Work-Family Balance in the Active Age Ethnic Hungarian Population in Romania

Abstract

Value shifts and labour market transformations of the recent past have increased the importance of the work-family balance in the active population. Work overload also means an increased health risk. This study aims at identifying the main demographic, social and work-related determinants of work-family balance in the ethnic Hungarian active age population of Mures County, Romania. Linear regression is performed to assess the controlled effects of variables. Single parents, parents with more children and shift workers are at increased risk of imbalance. Demographic agents account for more disparities in work-family balance than do work-related features.

1. Introduction

The aim of the paper is to examine the impacts of social, demographic and work related agents on work-family balance for active age ethnic Hungarian adults in a multi-ethnic region of Romania, namely Mures County. Hungarians have a somewhat unique situation within Romania, so that the new insights offered by this paper are to be interpreted first of all in terms of local knowledge about this population. There are, however, some broader implications for work-family balance to learn from the results for societies in transition, with typically full-time female employment and persisting traditional gender role division and with deficient social security for single parents and larger families.

In general, the ethnic Hungarian minority in Romania is facing the same demographic challenges as European societies do. Mean age is old, aggravated by the fact that the migration of the active age adults to Western Europe has been high in the past three decades. Therefore, the demographic consequences for the community and for the country in general are considerable (Kiss and Barna 2012). There are no significant differences in the life expectancy of the ethnic majority and no higher age-specific mortality rates are registered for Hungarians, except for suicide which in the regions inhabited by them is at least two times the country average (Radulescu 2014).

Full time female employment is typical, the working time of women is one of the highest in Europe and part-time jobs are almost non-existent for mothers (Leovaridis and Nicolaescu 2011). However—and this goes for the whole of the country—daycare facilities for very young children (1 to 3 years old) are scarce and only cover about 5% of the age group (Bragă and Iancu 2011), which means that care for young children is provided mostly within the family. With similar fertility rates to the average European one, Hungarians’ attitudes towards family life and gender roles, though their modernization, are still more traditional, even in the youngest age groups (Pongrácz 2009), the divorce rates and the number of single parents are lower (Földházi 2009).

The socioeconomic situation of the ethnic Hungarians is even worse than the country average. Although the ethnic composition of settlements has no effect on the number of businesses per capita, at the level of counties and smaller regions, in some smaller towns where the Hungarians are in a majority, there is a greater entrepreneurial activity. Nevertheless, in spite of this increased entrepreneurial willingness of Hungarians, asymmetric power differences and resource access differentials to the Romanian majority are typical (Szabó 2010). Social stratification analysis reveals that ethnic Hungarians are under-represented in better paying (technical, commercial) occupations and in “secure” private sector jobs (Veres 2013), university graduates get hired in the competitive private sector at a lower rate than the majoritarian Romanians (Csata et al. 2006). Among Hungarians, salaries are lower and income inequalities are smaller and both the proportion of the economic elite and the poor are behind the national average (Kiss 2010).

Although further research in this field is necessary, one can infer that the economic disadvantage of ethnic Hungarians implies greater financial instability for them, as well as the need to work hard, having further jobs or performing extra paid activities in order to attain economic prosperity, which is assumed to endanger work-family balance.

Theoretical Aspects. The Determinants of Work-Life Balance

The specific situation of ethnic Hungarians in Romania is under-researched, there are so far no representative studies on work-family balance in this population. Therefore we depart from more universal theories and empirical findings in order to point out those variables which one can assume to effect balance. Studying this phenomenon and its determinants is the more important as dissatisfaction with work-family balance is associated with higher levels of burnout and mental health problems (Shanafelt et al. 2012).

The changes in the gender roles in the recent past have accentuated the agenda of the work-family balance in active ages. The main structural reasons are the changes in the families’ lives, such as the emergence of the dual-earner family model and the changes in the family structure, primarily the increasing number of single parents (Allen et al. 2000). Also, with the increase of general welfare, a value shift to post-materialist values was also predicted to gain in importance (Inglehart 1997).

In this paper we examine the effect of those work-related, social and demographic agents that are supposed to impact upon the work-family balance of ethnic minority Hungarians in Mures County, Romania. The departure point for the selection of variables to be included in the analysis is the findings of previous international studies on the topic.

First and foremost, work-related aspects like the duration and timing of work predict the perceived work-life balance. Longer working hours generally lead to less balance (Anttila et al. 2015), however, there are a few findings that contradict this (Pereira and Coelho 2012). Dex and Bond (2005) assessed that in the British working population the work hours have the largest single effect in promoting imbalance. Altogether, a strong association between overtime work and lower levels of work-life balance is found in the literature (Albertsen et al. 2008). Working hours and work characteristics affect the perception of conflict (Stier et al. 2012). A Dutch study found that not only overtime work but also shift work was bad for balance (Jansen et al. 2003), whereas in a teleworker population, the flexibility of working time predicted better perceived work-family balance (Maruyama and Tietze 2012). In a physician subpopulation, Keeton and colleagues (2007) found that control over schedule and work hours were the most important predictors of work-life balance and burnout.

Second, there are demographic variables that show an association with work-family balance. Household work overload clearly has a negative impact. Having full responsibility for housekeeping and caring responsibilities at home are risk factors for the onset of work-family conflict (Jansen et al. 2003; Dex and Bond 2005). In a disproportionate division of household labour, this implies a higher imbalance for coupled women in the so-called “family age,” so that studies found the age group between 36 and 45 years to be characterized by most work-family imbalance (Dex and Bond 2005). For some special subpopulations like doctors, studies did not find either gender or age to have individual impact (Keeton et al. 2007).

Romanian single parents, first and foremost single mothers, find it hard to maintain balance between work and family life (Sănduleasa et al. 2011). Single-parent families also present a high risk of social marginalization and are overrepresented among the poor (Sănduleasa et al. 2011). State allowances to support single-parent families are scarce and particularly maternal single-parent families are households at social risk (Gotea and Busuioc 2016).

Social sciences have been addressing the issue of work-family balance particularly for women and foremost in connection to childbearing plans, childcare and household activities. Not only female but male roles have also been enriched in the last decades. Beside paid work, family and household roles have become increasingly accentuated in the lives of men, so that balance has become an issue for them, too. Although the research on the work-family balance for men is scarce, some studies argue that societal and personal expectations and pressures of being the main breadwinner and having to work long hours on the one hand and men’s desire to spend time with their families on the other, make men face even more stress than women (Evans et al. 2013).

Looking more specifically at the perceptions of work-family conflict, an Australian study found full-time working women as being significantly disadvantaged, facing more challenges and a lack of freedom in managing their work and non-work roles. For this reason, working women lead quite restricted lifestyles (Fujimoto et al. 2012). In the same manner, in European countries, women report higher levels of conflict than men, especially in Eastern Europe (Stier et al. 2012).

On average, the presence of children is associated with a sense of greater conflict for women but this effect is reduced in countries with higher coverage of day care facilities for very young children (Stier et al. 2012). In Romania, from all family types, the poverty rate is highest among families with more than 3 children and secondly, among single-parent families (Iovan and Oprea 2014).

However, not only coupled people but even singles may experience imbalance. A small-scale study conducted in a group of Romanian employees found that out of four types of marital status (unmarried, married without children, married with children under 18, married with children over 18) none had a significantly different level of work-family balance (Panisoara and Serban 2013). Young urban singles also perceived that workload and professional strains endanger their leisure time and work-life balance (Santha 2016). Typically, however, it is single parents who are at increased risk of work-family imbalance (Tausig and Fenwick 2001), single mothers being less happy and more stressed and fatigued in parenting than partnered mothers (Meier et al. 2016).

Third, educational attainment and financial situation are also associated with work-family balance. Here we have contradictory evidence from the United States, where work-family balance was higher among blue collar workers than among professionals (Tausig and Fenwick 2001) but there is also evidence that young people with low educational attainments and poor working conditions experienced more years of family-to-work interference (Ammons and Kelly 2008). When differentiating among professionals, it was found that workers in low white-collar occupations have a lower level of conflict than those in high white-collar occupations but there was no difference between the latter and blue collar workers (Stier et al. 2012).

According to the results of a comparative study on European countries (Anttila et al. 2015), which Romania was not a part of, country clusters showed a clear effect on perceived work-family balance even after controlling for flexibility measurements at the individual level. As such, country clusters gauge institutional differences that are not directly related to work-time flexibility at the individual level. The significance of each regime for the everyday lives of individual employees and families depends also on culture (Anttila et al. 2015) and cultural differences in work-related expectations and values also contribute to the relationships between working times and work-family balance (Pfau-Effinger 2005).

Although not steady in some cases, studies found the following variables as prominent to impact work-family balance: most remarkably, the work-related variables like working hours, shift work and the character of work, then, from demographic variables the gender, the family structure and sometimes the age and social status. The empirical analysis below departs from the assumption that these variables predict the work-family balance of the working age ethnic Hungarian population in Romania.

2. Materials and Methods

The analysis seeks to identify which variables impact upon work-family balance and assesses the magnitude of these effects. Departing from the theoretical framework offered by previous studies in the field, we proceed in a deductive way: variables found to be relevant for work-family balance are tested in order to identify which of them are relevant in the specific context of ethnic Hungarians in Romania.

Hypothesis 1.

It is assumed that work-related variables (working hours, character of work, shift work) cause the most significant differences in work-family balance.

Hypothesis 2.

It is assumed that demographic variables (gender, age, family structure, partnership status, number of children) impact upon work-family balance.

Hypothesis 3.

It is assumed that social status measured by educational attainment impacts upon work-family balance.

The empirical material originates from the survey called Hungarostudy, which was conducted on a sample of Mureș County’s ethnic Hungarian population in 2016, the sample making up 0.19% of the total Hungarian population of the respective county. The survey itself is a Hungary-based large-scale trend survey on physical and mental health. For the first time it was conducted in a Hungarian community outside the borders of Hungary, in the ethnically diverse region of Transylvania, Romania. The questionnaires were completed by interviewers at the respondents’ domicile, so that there are almost no missing data and the very few (two) missing cases were omitted from these calculations. For this analysis, a sample size of 383 persons was used with valid cases on all variable, which entails the working age active adults from the total sample. Data is only available for the Hungarian population and the sample is representative. However, the relatively small sample size calls for caution when drawing generalizing conclusions.

To confirm the aim of this study, in order to identify the determinants of work-family balance and assess their impact, descriptive and explanatory statistics are performed. The work-family balance measure was developed by Grzywacz and Carlson (2007) and first used and validated by Carlson et al. (2009). The six items contain the subjective evaluation of work-family balance and they refer to the extent to which an individual meets negotiated role-related expectations in both the work and family domains. Three items (number 1, 2 and 4) are measuring the individual’s own perspectives on his/her success in meeting work and family demands formulated by others and in other three (number 3, 5 and 6), the perspectives of an external party are tapped in order to capture what other people (people, supervisors, family members and co-workers) expect from the individual in achieving balance (Carlson et al. 2009). The questionnaire items are as follows: 1. I am able to negotiate and accomplish what is expected of me at work and in my family. 2. I do a good job of meeting the role expectations of critical people in my work and family life. 3. People who are close to me would say that I do a good job of balancing work and family. 4. I am able to accomplish the expectations that my supervisors and my family have for me. 5. My co-workers and members of my family would say that I am meeting their expectations. 6. It is clear to me, based on feedback from co-workers and family members, that I am accomplishing both my work and family responsibilities.

All items include a reference to the expectations or negotiation of roles and represent the definition of work-family balance developed by Grzywacz and Carlson (2007) which entails the perspectives of an external party, either evaluated by others or by the individual. This approach is the only measure of work-family balance in the questionnaire. No measure refers to the individual’s own subjective perceptions irrespectively of the expectations of others, like, for instance, does the scale developed by Grzywacz and Marks (2000). Certainly, there might be some limitation in the judgment of others and other more direct measures could have been used to avoid subjectivity of any kind (like Greenhaus et al. 2003). However, the author considers work-family balance a social construct and as such it seems adequate to follow a conceptualization that reckons balance as a result of interactions, as the accomplishment of role-related expectations that are negotiated and shared between individuals and their role-related partners in work and family (Grzywacz and Carlson 2007).

The work-family balance measure is assessed by computing the ratings given to the above six statements on a five-point Likert-type scale ranging from strongly disagree (1) to strongly agree (5). The measurement tool computed from the six items is a scale with values ranging from 6 and 30, where higher scores equate to higher levels of work-life balance. Principal Axis Factor Analysis was conducted on the six items which loaded at 0.73 or above on a single factor. The Eigen value was 4.94 and 81.5 percent of the variance was explained. The Cronbach Alpha for the work-family balance scale indicating the internal consistency of the measure was as high as 0.91. Our measurement tool can thus be considered as valid for this population.

Most demographic, social and work-related variables along which basic descriptive statistics and two-variable associations (ANOVA) are performed, are later included in a linear regression analysis to identify the determinants of work-family balance. Regression assesses the explanatory power of variables one by one, adjusted for the effect of all other covariates. Those situational (work-related), social and demographic agents are being searched for that account for the individuals’ work-family balance. These alleged explanatory variables and their measurement levels are: work hours (used both in its categorical and continuous forms), number of work shifts (continuous variable), the character of work (dichotomous categorical), educational attainment (ordinal level categorical), partnership status (dichotomous categorical), age (used both in its categorical and continuous forms), gender (dichotomous categorical), the number of underage children (continuous), family structure (dichotomous categorical) and partnership status (dichotomous categorical). Dichotomous variable values were recoded into 0 and 1, where 0 always indicates the situation which is supposed to be the unfavourable case: being a woman, being single, being single parent, performing physical work. For the only ordinal level variable, educational attainment it was examined as whether to use it as continuous or categorical variable in the form of dummy variables. The first one assumes linear effect, which might be too rigorous a restriction, while the latter does not use the hierarchy of categories. Technically this was done as follows: in the final model the ordinal variable was tested. Two models were fitted, in the first case the variable was used as continuous, in the second one as categorical variables using dummies. The more parsimonious model was chosen using the AIC information criteria indicator, so that in the final model, education was used in its original continuous form.

During the regression analysis we aimed to find the best fitting and most parsimonious causal model, using the variable selection procedure. Following Kleinbaum et al. (2007) the model was first simplified leaving out the non-significant interaction effects one by one. Thereafter an automatized variable selection algorithm was used and variables were allowed to be sorted out by the algorithm. Stepwise selection was used, setting the threshold at 5% for inclusion and at 10% for exclusion. The final model was subjected to multi-collinearity diagnostics, monitoring the VIF and Tolerance indicators.

It is one limitation of this study that the questionnaire does not contain items on the sharing of household work and caring activities, so that the impact of these upon the balance is impossible to assess. This deficiency cannot be overcome, by reason of the data used. Further, due to the research design, the study was carried out only in the Hungarian speaking population with a specific social and economic situation within the country. This way, our results allow for careful generalizations for the entire Eastern European working age population and the learnings from this understudied population could be relevant for other societies in transition with similar demographic and socioeconomic features.

3. Results

As a first indicator of work-family balance, the boxplot in Figure 1, Figure 2 and Figure 3 sums up the basic distributions of the balance scores across the most important work-related, social and demographic characteristics. The median, the minimum and maximum values, the upper and lower quartile are displayed.

In the figures above, the lower work-family balance scores and larger dispersions of the scores are obvious in the case of some groups like singles, single parents, parents with many children, low educated people, especially with lower secondary education and people with a precarious financial situation hardly coping on their income. Also, for employees working in three shifts as well as those above the weekly work hours of 45, the lowest scores are displayed which never reach the maximum possible score on the work-family balance scale.

To assess whether there are significant differences in the work-family balance across the most important demographic, social and labour market categories, analysis of variance (ANOVA) is performed. Eta-squared shows the percent of variance in the work-life balance scores explained by the independent variables below, measuring the two-variable associations between these and work-family balance scores.

Table 1. Work-family balance mean scores by work-related, social and demographic categories. Analysis of variance (ANOVA) and eta-squared values, N = 383.

	Mean	STD	p	Eta-Squared
Sample total	24.81	4.81
Character of work			0.869	0.000
Blue-collar	24.90	4.97
White-collar	25.03	4.55
Number of work shifts			0.023	0.071
One	26.45	4.26
Two	25.71	4.76
Three (night shift also)	24.38	4.34
Work hours/week			0.711	0.005
Less than 35	25.50	4.30
36–40	25.77	4.32
41–45	24.69	4.40
More than 45	24.15	5.51
Education			0.886	0.009
Primary	24.50	6.50
Lower secondary	24.84	4.91
Secondary	24.88	4.86
Tertiary	24.84	4.26
Subjective financial situation			0.307	0.008
Indebted, living on loan/external support	23.00	7.21
Living on savings	24.75	4.62
Hardly coping on present income	24.21	5.29
Coping on present income but no savings possible	24.52	4.85
Coping comfortably on present income and small savings possible	25.07	4.62
Coping comfortably on present income and substantial savings possible	26.38	3.27
Gender			0.565	0.002
Male	25.01	5.17
Female	24.62	4.29
Age			0.628	0.014
18–24 years	24.58	4.26
25–34 years	23.73	4.37
35–44 years	24.82	4.60
45–54 years	25.30	5.28
55–64 years	24.90	5.85
Partnership status			0.002	0.092
Single	23.26	5.22
Partnered	25.62	4.60
Family structure			0.000	0.162
Single parent	22.75	7.10
Two parents	25.53	4.57
Number of underage children			0.034	0.098
0	26.05	4.28
1	25.30	5.43
2	24.51	4.95
3 or more	23.50	4.50

reveals which population groups are satisfied and, most importantly, which ones suffer from the imbalance. In the two-variable analysis, significant differences are found with respect to the number of work shifts, partnership status, family structure and family size (number of underage children). Employees working in more than one shift, singles, single parents and people with larger families score significantly lower on the work-family scale. The measure of association (eta-squared) is for most variables very small, however, it reveals that work-life balance scores are weakly associated with partnership status (eta-squared = 0.092), the number of under age children in the household (eta-squared = 0.098) and the number of work shifts (eta-squared = 0.071) and moderately associated with family structure (eta-squared = 0.162).

Knowing the main disparities in work-family balance across the studied groups, the question is raised which variables impact individually upon work-family balance, increasing or decreasing it. In the hypotheses we formulated we assumed work-related variables to have the strongest impact and social and demographic variables to be associated with differences, too. Two-variable analyses do not offer a just picture on the issue, as relationships may be spurious, caused by further variables, whereas linear regression assesses the effect of variables controlled for other covariates. In the regression model all independent variables were entered regardless of the descriptive and two-variable results above. Initially, in a first attempt to build a regression model, all variables including financial status and character of work figured in the analysis. This way, however, the explanatory power of the model was very little, most probably due to the dependence of these variables on educational attainment. Thus, in the final model, in order not to reduce the explanatory power, solely educational attainment was included as the indicator of socioeconomic situation.

In the table below, positive numbers indicate a move towards bigger scores on the balance scale (positive association), negative numbers mean that moving away from the reference value or smallest value of the explanatory variable causes a decrease in the score (negative association).

Seen altogether, the regression model displayed in

Table 2. The determinants of work-family balance. Linear regression, final model, N = 383.

Explanatory Variables	Unstandardized B	p	Standardized β
Family structure (single parent–two parents)	3.218	0.000	0.282
Number of underage children in the family	−0.701	0.000	−0.223
Number of shifts	−0.539	0.002	−0.131
Educational attainment	0.188	0.021	0.116
Gender	−0.204	0.865
Age	−0.068	0.377
Partnership status	0.120	0.437
Work hours	0.192	0.347
Constant	30.720	0.000
Adjusted R2 = 0.16, F = 36.680, p = 0.021

is significant. Explanatory variables account for 16% of the variance in work-family balance (adjusted R² = 0.16). This explanatory power is not too strong, however, the most important agents have been identified that impact upon the balance in the individuals’ lives in this specific population group. Demographic variables explain by far the largest part of work-family balance variance and only secondly do work-related aspects play a role. Observing the absolute Beta (β) values it can be stated that family structure has the most powerful effect upon balance (β = 0.282), followed by the number of underage children (β = −0.223). The status single parent is most unfavourable for balance: on the scale ranging from 5 to 30, a partnered parent scores more than 3 points higher than a single parent (unstandardized B = 3.218). With every child in the household, balance decreases with 0.7 points (unstandardized B = −0.701).

From work-related determinants, only the number of working shifts has an effect, a negative one, upon work-family balance (β = −0.131). Compared to those working in one shift only, employees in two shifts obtain half a point less on the balance measure (unstandardized B = −0.539), whereas those in three shifts are even more disadvantaged.

Fourth in the raw, educational attainment also accounts for a small part of the work-family balance disparities (β = 0.116). With the raising of one level in educational attainment, the score on the balance scale increases slightly (unstandardized B = 0.188) and this means a difference of less than one scale point between the lowest and the highest educational levels. The explanatory power of educational attainment is quite low and it produces relatively small disparities, especially when compared to the leading demographic variables, family structure and number of children in the family.

None of the other variables which were supposed to have an impact on balance or in two-variable analyses were associated with disparities (like partnership status) proves to be significant when adjusted to covariates.

Hypothesis 1 was refuted by our data: not all work-related variables, only shift work, impact upon balance, and, contrary to the expectations, it is not the most powerful determinant. Hypothesis 2 was partially supported by our data: not all demographic variables play a role but family structure and the number of underage children did prove to have a significant controlled effect upon work-life balance. Data supported hypothesis 3 entirely: although educational attainment had the least explanatory power of all, it does differentiate among active adults, so that all other conditions being the same, people with higher education achieve higher balance than those with low education.

4. Discussion

Descriptive results and two-variable associations indicate some differences in the work-life balance across population groups. The results of the regression analysis (Table 2) partly reinforce our findings and those of previous studies, partly enrich this knowledge with new insights with respect to a specific population. Contrary to the findings of other studies that revealed the primacy of situational, work-related agents (Dex and Bond 2005; Keeton et al. 2007; Albertsen et al. 2008; Pereira and Coelho 2012; Albertsen et al. 2014), in the ethnic Hungarian adult active population of Mures county, Romania, demographic variables are the ones that have the strongest impact upon work-family balance. In spite the structural disadvantages which imply long working hours for this population, work overload in itself does not cause imbalance. It is single parents and families with more children that are at highest risk of imbalance.

The crucial issue in work-family balance and its most powerful determinant in the active age Hungarian population of Romania is whether a parent is single or partnered, that is, raising his/her child alone or not. Empirical studies on single parenthood have pointed out that single parents are at high risk of imbalance (Tausig and Fenwick 2001) and that they experience significant emotional detriments in parenting compared to other mothers, even after accounting for key social and demographic differences (Meier et al. 2016). Although in most cases single parents are mothers, our findings statistically demonstrate that it is not gender and not female work overload that causes imbalance but rather the status single parent and being a single parent is equally unfavourable for mothers and fathers

The number of underage children in the household is the second most powerful agent that appears in time management: the more caring responsibilities, the less balance. International studies have also found caring responsibilities to reduce balance (Jansen et al. 2003; Dex and Bond 2005; Stier et al. 2012). The hardships of single parents and larger families is an expected result, in line with the evidence from other studies, yet the primacy of family composition over labour market features is a new and obviously atypical phenomenon of undoubtable relevance.

Although based on the existing literature (Jansen et al. 2003; Dex and Bond 2005; Albertsen et al. 2008; Albertsen et al. 2014; Anttila et al. 2015) it was assumed that long working hours and shift work have a negative impact upon balance and that physical work is associated with less balance (Ammons and Kelly 2008), from work-related variables entered in the analysis only the number of work shifts have significant impact: work performed in changing times has a negative impact on work-family balance. Among those dissatisfied with their work-family balance, working in multiple shifts was more frequent. For balance, even long working hours are not as unfavourable as the change of shifts and most of all, working unsocial hours like night shift. The negative effect of unsocial work time was assessed by international studies (Pereira and Coelho 2012; Albertsen et al. 2014).

The explanatory power of educational attainment and the direction of the association was not found to be steady in international studies (Ammons and Kelly 2008; Tausig and Fenwick 2001). Although relevant for balance, the impact of the level of education is far behind that of the other, more powerful determinants. It seems, however, that educational investments pay off in this population, too, even with respect to work-family balance, as people with higher education achieve better scores.1

Those professions that do not allow for a wishful work-family balance and score under the mean of the scale are: construction worker, cleaning staff, lawyer, carpenter, dressmaker, unskilled factory worker, courier, saleswoman, agricultural worker. These jobs were listed that be articulately hard to reconcile with family life. The professions enumerated as problematic with respect to balance are mostly blue-collar jobs, however, there are some intellectual professions as well. The character of work (blue collar/white collar) was not included in the linear regression model, due to its strong correlation with educational attainment. Thus, the differences caused by character of work operate through educational attainment or shift work, the latter being predominant in blue collar jobs.

In the literature, the relationship between work hours and balance is marked by contradictory findings (Pereira and Coelho 2012). Work hours have stronger negative effect in some countries and weaker in others (Anttila et al. 2015). In some cases, work hours in themselves were found not to have a significant relationship with individual wellbeing. Our evidence is in line with these latter findings: although in descriptive analysis long working hours were associated with slightly less balance (the differences were, however, not significant), when adjusted to covariates in the linear regression, longer working hours do not decrease balance. Among those dissatisfied with their work-family balance, the average work time is about half an hour longer than that of their fellows.

Contrary to our expectations based on tremendous empirical evidence from the literature, within our ethnic Hungarian sample in Romania, there is no detectable difference between genders with respect to work-family balance: all conditions being the same, women are not disadvantaged and perceive to achieve work-family balance just as well as men do. It is, however, a fact that starting with marital age, women in Romania spend three times as long time with household and care activities than do men (INS 2013). In spite of the house and care work overload of women and the existing gender inequalities in these activities, gender in itself has no statistically significant impact on work-family balance.

In our sample, partnership status does not affect balance. Recently in Romania Panisoara and Serban (2013) also found the same independence of marital status of balance. Although in some previous research, the age was found significant and the most difficult time to reconcile work and family life was the “family age” between 36 and 45 years (Dex and Bond 2005), in our sample, balance scores are not significantly different in any age group and when controlled for covariates, age has no impact.

5. Conclusions

This study contradicts the scholarly knowledge on the determinants of work-family balance in the ethnic Hungarian population of Romania and most probably in other populations with similar characteristics, primarily in Eastern Europe. The social situation of this this specific ethnic group has not sufficiently been researched. The scientific relevance of having studied this population is that Hungarians in Romania have comparatively harder socioeconomic and working conditions to the Romanian majority. The findings contribute to better understand the inequalities in work-family balance in populations with similar features, that is, in transitional societies with almost integral and exclusively full-time male and female employment, where traditional and modern gender role expectations coexist. The main novelty of this study consists in revealing the primacy of family structure in reaching work-family balance as opposed to the majority of international studies which, besides acknowledging the importance of family features, emphasized work-related situational determinants as most powerful.

The explanatory power of variables was assessed independently controlling for covariates and it was predominantly the family composition variables that explained the differences in work-family balance in active adults. Some allegedly important predictors like being a woman or having longer working hours did not impact in themselves upon balance, while others led to considerable differences. Most importantly, there were immense differences between single and two-parent families, between childless people and parents with more children. From work-related variables, only shift work was differentiating among employees, so that shift workers are at a high risk of imbalance.

From expected predictors, the lacking effect of gender is notable. The literature on work-family balance has examined first and foremost the relationship between gender and balance and the results of different studies are conflicting. Independently from the societies where the topic was studied and of the type of workers’ samples used, in most cases it was the women who reported imbalance at higher rates (Torres et al. 2011) and in fewer cases there was no significant difference between genders (Keeton et al. 2007; Fapohunda 2014). However, from similar or even more traditional societies than Hungarians in Romania, there is also some evidence that both genders believed there was a negative spill over from work on to family life and men and women indicated alike that they had work-life imbalance (Doble and Supriya 2010; Fapohunda 2014). The results of the analysis presented in this paper report similar levels of work-family balance for men and women when controlled for the main social, demographic and work-related features. It is a key finding that all other conditions being the same, gender in itself does not cause lower levels of imbalance for women. In this particular case of the Hungarians in Romania, a population with economic hardship in a country in transition, despite the uneven distribution of household and childcare activities revealed by time use surveys, notwithstanding the long unpaid working ours for women, they account of no more imbalance than do men.

Explaining this contradictory finding would exceed the limits of this study. Nevertheless, it should be briefly pointed out that it has been a typical attitude of Hungarian and more generally, Eastern European women in the past decades to perceive their unpaid work overload as fair, basically as a manifestation of the care for their families and they have deemed this uneven division of household labour as satisfactory even in the conditions of their full-time employment. One possible explanation for the persistence of gendered household division of labour in this region is precisely the women’s volunteer household work overload that they claim as fair also in order to compensate—according to a psychoeconomic theory—for the higher earnings of their husbands (Shelton and John 1996; Blaskó 2006). Further discussion on the lack of expected statistical association between gender and balance in this population should also rely on qualitative findings on this topic.

Although no gender differences were revealed, the analysis sheds light on the notable role of family composition in work-family balance. The key findings of our study are that the status single parent carries by far the biggest risk for perceived imbalance and that raising one more child compared to fewer decreases balance, too. These insights are important for health, employment and family policy and they point far beyond scientific considerations. The difficulties of single parents and large families justify the necessity of further addressing their situation as a priority issue.