Gender Socioeconomic Inequalities in Intimate Relationships: Introducing the GSEI Index

Inspired by the materialist feminist theoretical approach, this study starts from an underlying assumption that research of gender socioeconomic inequalities in intimate relationships requires the analysis of specific variables that constitute the socioeconomic status of intimate partners. Based on five indicators relevant for contemporary Croatian society, the GSEI index was constructed and applied to a representative sample of Croatian women in intimate relationships. The results showed significant gender socioeconomic inequalities between intimate partners, mostly in favour of men. The highest level of inequality between intimate partners was discovered with regard to real-estate ownership, whereas the highest level of equality between partners was found in their level of education. Although socioeconomic inequalities between intimate partners in favour of men prevail regardless of the geographic and social context, they are more pronounced in regions with more traditional values, among couples who live in rural areas, and among those who are married (in comparison with those who live in domestic partnership). A recommendation for further research and GSEI index application is to consider which are the relevant indicators for measuring gender socioeconomic inequalities in intimate relationships depending on the wider social context.


INTRODUCTION
Gender socioeconomic inequalities are a type of social inequality that can be defined as unequal distribution of social and economic resources between women and men.Classical sociological approaches to social inequalities are mostly focussed on the analysis of social classes and social status (area of social stratification).Inequalities are considered in terms of social relations in which individ-uals are more or less privileged (Goldthorpe, 2012).The most common indicator of social position in social stratification is socioeconomic status.Graham claims that socioeconomic status is both structurally imposed and socially produced (Graham, 2007: 36).Feminist theoreticians (for example, Walby, 1990) point out that classical sociological theories ignore gender economic inequalities in the private sphere (such as gender division of labour within the household).Contemporary sociological thought gives much more attention to gender economic inequalities by, for example, introducing the concepts of the second shift in the private sphere (Hochschild and Machung, 2003) or glass ceilings and glass walls in the labour market (Reid, Kerr and Miller, 2003).Studies show that gender socioeconomic inequalities correlate with intimate partner violence (Costa et al., 2013) and poorer health (Hassanzadeh et al., 2014).
The inspirational theoretical source of this paper are feminist materialist perspectives which argue for the relevance of the socioeconomic dimensions of gender inequalities and focus on material social inequalities between women and men.Although the third wave of feminism was mostly characterised by a theoretical shift from the materialist to cultural and symbolical focus, not all feminist approaches neglected the material, everyday aspects of life in order to understand the role of gender in the social relations of modern capitalist society (Jackson, 2001).Accordingly, this paper starts from an underlying assumption that research of gender socioeconomic inequalities in intimate relationships requires an analysis of specific variables that constitute the socioeconomic status of women in comparison to their intimate partners.
In search of indicators of gender socioeconomic inequalities in intimate relationships, I begin with a brief discussion on measuring socioeconomic status (SES).Then I describe some basic social indicators of gender inequalities in Croatia, which constitute the social context within which I will propose a constructed index.Based on these insights I argue for the relevance of five indicators of gender socioeconomic inequalities in intimate relationships: level of education, monthly income, employment status, real-estate ownership and car usage of each partner.One main and three derived hypotheses have been set based on the findings of previous research.The main contribution of this paper is to assess gender socioeconomic inequalities in intimate relationships in Croatian society by means of specific SES indicators and a newly constructed measure called the GSEI index.

SOCIECONOMIC STATUS IN GENDER PERSPECTIVE -METHODOLOGICAL ISSUES AND THE CROATIAN CONTEXT
Composite socio-economic indicators aim to measure complex, multidimensional phenomena, which cannot be measured directly, and are constructed by aggregating a number of individual indicators into one composite measure (UNECE, 2019: 67).SES can be measured as a univariate or as a composite construct.The most common variables for measuring SES in general are level of education, profession, and income.Different studies of SES show the diversity of approaches depending on the aims and the subject of the research, and they often take into account different contextual variables.For example, in their study on women's status and health in India, Ghosh and Bharati (2005) use the following contextual variables as indicators of socioeconomic status of women: literacy, poverty level, employment status, type of work and distance from the workplace.In a study on socioeconomic impacts on health in the Caribbean and Central America, Ligeon, Jolly, and Jolly (2012) use income, education, owning a TV, access to sanitary facilities and clean water.In a study on the impact of SES and anomie on illegal behavior in Iran, Heydari et al. ( 2013) use different subjective measurements with ordinal assessment of the values for father's education and occupation, family income, value of property/ possibility of purchasing and self-evaluation of class affiliation.In their study on the association between intimate partner violence and socioeconomic status in different European cities, Costa et al. (2013) use educational level, occupation, unemployment time and frequency of worries about daily expenses as indicators of socioeconomic status.Gregurović and Kuti (2010) studied the effect of socioeconomic status on student educational achievement in Croatia.They used several different indexes of socioeconomic status: International Socio-Economic Index of Occupational Status (ISEI) (Ganzeboom, De Graaf and Treiman, 1992), Index of Family Educational Resources and Index of Cultural Possessions, along with four additional SES variables: average monthly household income, average monthly education expenditures, highest level of parental education, and number of books in the household.These examples illustrate how SES indicators can vary depending on the research context and their relevance for studying a topic of interest.
Gender socioeconomic inequalities in Croatian society are present in both the public and private spheres.Women in Croatia are generally less educated than men (DZS, 2018: 24), although there is a trend of reducing these inequalities, identified in the latest 2011 census that shows, for the first time, that the portion of highly educated women somewhat exceeds the portion of highly educated men, though there are still more women than men with the lowest educational status.Highly educated women are mostly educated in the field of social welfare services, edu-cation of teachers and educators, and educational and biological sciences, while they are the least educated in the most profitable areas: engineering and computer sciences.Women traditionally dominate in the Croatian educational system, but as the education level increases, the proportion of women among educators becomes smaller (DZS, 2018: 36).With regards to employment, women constitute the majority of the unemployed (55%) (HZZ, 2019: 10) and are in the minority when it comes to the employed population in Croatia (47%) (HZZ, 2018).Compared to men, they have lower average salaries (the proportion of women salaries in salaries of men is 86.8% (DZS, 2018: 43).Women in Croatia have less political power than men: they are extremely under-represented in all political bodies, both at local and state levels.Gender inequalities are particularly pronounced in family life.Women enter the institution of marriage three years earlier, on average, than men do (at the age of 28) (DZS, 2018: 15). 1 Distribution of housework between spouses/partners in Croatian families is highly consistent with stereotypical gender roles (Čulig, Kufrin, and Landripet, 2007;Bijelić, 2011;Klasnić, 2017).Gender inequalities are also evident in childcare -in Croatia, it is quite unusual for fathers to use the right to parental leave and childcare is predominantly perceived as female work (Čulig et al., 2007) -and in choices of career or leisure activities (Kamenov and Galić, 2011;Kamenov and Jugović, 2011).Gender-based violence in families and in intimate relationships is a significant problem in Croatian society.National representative studies show that between 21% (Otročak, 2003) and 23% (Klasnić, 2013) of women are victims of physical violence committed by their intimate partner.A newly recognised form of domestic violence is economic violence against women, which is even more present than physical -approximately 25% of women in intimate relationships are victims (Klasnić, 2013).One of the main reasons why women do not leave abusive relationships is economic dependence on their partners (Maslić Seršić, 2010).

CONCEPTUAL FRAMEWORK FOR GSEI INDEX
As already pointed out, SES measuring should take into account indicators relevant to a particular research context and topic of interest.Despite the lack of specific data on gender socio-economic inequalities between intimate partners, the conceptualisation of the GSEI index started based on available data on gender inequalities in Croatia.The development of the conceptual model for this index is therefore immersed in the described context of Croatian society.However, later on, I will also discuss how this index can be applied in other societies.

1
Data for 2016, according to Croatian Bureau of Statistics women enter the first marriage at the average age of 28.3, and men at the average age of 31.1 years.
In this study, five indicators were considered as relevant for the assessment of gender socioeconomic inequalities between intimate partners in the cultural and economic context of contemporary Croatian society.The term "relative" in the first three indicators refers to the comparison of intimate partners for each one.
1. Relative level of education -as a generic indicator of SES, education is thought to specifically capture the knowledge-related assets of an individual (Shaw et al., 2007).The level of education is often a critical determinant of labour market opportunities and wages and, as a consequence, the economic situation of a person (UNECE, 2009).A higher level of education also contributes to civic engagement, more personal freedom and choices, and better health and well-being (Brennan, Durazzi and Séné, 2013).
2. Relative amount of monthly income -income is the indicator of SES that most directly measures material resources.As with other indicators such as education, income has a predictive association with health.Income also has a cumulative effect over the course of life and is the SES indicator that can change the most on a short-term basis (Shaw et al., 2007).
3. Relative employment status -although employment status cannot, by itself, cover the complexities of gender differences in access to economic opportunities, the comparison of the partners' employment status is a rough indicator of potential differences in their access to earnings, professional reputation, and social opportunities.
4. Real-estate ownership -ownership of and control over assets play an important role in determining the socioeconomic status of women and men.Real-estate ownership represents a store of value, and can provide safety in periods of economic difficulty.Real-estate ownership provides decision making authority over assets which also enhances social status and increases bargaining power (UNECE, 2009). 5. Car usage -car usage is an indicator of mobility and has potential implications for the socioeconomic opportunities of the user.Car usage can facilitate in finding and maintaining employment, save travel time to work, schools, health care appointments, etc., thereby freeing time for other activities (UNECE, 2009).Car ownership and access (whether a person has access to a car or van in their household, even if they do not own it) is a commonly used asset-based indicator of socioeconomic position (Shaw et al., 2007).
The main hypothesis of this paper argues that intimate relationships in Croatian society are characterised by significant gender socioeconomic inequalities between intimate partners, mostly in favour of men.
To get a better understanding of how the gender socioeconomic inequalities in intimate relationships are distributed in Croatian society, three derived hypotheses are set and tested in this paper.H1: Gender socioeconomic inequalities are not equally distributed across different geographical regions of Croatia.
A higher degree of gender socioeconomic inequalities is expected in regions with more pronounced traditional values, a higher degree of sexism, lower awareness of gender inequalities and gender-based discrimination, and less pronounced egalitarian attitudes.Based on previous research of Croatian regions, we expect higher inequalities in the Slavonia and Dalmatia regions and lower in the Istria and Primorje region and in the Zagreb region.Ilišin and Gvozdanović (2016: 180) argue that traditionalism is a highly recognisable latent dimension of social values in Croatia, in which religion, family, and nation represent mutually interrelated identitary pillars.In their study of Croatian youth, they found that those traditional values were expressed the most in Dalmatia and eastern Croatia (that is, the Slavonia region), and the least in the Zagreb region and the Istria and Primorje region.According to the European Values Study data for Croatia from 2018 (Baloban, Črpić, and Ježovita, 2019), the value of religion is most important for residents of Slavonia (76.5%) and northeastern Croatia (71.2%), and the least important in Istria and Primorje (41.7%) and in City of Zagreb (47.7%).Galić (2004) found that residents of the Dalmatia and Slavonia region show a higher degree of a specific sexist orientation called "anthropologically grounded patriarchy in family and public gender relations" in comparison to residents of Zagreb and the Zagreb region.Additionally, residents of Zagreb and the surrounding region had the highest need for the suppression of patriarchy and establishment of gender egalitarianism in Croatian society.In another study from 2010 Galić (2012) found similar results.Residents of Dalmatia, Slavonia, Istria, Lika, and Banovina showed a higher degree of androcentrism than residents from Zagreb, Primorje, and Međimurje (Galić, 2011a).Regional differences also exist in the degree of awareness of gender inequalities and gender-based discrimination in Croatian society (Ajduković, 2011).Residents of Istria, Primorje and Gorski Kotar show a higher degree of awareness than residents in Dalmatia and Zagreb.
Furthermore, Jugović and Baranović (2011) found that residents of northern Croatia and Zagreb have more pronounced egalitarian attitudes about gender educational inequalities than residents from the south of Croatia (Istria, Primorje, and Gorski Kotar and Dalmatia).While analysing gender inequalities in the labour market, Galić (2011b) found that tendencies to gender discrimination in employment are less pronounced in Istria, Primorje and Gorski Kotar, the Zagreb region, and Slavonia than in northern Croatia and Dalmatia.H2: Gender socioeconomic inequalities are not equally distributed in rural and urban areas.
Based on the findings of previous research about gender division of household labour, gender discrimination in family relations and educational system, a higher degree of gender socioeconomic inequalities in favour of men is expected among couples from rural areas.
Gender inequalities in the private sphere are the most obvious when it comes to division of labour in the household, and this particularly refers to the rural population of Croatia.Studies show that among the rural population men still play a larger role in making important family decisions and managing the family money -rural areas retained the traditional understanding of a man as the "head of the family" (Kamenov and Jugović, 2011).Rural populations also experience more gender discrimination against women in family ( Kamenov and Jugović, 2011), they show a lower degree of awareness of gender inequalities and, consequently, are less sensitive to gender discrimination (Ajduković, 2011).Traditional attitudes of rural populations also reflect in lower support for women's rights.Urban populations are, for example, more prone to support a woman's right to in vitro fertilisation regardless of the woman's marital status, while residents of rural areas are more prone to support the idea that embryos should be protected by the state regardless of the woman's will and that the decision on pregnancy should be made by men (Galić, 2011a).Residents of rural areas experience gender-based discrimination in the educational system more often than those in urban areas.More precisely, people from rural areas are often being advised not to engage in a certain business due to their gender, are more likely to have experienced offensive comments based on their gender and sexual harassment by their teachers (Jugović and Baranović, 2011).

H3: Gender socioeconomic inequalities are not equally distributed across different types of intimate relationships (marriage vs. domestic partnership).
A higher degree of gender socioeconomic inequalities in favour of men is expected among married couples in comparison to couples living in domestic partnerships.The possible relation between the type of intimate relationship and gender socioeconomic inequalities may seem less self-explanatory than the aforementioned differences based on geographical region and type of place of residence.However, the underlying confounding reason for this relation is assumed to be the fact that married people in Croatia generally have more pronounced traditional values and attitudes in comparison to the ones living in domestic partnerships.In addition, domestic partnership in Croatia is very unusual in older generations, 2 which means that age could also be the confounding variable in this relationship.
In a study on the socio-cultural identity of the Croatian population, Cifrić (2013) found that people who are married and ones who are widows/widowers value their homeland and territory as more important components of their identity than people who are single.In addition, married people and widows/widowers have a more pronounced national-religious dimension of their socio-cultural identity than people who live in domestic partnerships or people who are single.Galić (2012) found that people who are married generally accept patriarchal values more than the ones who live in domestic partnerships.

Method
The data used in this paper were obtained as part of a broader project on the socioeconomic status of women in intimate relationships and women's experiences of economic violence carried out by their intimate partners 3 .Empirical research was conducted using a survey method during September and October of 2012 on a representative sample of adult Croatian women with a minimum of one-year experience of living in the same household with a current or a former intimate partner.The institutional research ethics committee 4 had approved the study.
The survey technique was computer-assisted personal interviewing (CAPI), with only female professional interviewers, additionally trained for interviewing women on sensitive gender topics.Informed consent was obtained from all individual participants included in the study.
For a 95-percent confidence interval and a margin of error of +/-4%, the required sample size was estimated at 600 participants. 5The sample was weighted 2 According to Census 2011, in the age group 60 years and older, 1.6% of men and 0.8% of women live in domestic partnership, while 74% of men and 40% of women are married.(DZS, 2011) 3 Project "Naming, framing and blaming the economic violence against women in intimate relationships".Contracting Authority: European Commission -Western Balkans and Turkey; Civil Society Facility: Partnership actions -Empowerment of Women.

Study participants
The study participants were 601 women from 18 to 89 years of age, with an average age of 50 years (sd=16.6).The majority of women were referring to their current marital partner (69.4%), 19.8% to their ex marital partner, 7.5% to their current domestic partner, and 3.2% to their ex domestic partner.The duration of living in the same household with their partner ranged from one to 66 years, with an average duration of 16 years (sd=16 years).For women who were referring to their ex-partners (23% of all women), the relationship ended between 1 and 55 years ago, with an average of 12 years ago (sd=10.6 years ago).The majority of women in the sample (83%) had a child/children with their partner.Less than 48% of women were employed (of which 80% were employed on a permanent basis, 15% on a fixed-term contract, 2% were self-employed private entrepreneurs, and 3% worked on an agricultural family farm).One in five women (21%) had no monthly income.These were mostly unemployed women (57%) and housewives (39%).About 9% of women in the sample had a very low monthly income (less than 1500 Croatian Kuna, approximately 200 Euros).Most of them were unemployed (22%), retired (26%) or housewives (26%), but almost one-quarter of them were employed (24%).Approximately 26% of women had a monthly income between 1500 and 3000 Croatian Kuna (200 to 400 Euros), the majority of which were employed (45%) or retired (33%).
or domestic partners, divorced women and widows was estimated to 1,365,000 women in Croatia.This estimation did not take into account the criteria of a one-year minimum of living in the same household.

6
The data used for correction of the demographic characteristics of the sample were obtained through another survey, conducted on a representative sample of the Croatian population.All women were asked a question: "Are you now, or have you ever, during your life, been at least one year in marriage or domestic partnership?".Through this question, data were obtained on demographic characteristics (age, education, size of place of residence and region) of women over 18 years old who have at least once in their lifetime cohabitated with an intimate partner for at least one year.These data were used for weighting of data in our sample.

Self-assessment of family material status
Much worse than the majority 2.7 Somewhat worse than the majority 7.4 Like the majority 70.9 Somewhat better than the majority 15.4 Much better than the majority 2.7 User or system missing 1.0 Although the participants in this study were women only, since the women were providing answers on objective SES indicators referring to both themselves and their intimate partners, the unit of analysis in this paper is an intimate heterosexual relationship.

Measuring instruments
In order to measure the gender socioeconomic inequalities (GSEI) in intimate relationships I designed the GSEI index consisting of five indicators, relevant in the cultural and economic context of contemporary Croatian society: 1. relative level of education, 2. relative amount of monthly income, 3. relative employment status, 4. real-estate ownership, 5. car usage.
All five variables represent a comparison of a man and a woman who are in an intimate relationship sharing a household for a minimum of one year.It is important to note that these were not the original variables in the questionnaire.These variables were constructed by either recoding the original variables from the questionnaire (in this way the relative amount of monthly income, real-estate ownership and car usage indicators were obtained) or by combining responses to two original variables from the questionnaire (in this way the relative level of education and relative employment status indicators were obtained).The original variables with a description of collected data are shown in Table 3.
The results on the GSEI index were computed by summing up the data on five constructed variables with the following coding: The result on the GSEI index (GSEI score) indicates the direction and the strength of gender socioeconomic inequalities between intimate partners.The value of zero describes equal socioeconomic status between intimate partners.Negative results (from -1 being the lowest degree of inequality to -5 being the highest degree of inequality) describe intimate relationships in which the woman has a superior socioeconomic position in relation to her partner.Positive results describe intimate relationships in which the man has a superior socioeconomic position in relation to his female partner (from 1 being the lowest degree of inequality to 5 being the highest degree of inequality).

Metric Characteristics of the GSEI Index
The factorial validity of the index was determined by an exploratory factor analysis with a principal component method of extraction of factors, Guttman-Kaiser ('eigenvalues greater than one') criterion for extraction of common factors and direct oblimin transformation of the component matrix.Two retained factors explained 56% of total instrument variance.The first factor was mostly saturated with the variables relative employment status and relative amount of monthly earnings, with an eigenvalue of 1.599, while the second factor was mostly saturated with variables car usage, relative level of education and real-estate ownership with an eigenvalue of 1.275.Inter-factor correlation was r=0.179.The assessment of criterion-based validity of the constructed GSEI index based on three derived hypotheses will be tested and discussed further in this paper.The rationale for this approach is the premise that if the GSEI index measures what it is supposed to, then the analyses should result in expected differences between groups according to previous studies of related social phenomena.
Regarding reliability, the constructed GSEI index did not have high internal consistency (Cronbach's Alpha was 0.449), but since it consists of only five items with very small variability (only three possible values on each item), it was not considered as a statistical or methodological problem.Another possible reason for relatively low internal consistency is the fact that these five variables can be seen as two-dimensional space in which the relative amount of monthly income and relative employment status are mutually highly correlated and form one factor, while the remaining three variables are grouped together in the second factor.
Since there was a statistically significant deviation of the results on the GSEI index from the normal distribution (one-sample Kolmogorov-Smirnov test, p<0.001)only nonparametric statistical tests were used (Kruskal-Wallis and Mann-Whitney U).

RESEARCH RESULTS
Table 3 shows a description of answers on the original variables used to construct indicators of gender socioeconomic inequalities between intimate partners.By combining and recoding these variables, the main five indicators of gender socioeconomic inequalities in intimate relationships were obtained (Figure 1).Car is equally driven by woman and her partner 144 23.9 Each have their own car 79 13.1 User or system missing 2 0.4 The data show that approximately half of women have the same educational level as their intimate partner, and the proportion of women who are more educated than their partners is approximately the same as the proportion of women who are less educated than their partners.However, the distribution of monthly income between intimate partners is very asymmetrical.Only 7% of women have a higher monthly income in comparison to their partners, and in more than 40% of relationships, men have a higher income.Based on employment status, the majority of intimate relationships are characterised by gender equality (65%).Still, for the remaining 35%, it is clear that men usually have a superior position in relation to their female partners.Ownership of the real-estate in which the couple lives is the most pronounced indicator of gender inequality in intimate relationships in Croatia.The majority of couples (53%) live in a space that is owned by the man or a member of his family.In 28% of intimate relationships, partners have mutual ownership of their living space, or some other equally favourable arrangement, and only 19% of cou-ples live in a space that is owned by the woman or a member of her family.Finally, car usage also shows considerable gender inequalities in intimate relationships.Although 59% of couples either do not own a car (22%) or they use it equally (in 13% of relationships each partner has their own car, and in 24% they equally share one), in the remaining 41% of couples the vast majority shows that the car is mostly or exclusively driven by men.
Figure Based on the results on the GSEI index (M = 1.18,SD = 1.80,Mdn = 1, IQR = 0 -2), this study shows significant gender socioeconomic inequalities between intimate partners in Croatian society, mostly in favour of men.Taking into account that negative values on the GSEI index describe the relationships in which the woman has a superior socioeconomic position in relation to her partner, the obtained results show that this is the case in only 18.4% of intimate relationships.On the other hand, positive results describe the relationships in which the man has a superior socioeconomic position in relation to his female partner, and this was the case in 63.8% of relationships.Only 17.8% of relationships have the score zero, which indicates equal socioeconomic status between intimate partners.
If we take a milder criterion and declare relationships with results in the range from -1 to 1 (56.8%) as socioeconomically equal, a notable difference in the remaining distribution persists: 6.6% in favour of women and 46.3% in favour of men.To test the hypothesis that gender socioeconomic inequalities are not equally distributed across different geographical regions of Croatia due to their cultural and social differences, the Kruskal-Wallis test was used with the GSEI score as a dependent variable (see Table 3).The test results led to the rejection of the null hypothesis (χ 2 (5) = 33.58,p < 0.001).Pairwise comparisons showed statistically significant differences between six pairs of regions (see Figure 3).Slavonia and Dalmatia have a higher degree of gender socioeconomic inequalities in intimate relationships in favour of men than the Istria, Primorje and Gorski Kotar region (p = 0.001 and p = 0.002 respectively), northern Croatia (p = 0.005 and p = 0.006 respectively), and the Zagreb region (p = 0.003 and p = 0.004 respectively).The third hypothesis argued that gender socioeconomic inequalities are not equally distributed across different types of intimate relationships (marriage vs. domestic partnership).It was tested using the Mann-Whitney U test with the GSEI score as a dependent variable.The test result led to the rejection of the null hypothesis (U = 14550, p = 0.021).It has been found that gender socioeconomic inequalities in favour of men are higher in married couples (Mdn = 1, IQR = 0 -3) than in domestic partnerships (Mdn = 1, IQR = 0 -2).
Figure 5. Box-and-whisker plot for GSEI score by type of intimate relationship

DISCUSSION
Inspired by the materialist-feminist theoretical approach, this study starts from an underlying assumption that research of gender socioeconomic inequalities in intimate relationships requires an analysis of specific variables that constitute the socioeconomic status of women in comparison to their intimate partners.Thus, I analysed five indicators of gender socioeconomic inequalities in intimate relationships relevant in a cultural and economic context of contemporary Croatian society: relative level of education, relative amount of monthly income, relative employment status, real-estate ownership, and car usage; and finally created the GSEI index by summing the results for these five variables.The unit of analysis was a (heterosexual) intimate relationship.
The main hypothesis of this paper argued that intimate relationships in Croatian society are characterised by significant gender socioeconomic inequalities between intimate partners, mostly in favour of men.This hypothesis was confirmed.The highest level of inequality between intimate partners was detected with regards to real-estate ownership, followed by monthly income, whereas the highest level of equality was detected when considering the level of education.Based on the results on a constructed GSEI index, this study showed significant gender socioeconomic inequalities between intimate partners in Croatian society, mostly in favour of men.The results showed that in only 17.8% of intimate relationships the partners have equal socioeconomic statuses; in 18.4% women have a superior socioeconomic position and in 63.8% men have a superior socioeconomic position.If we take a milder criterion for the interpretation of the results on the GSEI index and declare relationships with results in the range from -1 to 1 (56.8%) as socioeconomically equal, a notable difference in the remaining distribution persists: 6.6% in favour of women and 46.3% in favour of men.These results are in accordance with all other indicators of socioeconomic status of women in Croatian society, in both the private and public sphere, which can all be seen as symptoms of deeply rooted gender inequality.
To get a better understanding of how gender socioeconomic inequalities in intimate relationships are distributed in Croatian society, I tested three derived hypotheses, grounded in previous sociological research.These tests also served as validation for a constructed index -if the GSEI index measures what it is supposed to, then the analyses should result in expected differences between groups according to previous research of related social phenomena (criterion-based validity).
The first hypothesis argued that gender socioeconomic inequalities are not equally distributed across different geographical regions of Croatia.This hypothesis was partly confirmed -I found differences between some regions, but not all of them.The results showed that Slavonia and Dalmatia have a higher degree of gender socioeconomic inequalities in intimate relationships in favour of men than the Istria, Primorje, and Gorski Kotar region, northern Croatia, and the Zagreb region.Previous sociological studies showed that Slavonia and Dalmatia have a higher degree of sexism (Galić, 2004;Galić, 2012) and androcentrism (Galić, 2011a) compared to some other regions, residents of Dalmatia show lower awareness of gender inequalities (Ajduković, 2011), while the value of religion is most important for residents of Slavonia (Baloban, Črpić, and Ježovita, 2019).Since residents of these Croatian regions score higher on traditional values and discriminatory attitudes, these findings indicate that gender socioeconomic inequalities in intimate relationships are just one form of inequality in the network of unequal gender relations.A possible explanation for these regional differences could be found in the fact that regional differences exist in Croatia in prevailing social identities (Cifrić and Nikodem, 2008).In Dalmatia, people have a pronounced cultural dimension of their social identity (they highly value their customs, traditions, ancestry, and landscape), for residence of the Slavonia region the most pronounced dimension of social identity is national-religious identity.These identities could perpetuate gender inequalities in intimate relations.
The second hypothesis argued that gender socioeconomic inequalities are not equally distributed in rural and urban areas.This hypothesis was confirmed.The results showed higher gender socioeconomic inequalities in favour of men in couples living in rural areas in comparison to the ones living in urban areas of Croatia.This finding is consistent with previous studies that showed that the population of rural areas expresses more traditional values than the urban population, experience more gender discrimination against women in family ( Kamenov and Jugović, 2011), show a lower degree of awareness of gender inequalities and are less sensitive to gender discrimination (Ajduković, 2011).
The third hypothesis argued that gender socioeconomic inequalities are not equally distributed across different types of intimate relationships (marriage vs. domestic partnership), and this was confirmed.The results showed that gender socioeconomic inequalities in favour of men are higher in married couples than for people living in a domestic partnership.This finding is consistent with previous studies that showed that married people generally accept traditional and patriarchal values more than people who live in domestic partnerships (Cifrić, 2013;Galić, 2012).

CONCLUSION
Although there have been some valuable empirical contributions in the recent years with regard to research of different dimensions of gender socioeconomic inequalities in Croatian society, such as gender inequalities in household chores and child domestic work (e.g.Klasnić, 2017, Tomić-Koludrović, Puzek andPetrić, 2019), gender inequality and discrimination in family relations, education, at work and in the labour market, in politics (e.g.Kamenov and Galić, 2011), this study, with its focus on gender socioeconomic inequalities between partners in intimate relationships in Croatian society, deals with an unexplored topic.The results showed significant gender socioeconomic inequalities between intimate partners, mostly in favour of men, particularly in regions with more traditional values, among couples who live in rural areas, and among people who are married.In addition to these new empirical insights into gender socioeconomic inequalities in intimate relationship in Croatian society, this paper introduces a new way of measuring this complex research subject using a survey method -the GSEI index.The test results of three derived hypotheses are in favour of validating the constructed GSEI index on a sample of Croatian women.The main strength of this index is the fact that it consists of only five indicators for which seven variables in the questionnaire are needed, which makes it straightforward and simple to use.Another strength of this index refers to the fact that it can be used on samples of women only, men only, or men and women together, and still have intimate relationship as the unit of analysis.This is possible because survey participants can provide answers on objective SES indicators referring to both themselves and their intimate partners.The only criterion for application of the GSEI index is that participants must be in an intimate relationship, living in the same household with their intimate partner (or, depending on the research problem, that they had such experience prior to conducting the research).
The main limitation of this study is due to the fact that this index has only been validated on a sample of women.More research is needed to evaluate the GSEI index on a sample of men, and on a sample of both sexes together.Moreover, although I consider the fact that the GSEI index consists of only five indicators as an advantage, it can also be seen as its limitation because of the possibility for omission of relevant indicators in different social environments.All of the pros and cons of composite socioeconomic indicators that have been well established and methodologically grounded (for practical review see OECD, 2008: 13-14) can be applied to the GSEI index.
Although it could be conceived that the empirical data presented in this paper can be used to describe socioeconomic inequalities in Croatian society, there are two major limitations for such a venture.The first limitation refers to the nature of our sample: 23% of surveyed women referred to their former partners and relationships that ended on average 12 years ago.Not only does this fact prevent me from considering these results as the exact representation of current social reality, but it could also make the process of index validation questionable.However, this is the case with only less than a quarter of the whole sample and additional analyses showed that there was no statistically significant difference in the GSEI scores between women who referred to their former partners and those who referred to their current partners.Thus, I believe that this limitation had a significant impact neither on the results nor on the process of index validation.The second limitation for generalisation and description of the current state of gender socioeconomic inequality in Croatian society refers to the fact that the data was collected in 2012, which could by now be considered obsolete.However, since gender inequality in Croatia is practically a stable social phenomenon, or more precise, a social phenomenon that shows very slow positive trends -most notably from the Gender Equality Index, which rose from 50.3 in 2005 to only 53.1 in 2015 (European Institute for Gender Equality, 2017a) -this is a strong argument in favour of considering these data relevant and not outdated.
Recommendation for further research and GSEI index application is to consider which are the relevant indicators for measuring gender socioeconomic inequalities in intimate relationships depending on the wider social context and to validate the index on different samples.For example, in countries like Sweden, Iceland, Portugal, and Norway, where the incidence of parental leave-taking is relatively evenly shared between women and men (OECD, 2019), the intensity of parental leave use (duration of paid leave in days or mounts) could also be a relevant indicator of gender socioeconomic inequalities in intimate relationships, but one that applies only to couples with children.Inheritance norms (Goodnow and Lawrence, 2010) and gender trends in atypical and precarious work (European Institute for Gender Equality, 2017b) could also be considered while searching for relevant indicators in a particular social and cultural environment.Also, the potential international application of this new measure could be used to explore the relation between national and regional GSEI scores and other gender inequality measures like the United Nations' Gender Development Index (GDI) or the World Economic Forum's Global Gender Gap Index.

4
The Committee of the Sociology Department for evaluation of research ethics at the Faculty of Humanities and Social Sciences, University of Zagreb.based on age, education, size of place of residence and region.6

001- 1 01
Woman has much greater income or is the only provider 0 Equal incomes or small difference in incomes 1 Man has much greater income or is the only provider Relative employment status -1 Woman is employed, man is other (unemployed, retired or other) Both are employed or neither is employed 1 Man is employed, woman is other (unemployed, takes care of household, retired or other) Woman or a member of her family Mutual or other (rent, someone else, unresolved ownership) Man or a member of his family Car usage Mostly or exclusively woman No car, equal use or each have their own car Mostly or exclusively man

Figure 2 .
Figure 2. Distribution of results on the GSEI index

Figure 3 .
Figure 3. Box-and-whisker plot for the GSEI score by region

Figure 4 .
Figure 4. Box-and-whisker plot for GSEI score by type of place of residence

Table 3 .
Description of answers on original variables (N=601)