparable as possible, the age of the respondents is slightly different. Given the fact that both studied samples were selected from secondary school students, living in most cases with their parents, the age difference should not be so problematic. Furthermore, the age difference can only explain our results if the SES gradient depends on the adolescents' age, and this age effect would have to be very large to explain our results.