Partner choice in Sweden: How distance still matters

Abstract

Spatial homogamy, or the geographical closeness of life partners, has received little attention in recent decades. Theoretically, partners may be found anywhere in the world, as increases in educational participation, affluence, mobility and internet access have reduced the meaning of geographical distance in general. This paper examines whether geography still matters in the Swedish partner market, by examining distances between partners before co-residence over time. Register data are used to track the residential histories (1990–2008) of couples who married or had a child in 1996, 2002 or 2008 (N = 292,652). With the couple as the unit of analysis, the distance between partners before co-residence is explained by geographical, socio-economic and demographic indicators. I find that although the distance between partners has increased over time, it is still the case that half of all partners lived within 9 kilometres of each other before moving in together. Demographic and socio-economic characteristics explain some of the variation in spatial homogamy, but geographical factors, such as previous place of residence, spatial mobility, degree of urbanization and nearness of parents, are crucial. Even in a globalized society, most people still find their partners very close by. The findings are relevant to the family migration literature, where residential mobility at the beginning of co-residence has received little attention, despite long-lasting consequences of partner choice on social ties and people’s socioeconomic careers. The results exemplify the importance of short geographical distances for intimate relationships.

Keywords

Partner choice marriage co-residence spatial homogamy Sweden

Introduction

It is well known from intermarriage studies that people tend to marry partners much like themselves; this is known as homogamy. Traditionally, people tended to marry locals, which can be termed ‘spatial homogamy’. Historical studies on distances between marital partners show patterns of ‘distance decay’: with increasing distance between potential partners, the probability of a partner match decreases (e.g. Bossard, 1932; Coleman and Haskey, 1986).

Nowadays, however, substantial increases in social and spatial mobility, individual affluence, educational participation, travel abroad, and, not least, the use of the internet, may have led to geography no longer being a pivotal factor in partner choice. This fits in with visions of space becoming less important as a result of globalization, and with the ‘death of distance’, suggesting a world in which distance does not matter, where people are decoupled from place, and where societies are based on networks (e.g. Cairncross, 1997; Castells, 2010).

However, there are also indications that geography still matters. Samers (2010) argues that ‘fluctuations in the cost of energy over the last few years suggest that the problem of “overcoming distance” will not disappear’ (Samers, 2010: 19). Space-time constraints still restrict our daily activity patterns (Hägerstrand, 1970; Hanson and Hanson, 1993) and limit our opportunities to meet partners who come from far away. Being in close proximity to someone increases the probability of forming a relationship with him or her. Moreover, the opportunities to meet partners are restricted by geography, and potential partners with specific characteristics are unevenly distributed across space, with highly segregated local marriage markets (Haandrikman, 2010). Recent studies in China (Ouyang et al., 2009), Uruguay (Pullum and Peri, 1999) and the Netherlands (Haandrikman et al., 2008a) find that spatial homogamy is not a phenomenon of the past, which may be explained by the fact that bridging a distance still involves time, energy and cost.

The aim of this paper is to analyse the spatial dimensions of the partner market in contemporary Sweden, using a longitudinal perspective. I use the term ‘spatial homogamy’ for the geographical closeness of future partners, operationalized as the geographical distance between partners before co-residence for partners who both lived in Sweden. In analysing spatial homogamy over time, I contribute to an increased understanding of the role of distance in partner choice in contemporary society, and I shed light on the personal and geographical characteristics that influence the geographical distances between partners, as migration fields and marriage fields mirror the contacts that people have, and give a notion of what surroundings mean to people (Hägerstrand, 1951, 1957). Changes in distances between partners tell us something about social and cultural changes over time: they have been related to the erosion of cultural peculiarities (Bylund, 1977), the integration of smaller communities into larger ones (Knippenberg and De Pater, 1988) and the weakening of regional identification (Häkli and Paasi, 2003). The paper also fills a gap in homogamy research: the spatial dimension of homogamy is a relatively under-researched component, and many of the existing studies on this topic are outdated or were based on historical data, and were often concerned with islands, cities, or small areas.

Most of the family migration literature focuses on couple migration, while few studies focus on the residential relocations around union formation (Brandén and Haandrikman, 2018; Mulder and Wagner, 1993). Knowledge about the changes in and determinants of the geographical distances between partners are relevant to this literature. Studies on residential mobility have shown that most people move short distances, and that moves are triggered by life events, such as starting a co-residential union (see e.g. Coulter et al., 2016). Life course trajectories are not only influenced by individual choices but also by so-called “linked lives,” or the trajectories of persons tied to the individual (Elder et al., 2003). Partners moving in together is the result of bargaining, negotiation and trade-offs between two people (Coulter et al., 2016); migration decisions likely have consequences for daily activities and social relations, and for later life socio-economic outcomes, which often favour the man (Brandén and Haandrikman, 2018; Cooke, 2008; Løken et al., 2013). A spatial perspective on union formation is warranted and may give new insights into migration patterns, processes and theory (Wright and Ellis, 2016).

From a geographical perspective, Sweden is an interesting case as it is a large country, covering an area of almost 450,000 square meters, with a population of only nine million. About 40% of the population resides in one of the three metropolitan areas of Stockholm, Gothenburg and Malmö. Sweden is considered to be one of the world’s most globalized countries (Dreher, 2006), foreign travel is very common (Frändberg, 2009), and the number of marriage migrants has increased greatly in the 1990s and 2000s (Niedomysl et al., 2009). On the other side, Swedes live relatively close to their family members (Kolk, 2017). In addition to this interesting geographical context, the empirical side of the study is another of its strengths: longitudinal register data with excellent geographical features are used for the whole Swedish population for the period 1990–2008.

Preferences, norms and opportunities

In order to understand the mechanisms behind spatial homogamy, the metaphor of the partner market can be used. Contact opportunities determine the supply of partners, whereas partner preferences stimulate the demand side. Sociological studies generally assume that three factors influence partner choice: preferences, social norms, and opportunities (e.g. Kalmijn, 1998). This study applies these concepts in order to understand the geographical dimensions of partner choice, and the role of geography in changing patterns of spatial homogamy. First, abundant studies have shown that people tend to choose partners similar to themselves. The resemblance between partners is usually interpreted to be a result of a preference for similarity, because by sharing values and worldviews each person confirms the other’s views and behaviour, making the other more attractive (e.g. Kalmijn, 1991).

Modernization theory generally states that increased educational levels and economic independence lead to romantic love becoming the most important criterion when choosing a (marriage) partner. A shift from consanguinity to conjugality as a central principle of social relationships started as early as the fifth century (Goody, 1983). According to Shorter (1975), the increasing importance of love as a criterion for partner choice should lead to a change in mentality and the disappearance of (social) homogamy, though most studies have refuted this hypothesis (e.g. Van de Putte, 2003; Van Leeuwen and Maas, 2002). Goode (1963) associated increasing industrialization with changing family systems, in which individualistic values and personal fulfilment become increasingly important in mate selection. These developments imply increasing distances between potential partners. The choice of a partner from the same neighbourhood or village nowadays might partly result from a preference for cultural similarity, given that language, religion and family values are often shared over such distances, as studies for the Netherlands have found (Haandrikman and Hutter, 2012; Van Poppel and Ekamper, 2005).

Second, the social and cultural norms of parents, family, peers, the church, and fellow-villagers, and the sanctions imposed when a partner is found outside the group, influence partner choice and homogamy. In most societies, including in Europe, parents usually had an influence on their children’s marriages, often because of property concerns (Goody, 1996). Parental control was mostly directed towards influencing the meeting process: making sure that children met the right kind of partners. Contemporary research shows that parents still influence their children’s life course decisions (Barber, 2000; Jennings et al., 2012; Liefbroer and Billari, 2010). Such norms may lead to increased spatial homogamy if cultural homogamy is the norm. High endogamy rates among ethnic groups have, for instance, been explained by third party influences (Kalmijn and Van Tubergen, 2010), and, in a qualitative study, the partner’s place of origin and the assumed status of the partner’s family were perceived to be of primary importance in selecting a partner (Haandrikman and Hutter, 2012), suggesting an association between norms and geographical origin. For Sweden, the only relevant studies on recent partner choice focus on migrants, which evidence the importance of parental involvement in partner choice (Çelikaksoy, 2014; Çelikaksoy et al., 2010; Huschek and De Valk, 2012).

Third, the opportunities one has to meet potential partners may be the most obvious mechanism behind spatial homogamy. Meeting a person is normally assumed to be a precondition for falling in love with them, and as we spend most time at or close to home – Ellegård and Vilhelmson (2004) found that Swedes spend two-thirds of their time at home or in the immediate environment of the home – the chance of forming a relationship with someone nearby is increased. Our opportunities to meet partners from far away are limited, because bridging a distance still involves time, energy and cost. Moreover, people tend to meet partners in local marriage markets that are segregated by, for instance, education and religion (Chudnovskaya, 2017; Haandrikman, 2010). Another reason to expect spatial homogamy is that potential partners with preferred characteristics, such as a particular age, education or ethnicity, are unevenly distributed across space. Living in a remote area may significantly reduce the everyday probability of meeting an eligible partner.

In most of human history, marital distances were very short, as most people had hardly any information about potential marriage candidates who lived far away from their ‘information field’ (Marble and Nystuen, 1963; Morrill and Pitts, 1967). Then, throughout the world, different rules concerning endogamy and exogamy determine who is an eligible marriage partner, as described by Goode (1963). In India, the exogamy rule meant that men could not marry women from adjacent villages or from the same lineage, leading to increased marital distances and distinct spatial marriage networks (e.g. Mayfield, 1972; Sivaram et al., 1995), while marriages between two people with the same surname or from the same clan were prohibited in China, which led to a substantial reduction in the availability of marriage partners and to marital distances that were quite long. Processes of modernization have been associated with weakening kinship ties, changes in mate selection practices, an increase in the geographical dispersion of social relations, increasing geographical mobility, and, hence, increasing contact opportunities between people of different backgrounds. As a result, the area from which eligible partners can come has been substantially enlarged (Goode, 1963), and we can suppose that geographical distances between eligible partners will increase with time.

Spatial homogamy in Sweden and beyond

The majority of studies examining spatial homogamy have established that, across the globe, most partners are found close to home, but that, over time, there has been an increase in the number of partners found outside the parish, city or other administrative area (Bossard, 1932; Clegg et al., 1998; Coleman, 1979; Coleman and Haskey, 1986; Fornasin, 2011; Gueresi et al., 2001; Kok and Mandemakers, 2008; Maas and Zijdeman, 2010; Van de Putte, 2003; Van Poppel and Ekamper, 2005). The studies were generally based on very small samples, were conducted in small areas and were mostly cross-sectional.

Published Swedish research on marriages fields is dated and rare (Bylund, 1977), and most studies are case studies of villages or parishes. Kvillner (1969) showed that marriage fields had a radius of about 10 kilometres in previous centuries, and he associated this with limited opportunities to meet partners from far away. Mobility was highly restricted for non-nobility. Beckman (1959, 1960) found substantial regional differences in ‘inbreeding’ in northern Sweden in the period 1860–1920, linking patterns to community size, population density and traditional marriage practices. Bylund (1977) found that mobility within Sweden hardly changed and that migration distances were relatively short in the period 1861–1960. Following Shorter’s (1975) propositions, Van Leeuwen and Maas (2002) found no proof of a decline in social homogamy in a northern industrial area in the 19th century. Using historical ethnographic studies, they describe courtship practices regulated by rural fraternities, consisting of communal nightly visits, which was also found in Japan (Goode, 1963). These customs slowly disappeared from society, though parental control remained quite strong.

Developments in Sweden can be placed in a wider context of modernization processes leading to a weakening of kinship ties, an increased number of individuals choosing their partners freely and from a broader pool of potential partners, and an increased prevalence of separate living arrangements (Goode, 1963). Current average distances between partners in Sweden may be expected to be considerable (hypothesis 1) and to have changed over time (hypothesis 2). There has been a substantial increase in the number of people completing higher education in the last few decades (Kupiszewski et al., 2001), and the more highly educated tend to find partners further away (e.g. Haandrikman et al., 2008b). Swedes are known for their frequent visits and stays abroad, for work, study or leisure (Frändberg, 2009). The use of the internet for finding a partner is also on the rise, and Sweden is very well connected to the internet (Ellegård and Vilhelmson, 2004). Increasing numbers of Swedes marry foreign-born partners (Haandrikman, 2014), both migrants living in Sweden and marriage migrants (Niedomysl et al., 2009).

Factors influencing spatial homogamy

From studies on spatial homogamy, three main groups of factors that influence the level of spatial homogamy can be identified: socio-economic status, demographic factors and geographical factors. These are discussed below, which lead to the hypotheses as shown in Table 1.

Table 1.

Hypotheses.

1	In contemporary Sweden, partner choice is still characterized by distance decay.
2	The distance between partners before co-residence has increased over the last two decades.
3	Higher socio-economic status is related to increased distances between partners before co-residence.
4	With increasing age, distances between partners decline.
5	Spatial homogamy is stronger for single parents and those living with their parents prior to union formation.
6	The higher the level of urbanization, the lower the mean distances between partners before co-residence.
7	Individuals who are relatively mobile will have longer distances to their partner prior to union formation.
8	Spatial homogamy is strongest for people living close to family members.
9	There are no large differences in spatial homogamy across different areas.

Socio-economic status is the most studied dimension in spatial homogamy research, especially in historical studies. Typically, the higher social classes find partners at longer distances (see for instance Clegg et al., 1998; Haandrikman et al., 2008b; Van Poppel and Ekamper, 2005). This has been explained by a combination of factors: those in higher classes might prefer partners with a similar high status; there might be strong norms to marry endogamously; and, especially in the past, the higher social classes have had much more opportunity to search for partners in a wider geographical area. At present, the better educated and those with higher incomes might still have broader geographical networks and more opportunities to travel (the higher educated tend to be more likely to be long-distance movers), increasing the chance of finding a partner from further away (hypothesis 3).

Second, spatial homogamy is affected by demographic factors. Generally, with increasing age, distances between partners decrease; this is often associated with diminishing affluence for older people (Clegg et al., 1998; Coleman and Haskey, 1986; Haandrikman et al., 2008a). Therefore, we expect that spatial homogamy decreases with increasing age (hypothesis 4). In addition, marital distances have been found to be shorter for those who are single parents or who live with their parents before they move in with their partner (Haandrikman et al., 2008a), which is in line with family ties deterring migration (Mulder, 2007) (hypothesis 5).

Third, spatial homogamy is influenced by geographical factors. First, the opportunities to meet partners may be reduced because of geographical barriers. In addition, distances between partners may vary by place of residence because of varying degrees of population density (hypothesis 6). In urban areas, spatial homogamy has been found to be stronger than in peripheral areas (Haandrikman and Van Wissen, 2011; Haandrikman et al., 2008a). Urban areas have higher densities of people, schools, jobs, and places of entertainment, and thus offer plenty of eligible partners. Large parts of Sweden are uninhabited, and this might influence spatial homogamy levels in different parts of the country. Second, geographical mobility changes one’s dating pool, and those who travel more have better chances of meeting more people (Smaldino and Schank, 2012). In Sweden, commuting has increased over time, leading to the enlargement of functional regions (Scholten et al., 2012). This leads to hypothesis 7: increased geographical mobility is associated with an increasing distance at which partners are found. Third, regional variations in cultural factors, such as religion, mentality, and dialect, may lead to regional differences in spatial homogamy. Sweden is generally seen as one of the most equal countries in the world, with some arguing that class boundaries do not exist (Ehn et al., 1993; Pred, 1997). There is a traditional division between Norrland (north), Svealand (middle) and Götaland (south), based on differences in agriculture and, for instance, construction types; though today some may argue that there are hardly any cultural differences left, other divisions are based on ethnological and linguistic material (Edlund, 1994; Öberg and Springfeldt, 1991). Sweden’s bible belt is located around Jönköping, and, together with the south of Sweden, this area is often considered to be more traditional. This is for instance revealed when mapping total fertility rates across regions, where the highest rates are found in the south and especially around Jönköping (Statistics Sweden, 2018), revealing more traditional views on marriage (Sandström, 2011). Linguistically, we might expect more endogamous partner choice in areas with minority dialects or languages, as was found for the Netherlands (Haandrikman and Van Wissen, 2011). Finnish and Sami are practiced in some Swedish areas, but dialectal and other cultural differences are fairly small. I therefore do not expect to find large differences in the distances between partners in different areas (hypothesis 9). Fourth, the distance to family members might influence the distance between partners before living together (hypothesis 8), as studies have shown that Swedes live relatively close to their family members (Kolk, 2017) and that people tend to be unwilling to move far away from family members (Mulder, 2007). Swedish men tend to live closer to their parents than Swedish women do (Brandén and Haandrikman, 2018; Malmberg and Pettersson, 2007), and couples with children tend to live closer to the paternal than to the maternal grandparents (Løken et al., 2013).

Data and method

I used register data from the PLACE database, which is administered by Statistics Sweden and managed at Uppsala University, and includes all registered individuals in Sweden for the period 1990–2008. PLACE contains a wide range of variables, including demographic and socio-economic features together with information on residential history. Moreover, it has excellent geographical attributes, namely location coordinates for each 100 by 100 meter square, implying that analyses can be done independently of administrative borders.

In Swedish registers, individuals are registered at properties (fastigheter in Swedish) rather than in households. I define a couple as two individuals living in the same property who are either married to each other or have a common child. Individuals who cohabit without having a common child cannot be captured. Place of residence is approximated by the geographical coordinates of the 100 by 100 meter square in which the property is registered.

As I am interested in distances between partners before they move in together, I chose to focus on new couples and the distances between their previous addresses. Given the definition issues discussed above, I limited the study to couples who got married, or who had a child together as cohabitants. The procedure for selecting couples was based on a transition in family position. I selected those persons who were linked to each other by living in the same property and by being married or having a child in common in a certain year. Then I only selected those who were not linked to that partner through marriage or a common child in the previous year. For each couple, I then went back year by year, until 1990 at the earliest, to find the time when the two partners did not live in the same square. In the year before co-residence in the same square, I measured the Euclidean distance between the partners’ addresses. If two individuals lived in the same square before they became a couple, the timing of co-residence is mistakenly measured prior to the start of actual co-residence. However, the determination of couples is not affected. If two individuals moved in together and transitioned to marriage or having a child the same year, the distance measured is the distance between the two individuals in the year prior to the move. In those cases, the distance measured will be overestimating the distance between these partners at the moment of meeting. Couples for which one or both partners lived abroad in the year before cohabitation were excluded from the analysis, as, in order to measure the distance between partners, both addresses needed to be in Sweden. Another limitation is that I only measured transition to marriage or having a child with a cohabitant; I did not know if a person had been married before, or whether a person already had children with another partner. In the analysis, I do control for age on first living together and previous household position. A last limitation is that I only had access to annual data on place of residence, so that the exact order of the events of interest is unknown in the case of several events occurring within the same year (see above).

I selected three years for data analysis, 1996, 2002 and 2008, resulting in 313,200 individuals or 156,600 new couples, who either married (42%) or had a child while cohabiting (58%) in the selection year.¹ The numbers of individuals included for each selection year, by selection mechanism, is displayed in Table 2. The proportion of partners not captured because they started living together before 1990 ranges from 11% for those selected in 1996, to 6% for 2002 and 4% for 2008 couples. The total number of individuals analysed was 292,652.

Table 2.

Data and couple selection characteristics.

Selection year	Year in which persons started living together	Selected because of marriage	Selected because of cohabiting with common child	Total included N	Excluded cases
1996	1990–1996	31,612	46,646	78,258	9618
2002	1990–2002	42,364	58,588	100,952	6384
2008	1990–2008	49,410	64,032	113,442	4546
Total	1990–2008	123,386	169,266	292,652	20,548

Note: A total of 20,548 cases out of 313,200 or 6.5% of all individuals were excluded because they were co-residing with their partner for the whole period until the selection year.

The method centres on couples as units of analysis, thereby recognizing the mutual nature of partner choice. Each record includes (1) variables to describe the man; (2) variables to describe the woman; and (3) variables describing the couple, as described below. The modelling strategy consisted of multiple regressions explaining the distance between partners before co-residence as the dependent variable. Two regression models are specified, where the one for men uses predictors of types (1) and (3), while the one for women uses predictors of types (2) and (3). In order to separate the geographical effects from the socio-economic and demographic indicators, stepwise regressions are conducted. The choice of the variables included in the analysis was inspired by previous findings and the rich geographical variables available in the data.

To test how spatial homogamy has developed over time, the year in which couples started to co-reside was included. In terms of socio-economic class, level of education, income and whether the person was in education were included. Level of education was measured in June of the year before a person started living with a partner, and was categorized as primary school (nine years of schooling), upper secondary school (an additional two or three years), and tertiary education. Income was also measured in the year before co-residence, and includes the total income from employment and self-employment. The variable in the analysis represents relative income: very low or missing income (very low meaning up to 133,800 SEK/14,150 euro per year), medium income (133,800–214,600 SEK/14,150–22,700 euro per year) and high income (more than 214,600 SEK/22,700 euro per year). An unknown number of young people in Sweden, especially students, are believed not to be registered at the right address just after leaving home (Statistics Sweden, 2008). If these people are wrongly registered at their parents’ address, for instance, this could bias the distances measured in this paper. To partly control for this, an indicator for whether a person received student allowances in the year the person started to co-reside, was included.

The demographic indicators incorporated were age, household position, and marital status. Age was measured in the year the person started living with a partner. Household position in the year before moving in with a partner was categorized as living with parents, living as a single parent, or living in another household position, including living alone. Marital status was measured in the year the couple started living together, distinguishing those who immediately married from those who did not.

To capture possible geographical influences on spatial homogamy, five indicators were included. First, country of birth was categorized as born in Sweden, born elsewhere in Europe or born outside Europe. Second, the area in which people lived before co-residence was included, as this might reveal regional cultural differences in distances to partners. Area of residence was approximated by local labour markets, a Statistics Sweden geographical unit that uses annual commuting patterns between municipalities (Statistics Sweden, 2010), that is also useful to display regional variation in cultural factors. For this study, I used the classification from the year 2000 (N = 90) and applied it to the municipality of residence before co-residence for all years of observation. Third, degree of urbanization of the property in the year before co-residence was based on a unique measurement of the radius of the circle centred on an individual in which 20,000 people live, calculated for each 100 by 100 meter square in the whole country using the EquiPop software (Östh, 2014; Östh et al., 2014). I chose to use this definition rather than other measures such as the more widely used OECD typology (OECD, 2010), that define degree of urbanization for each municipality, based on a combination of population density and the presence of urban centres, which implies that, apart from Stockholm, Gothenburg and Malmö, the rest of Sweden is rural. The new urbanization index better captures the heterogeneous nature of rural areas. Five categories were used in the analysis: urban areas, areas adjacent to urban areas, accessible countryside, peripheral countryside and remote countryside, based on the distance in which 20,000 people are captured – respectively less than 11 km, 11–23 km, 23–48 km, 48–93 km and more than 93 km (Hedberg and Haandrikman, 2014).

Fourth, a mobility variable was included to measure the extent to which individuals have access to a larger partner market. The variable measures whether a person was living in a different county (there are 21 Swedish counties, or län in Swedish) five years before the year that the person started to co-reside, to indicate a certain degree of residential mobility. Finally, the distance to a person’s mother’s address in the year of co-residence was included.² If the address of the mother was not available, the father’s address was taken instead. The Euclidean distance between the geographical coordinates of the addresses of the parent and the child was then calculated.

Results

The analysis of register data shows that most couples live very close to each other before they start living together, while only a few live a long distance apart. Table 3 shows mean and median distances between partners before co-residence for the period under study. The mean distance between partners before co-residence was 57 kilometres, while half of all partners lived within 9.2 kilometres of each other. About 14% of all couples lived within 1 kilometre of each other before co-residing, and 8% of these lived in the next 100 by 100 meter square.³ This confirms hypothesis 1 that, in contemporary Sweden, partner choice is characterized by substantial distance decay. The longest distance between two partners was 1417 kilometres – a couple of whom one partner lived in Kiruna in northern Lapland and the other lived in southern Malmö before they moved in together. The level of spatial homogamy varies by selection mechanism: those who were included in the study because of marriage lived further apart than those included because they had a child while cohabiting.

Table 3.

Indicators of spatial homogamy for the period 1990–2008.

	Included because of marriage (km)	Included because of having a child with cohabitant (km)	All (km)
Mean distance between partners before co-residence	65.4	51.0	57.0
Median distance between partners before co-residence	10.6	8.2	9.2

Figure 1 shows that the distance between partners before co-residence has increased over time, thereby confirming hypothesis 2, although the increase has not been linear and many fluctuations have taken place over the last two decades. It is also evident that, over the whole time period, distances between partners before living together for those who were selected because they married have been consistently higher than distances for those who had a child together while cohabiting.

Figure 1.

Median and mean distances between partners before co-residence, by year of co-residence and selection type.

Unfortunately there are no recent studies for Sweden to use for comparison, but the median distance of 9 kilometres before co-residence could be compared to the median distance of 6 kilometres that was found for couples who started to live together in 2004 in the Netherlands (Haandrikman et al., 2008a). Given the fact that Sweden is 10 times larger than the Netherlands, the median distance between partners found here could be perceived as relatively short, although the mean distance, influenced by the small number of very long distances, is much longer than the 23 kilometres found for the Dutch study, but is not unexpected given the much larger surface area of Sweden.

The results of the multivariate regression analyses on the distance between partners before co-residence, with the log distance between partners as the dependent variable, are shown in Table 4. Model 1 mainly contains socio-economic and demographic indicators to test hypotheses 2–5; in Model 2, geographical variables are added to the explanation of spatial homogamy in Sweden, testing hypotheses 6–9. The model fit statistics show that the explanatory power of the model greatly increases when geographical attributes are added to the model, especially for women, but that socio-economic and demographic characteristics of partners, together with changes over time, are also significant in explaining distances between partners before co-residence in Sweden.

Table 4.

Regression results of the log distance between partners before co-residence, men and women.

	Men				Women
	Model 1		Model 2^a		Model 1		Model 2^a
Variable	Β	t	Β	t	Β	t	Β	t
Constant		506.2		364.2		491.8		366.1
Selection because of having a child while cohabiting vs. marriage	−0.08***	−25.4	−0.06***	−21.9	−0.07***	−22.7	−0.05***	−17.8
Year of co-residence
1996–2000 vs. 1990–1995	−0.01*	−2.4	−0.01**	−3.1	0.00	−1.1	0.00	−0.9
2001–2004 vs. 1990–1995	0.02***	5.7	0.02***	4.6	0.02***	7.2	0.02***	6.8
2005–2008 vs. 1990–1995	0.04***	12.6	0.05***	13.3	0.05***	13.3	0.05***	14.6
Educational level before co-residing
Primary vs. upper secondary	−0.02***	−7.4	−0.02***	−7.3	−0.02***	−6.8	−0.02***	−5.7
Tertiary vs. upper secondary	0.04***	13.8	0.01*	2.0	0.04***	13.1	0.00	−0.7
Educational level missing vs. upper secondary	0.00	−0.8	0.00	0.4	0.00	−1.2	0.00	−0.1
Income before co-residing
Low or missing income vs. medium income	0.02***	5.0	0.01	1.6	0.01^†	1.9	0.00	−0.9
High income vs. medium income	0.03***	7.2	0.02***	5.5	0.00	0.8	−0.01^†	−1.7
Age at co-residence
<23 vs. 24–28	−0.03***	−8.5	−0.02***	−6.0	−0.05***	−15.1	−0.04***	−10.6
29–34 vs. 24–28	0.00	−1.0	−0.01^†	−1.8	0.01*	2.1	0.01^†	1.9
35–45 vs. 24–28	−0.02***	−4.8	−0.01***	−4.6	0.02***	5.3	0.02***	6.3
46–55 vs. 24–28	0.00	0.8	0.00	0.3	0.02***	6.4	0.02***	8.2
56–65 vs. 24–28	0.01***	4.1	0.01***	4.2	0.02***	5.6	0.02***	7.6
>65 vs. 24–28	0.01***	4.2	0.01***	4.1	0.01***	4.4	0.02***	5.8
Household position before co-residence
Living with parents vs. living alone or other household position	0.10***	33.5	0.11***	39.7	0.11***	33.8	0.14***	45.7
Single parent vs. living alone or other household position	0.00	−0.6	0.01*	2.0	−0.06***	−20.4	−0.03***	−10.8
Receiving study allowances vs. not receiving allowances	0.05***	15.3	0.02***	7.2	0.04***	13.1	0.02***	5.8
Married in year of living together vs. unmarried	−0.03***	−9.4	−0.02***	−7.7	−0.02***	−7.5	−0.02***	−5.7
Country of birth
Europe (outside Sweden) vs. Sweden			−0.02***	−6.6			−0.01***	−4.9
Outside Europe vs. Sweden			−0.02***	−5.7			−0.01***	−3.6
Degree of urbanization before co-residence
Urban area vs. area adjacent to urban areas			−0.09***	−29.9			−0.05***	−16.2
Accessible countryside vs. area adjacent to urban area			0.01*	2.0			0.00	−0.5
Peripheral countryside vs. area adjacent to urban area			0.00	0.6			0.00	1.1
Remote countryside vs. area adjacent to urban area			0.01*	2.4			0.00	0.7
Residential mobility: Living in another county 5 years ago vs. same county			0.12***	43.5			0.13***	46.4
Distance to parents in year of co-residence
<5 km vs. 25–100 km			−0.19***	−48.6			−0.23***	−63.4
5–25 km vs. 25–100 km			−0.09***	−23.4			−0.11***	−30.7
>100 km vs. 25–100 km			0.04***	11.9			0.06***	16.6
no parents alive or in Sweden vs. 25–100 km			−0.07***	−18.2			−0.08***	−22.0
Adjusted R²	0.023***		0.094***		0.028***	0.119***

*p < .05, **p < .01, ***p < .001, ^†< .1.

^aModel 2 also controls for labour market areas. The coefficients are displayed in Figure 2.

The first model shows that there has been a significant increase over time in distances between partners before co-residence, although this is only true for the first decade of the 21st century and not for the 1990s, thereby partly confirming hypothesis 2 that spatial homogamy has decreased over time. The increase in distance has been greatest for the most recent period of 2005–2008. Those included in the study because they had a common child while cohabiting had lived closer to each other before co-residence than those selected because of marriage. In Sweden, virtually all couples cohabit before marriage, and most people eventually get married (Andersson and Philipov, 2002), with a marriage revival being observed in the period under study (Ohlsson-Wijk, 2011). In a survey among young Swedes about attitudes towards marriage, reported advantages of marriage compared to unmarried cohabitation included sending a signal to others that one is serious about their relationship, that it is romantic to get married, and that marriage is more secure. More positive marriage attitudes were found among older individuals, those with higher education, those living in cities, those with partners, and those who were more religious and family-oriented (Bernhardt, 2001). Those who choose to cohabit and to remain unmarried may start living together more easily and instinctively with a partner who they meet close by, while those who (eventually) marry might make a more conscious decision, given their labour market attachment and family ties to given locations (following Baker and Jacobsen, 2006; Fischer and Malmberg, 2001).

Based on previous studies, I expected that those with higher socio-economic status would find life partners at longer distances (hypothesis 3). Without geographical features in the model, this is partly the case. Less well educated men and women find partners at significantly shorter distances, while the better educated find partners at longer distances. However, when geographical indicators are added to the model, I find no significant effects on spatial homogamy for better educated women and only a small effect for men, though the effects for the less well educated as compared to those with average education remain.

Taking into account geography (model 2), men with higher incomes find partners at longer distances, while women with higher incomes meet partners closer to home, though this effect is small; this is in contrast to previous studies.

The demographic correlates show an irregular pattern. Both men and women who start co-residence before age 24 find partners at shorter distances. Men who start living with a partner between the ages of 29 and 45 live closer to their partners than those who start to co-reside between the ages of 24 and 28, while late cohabiters find partners at significantly longer distances. For women, co-residence from age 29 onwards is related to finding a partner at a significantly longer distance. These findings are contrary to hypothesis 4 that the distance between partners decreases with increasing age. One possible explanation could be that the studies that influenced this hypothesis were based on bivariate analyses. Another could be that, with age, women in particular place more value on socio-economic status in partner choice (Schwarz and Hassebrauck, 2012), which may imply longer distances to find such a partner. Third, women more often move longer distances than men, because men (who are generally a few years older than their partners) are more established in the labour and housing market than women of a similar age (Bielby and Bielby, 1992; Mulder and Malmberg, 2012), and because of a gender culture that supports the systematic adaptation of women to their partners’ careers (Brandén and Haandrikman, 2018). Finally, women tend to move away from rural areas much more than men (Nilsson and Strandh, 1999), which has been linked to men being more interested in activities typical for and tied to rural areas, and women more inclined to search for education and employment opportunities that are located further away (Dahlström, 1996).

Single mothers find new partners at shorter distances than those living alone before moving in with a partner, as expected (hypothesis 5), while single fathers are slightly more inclined to find partners further away. This would seem to be related to the fact that mothers live closer to their children after separation (Stjernström and Strömgren, 2012). Living in the parental home before moving in with a partner is an important explanatory factor for spatial homogamy: living with one’s parents is associated with significantly longer distances to partners. This is in contrast to hypothesis 5 and the expectations and findings from the Netherlands, where, in a bivariate analysis, those living with their parents were more likely to find partners at shorter distances (Haandrikman et al., 2008a). Given the rural nature of the country, many Swedish youngsters living with their parents will obviously have to travel much longer distances to find partners than their peers in densely populated areas.

Unsurprisingly, I find that those enrolled in higher education find partners at longer distances than those who are not enrolled. These students might meet their partners in the towns where they study; in these places, many people of different origins come together and form a big dating pool. Lastly, those getting married in the same year as they started living together are more likely to have found their partners at shorter distances than those who do not get married in the same year. Further statistical inspection reveals that these partners are more often migrants, were older when they started living together, more often lived alone instead of living with their parents before union formation, and more often lived in Stockholm. Spatial homogamy might be stronger for these individuals, who may be more inclined to search for commitment, prompted by cultural background or for other reasons that are not revealed by these data.

Residential mobility, degree of urbanization and the nearness of parents are the most important explanatory geographical factors in explaining spatial homogamy, and lead to a substantial increase in the explanatory power of the model. Corroborating previous studies and confirming hypothesis 6, spatial homogamy is strongest in urban areas, while men in the remote countryside find partners at longer distances. The effect of the mobility variable is as expected (hypothesis 7): the more mobile individuals are, the longer the distances at which they find a partner. The analysis, furthermore, shows a strong association between geographical closeness to parents and spatial homogamy. The closer the parents, the shorter the distance at which a partner is found (confirming hypothesis 8). Living close to one’s parents might indicate that there is a socio-cultural norm of finding a partner from the same region. On the other hand, those who do not have their parents nearby, because their parents are deceased or live abroad, also tend to find partners nearby. The same applies to partners born outside Sweden. These people might have geographically restricted networks within Sweden, and as a result also choose partners from a geographically limited pool of eligible partners.

Figure 2 shows the standardized betas (Β) of the regression analysis for the different labour market areas of Sweden on a map. The most remarkable finding is that both men and women in northern areas find partners at significantly longer distances than men and women in the Stockholm region, even when taking socio-economic, demographic and other geographical factors, such as population density, into account. Spatial homogamy is strongest in the area west of Stockholm and in Skåne in the south. We may conclude that there is a north-south divide, with longer distances between partners in some areas in the north and shorter distances in some areas in the south. The exceptions in the north are that there are relatively short distances between partners for men in Haparanda near the Finnish border and for both men and women in the most northern area, Kiruna. This might be related to linguistic preferences, or to concentrations of people in small places surrounded by largely uninhabited areas. Southern Sweden and the area known as the bible belt are generally seen to be slightly more traditional in terms of gender relations and demographic behaviour, and we also see that spatial homogamy is strongest in those areas, thereby partly refuting hypothesis 9.

Figure 2.

Regression results of the log distance between partners before co-residence for labor market areas (classification from the year 2000). The map shows standardized betas, corrected for all other variables as listed in Table 4. White areas are those for which the standardized betas were non-significant.

Conclusions and discussion

This study has shown that geography still matters in present-day partner choice in Sweden. Likes like like, also in terms of geography. Although distances between partners before co-residence have increased over time, the level of spatial homogamy is still relatively high. Partners are, in general, still found close to home, most probably because most people still study, work and spend most of their time close to their home base. We can certainly not speak of the death of distance so far as partner choice is concerned. On the other hand, recent changes in mobility and technology and, in particular, the enormous increase in internet use are likely to have led to slight increases in distances between partners, but only in the last couple of years. The effects of space-transcending technologies have not yet led to substantial changes in the distances people cross in their everyday life (Thulin and Vilhelmson, 2010) or to substantial changes in the distance between partners.

Geographical attributes of individuals, especially the degree to which individuals are residentially mobile, the degree of urbanization of their surroundings, and the nearness of parents, explain the major differences in distances between partners in new couples. Distances to partners are shortest in urban areas, for less mobile partners, for people whose parents either live nearby or are not in Sweden, and for those born abroad. The longest distances are found for those who are geographically mobile, for men who live in the remote countryside and for those living far away from their parents. These findings suggest that the place where one meets one’s future partner is largely determined by geographical opportunities. In addition, shorter distances between partners, such as those found in the south and in the bible belt, might reflect a cultural preference and/or cultural norms to choose a partner from the same region.

In addition, socio-economic and demographic characteristics play a role in the varying levels of spatial homogamy. Spatial homogamy is strongest for single mothers, young cohabiters, the less well educated, women with higher incomes, and migrants. Distances to partners are longest for people who start co-residing at later ages, those moving directly from the parental home, the better educated, students and men with higher incomes.

Presently, most migration scholars are concerned with international migration, which is usually concerned with movement between very distant places (Wright and Ellis, 2016). However, the numbers of people involved in internal migration are much larger than the volumes of international migration. This study is one of a few that investigated mobility at the beginning of co-residence and reveals that half of all people in Sweden meet their partner within 9 km distance. The role of geography and of local and social ties, such as the nearness of family members, in life course trajectories is paramount. We know that if people move, the majority moves short distances. This paper adds that also at union formation, distances are short. A consequence of moving short distances when moving in with a partner is that local ties are strengthened (Løken et al., 2013). Such ties may then widen to include local ties to friends, employment, housing and other ties, thereby further decreasing the likelihood of moving (Fischer and Malmberg, 2001). On an individual level, women more often move in with their male partner, and more often move longer distances (Brandén and Haandrikman, 2018). Women’s careers more often suffer than men’s after couple migration (Bielby and Bielby, 1992; Boyle et al., 1999). On the other hand, the choice to move in with a partner living close by might prevent moving to places with more employment opportunities, and thereby hindering progress in individual employment careers (Løken et al., 2013). How couples negotiate the decision where to live throughout the life course is increasingly being studied (e.g. Brandén and Haandrikman, 2018; Stockdale, 2017), but deserves more attention within family migration, not the least as such decisions have consequences for later-life socio-economic outcomes, networks and perhaps, well-being (Nowok et al., 2013).

Although the choice of a partner is a very personal decision, the kinds of partner choices that are made reflect the intensity of ties between and within groups in society. Previous studies on geographical exogamy pointed to the relevance of such studies for gene flows (e.g. Relethford and Mielke, 1994), whereas today, patterns of spatial homogamy may reveal convenience, cultural factors and local activity patterns. Changing patterns of partner choice signal social change, with increased connections between different regions, the integration of small communities into larger ones (Knippenberg and De Pater, 1988), weaker regional identification, and increasing social openness and integration (Smits et al., 1998). Increasing distances between partners may suggest that localities are increasingly connected, and that cultural and social distances between groups are declining. The results of mapping the geographical dimensions of partner choice reveal a quite local pattern of contacts, implying that the immediate surroundings of individuals are still crucial in shaping social relationships. In places with relatively short distances between partners before co-residence, communities might be closer-knit as locals tend to marry locals, whereas in many urban areas, patterns of spatial homogamy might be based on people meeting locally, but born in places far apart. In any case, neighbourhood effects on partner choice seem crucial, even today.

A study that links actual meeting places to geographical distances between partners would add much to the explanations found in this paper. It would also be interesting to examine whether spatial homogamy is related to relationship stability, and how spatial homogamy relates to other forms of homogamy. Further research would, ideally, study all couples who move in together, including those who separate later or who do not have children. Another avenue to be continued pertains to the gender differences in residential mobility related to union formation. Swedish women leave the parental home earlier, study more often than men, and move more often, and in the Nordic countries, youngsters tend to live on their own when they leave the parental home instead of forming a union (Mulder, 2009; Mulder and Wagner, 1993). Those who do leave the parental home to form a union find partners at longer distances, and these couples are mostly less well educated and relatively older.

Footnotes

Acknowledgements

I thank Maria Brandén, Bo Malmberg, Gunnar Andersson and Marie Evertsson for insightful comments on the paper.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: I acknowledge the financial support from the Swedish Research Council (Vetenskapsrådet) via the Swedish Initiative for Research on Microdata in the Social and Medical Sciences (SIMSAM): Stockholm University SIMSAM Node for Demographic Research, grant registration number 340–2013-5164.

Notes

References

Andersson

Philipov

(2002) Life-table representations of family dynamics in Sweden, Hungary, and 14 other FFS countries: A project of descriptions of demographic behavior. Demographic Research 7(4): 67–144.

Baker

Jacobsen

(2006) A human capital-based theory of postmarital residence rules. The Journal of Law, Economics, and Organization 23(1): 208–241.

Barber

(2000) Intergenerational influences on the entry into parenthood: Mothers’ preferences for family and nonfamily behavior. Social Forces 79(1): 319–348.

Beckman

(1959) The frequency of regional intramarriage in North Sweden. Acta Genetica Et Statistica Medica 9: 9–17.

Beckman

(1960) Breeding patterns of a north Swedish Parish. Hereditas 47: 72–80.

Bernhardt

(2001) Att gifta sig – Eller bara bo ihop? VälfärdsBulletinen 4: 4–5.

Bielby

(1992) I will follow him: Family ties, gender-roles beliefs and reluctance to relocate for a better job. The American Journal of Sociology 97(5): 1241–1267.

Boyle PJ, Cooke T, Halfacree KH and Smith D (1999) Gender inequality in employment status following family migration in GB and the US: The effect of relative occupational status. International Journal of Sociology and Social Policy 19(9/10/11): 109–143.

Bossard

(1932) Residential propinquity as a factor in marriage selection. American Journal of Sociology 38: 219–224.

10.

Brandén

Haandrikman

(2018) Who moves to whom? Gender differences in the distance moved to a shared residence. European Journal of Population, pp. 1–24. doi: 10.1007/s10680-018-9490-4.

11.

Bylund

(1977) Kulturregioner och giftermålsfält. Preliminära funderingar. In: Björklund J, Gaunitz S and Lundberg O (eds) Samhällsforskning kring Historiska Problem. Uppsatser tillägnade Professor Gustaf Utterström. Umeå: Studies in Economic History nr 2, pp.56–66.

12.

Cairncross

(1997) The Death of Distance. How the Communications Revolution Will Change Our Lives. Boston, MA: Harvard Business School Press.

13.

Castells

(2010) The Rise of the Network Society. Chichester: Wiley-Blackwell.

14.

Çelikaksoy

(2014) Parental background and union formation behavior of native born individuals in Sweden with a foreign background. Societies 4(3): 351–362.

15.

Çelikaksoy

Nekby

Rashid

(2010) Assortative mating by ethnic background and education among individuals with an immigrant background in Sweden. Zeitschrift För Familienforschung 22: 65–88.

16.

Chudnovskaya

(2017) Higher Education and Family Formation: A Story of Swedish Educational Expansion PhD Dissertation, Stockholm University, Stockholm.

17.

Clegg

Ringrose

Cross

(1998) Some factors affecting marital distances in the Outer Hebrides. Journal of Biosocial Science 30: 43–62.

18.

Coleman

(1979) A study of the spatial aspects of partner choice from a human biological viewpoint. Man 14: 414–435.

19.

Coleman

Haskey

(1986) Marital distance and its geographical orientation in England and Wales, 1979. Transactions (Institute of British Geographers: 1965) 11: 337–355.

20.

Cooke

(2008) Migration in a family way. Population, Space and Place 14(4): 255–265.

21.

Coulter

Van Ham

Findlay

(2016) Re-thinking residential mobility: Linking lives through time and space. Progress in Human Geography 40(3): 352–374.

22.

Dahlström

(1996) Young women in a male periphery – Experiences from the Scandinavian North. Journal of Rural Studies 12(3): 259–271.

23.

Dreher

(2006) Does globalization affect growth? Evidence from a new index of globalization. Applied Economics 38(10): 1091–1110.

24.

Edlund

L-E

(1994) Kulturgränser – Myt eller verklighet? Umeå: Diabas.

25.

Ehn

Frykman

Löfgren

(1993) Försvenskningen Av Sverige. Stockholm: Natur och Kultur.

26.

Elder

Johnson

Crosnoe

(2003) The emergence and development of life course theory. In: Mortimer

Shanahan

(eds) Handbook of the Life Course. New York: Kluwer Academic/Plenum Publishers, pp.3–19.

27.

Ellegård

Vilhelmson

(2004) Home as a pocket of local order: Everyday activities and the friction of distance. Geografiska Annaler 86B: 281–296.

28.

Fischer

Malmberg

(2001) Settled people don’t move: On life course and (im)mobility in Sweden. International Journal of Population Geography 7: 357–371.

29.

Fornasin

(2011) Determinants of territorial exogamy in Friuli (North-East Italy) in the second half of the nineteenth century. Journal of Biosocial Science 43: 453–467.

30.

Frändberg

(2009) How normal is travelling abroad? Differences in transnational mobility between groups of young Swedes. Environment and Planning A 41: 649–667.

31.

Goode

(1963) World Revolution and Family Patterns. London: The Free Press of Glencoe.

32.

Goody

(1996) Family systems in Europe and Asia: Are there different sets of rules? Population and Development Review 22(1): 1–20.

33.

Goody

(1983) The Development of the Family and Marriage in Europe. Cambridge, UK: Cambridge University Press.

34.

Gueresi

Pettener

Veronesi

(2001) Marriage behaviour in the Alpine Non Valley from 1825 to 1923. Annals of Human Biology 28(2): 157–171.

35.

Haandrikman

(2010) The Geographical Dimensions of Partner Choice. Amsterdam: Rozenberg.

36.

Haandrikman

(2014) Bi-national marriages in Sweden: Is there an EU effect? Population, Space and Place 20 (2): 177–199.

37.

Haandrikman

Harmsen

Van Wissen

et al . (2008a) Geography matters: Patterns of spatial homogamy in the Netherlands. Population, Space and Place 14: 387–405.

38.

Haandrikman

Harmsen

Van et al . (2008b) De geografische dimensie van partnerkeuze. Bevolkingstrends 56: 19–28.

39.

Haandrikman

Hutter

(2012) ‘That’s a different kind of person’ – Spatial connotations and partner choice. Population, Space and Place 18: 241–259.

40.

Haandrikman

Van Wissen

(2011) Regional variation in short distance homogamy. Genus 67(1): 45–59.

41.

Hägerstrand

(1951) Omflyttningen och uppkomsten av kulturregioner. In: Enequist

(G)

(ed) Tätorter Och Omland. Uppsala: AB Lundequistska Bokhandeln, pp.100–110.

42.

Hägerstrand

(1957) Migration and area. Survey of a sample of Swedish migration fields and hypothetical considerations on their genesis. Migration in Sweden, Lund Studies in Geography Series B 13: 27–158.

43.

Hägerstrand

(1970) What about people in regional science? Papers in Regional Science 24(1): 6–21.

44.

Häkli

Paasi

(2003) Geography, space and identity. In: Öhman

Simonsen

(eds) Voices from the North: New Trends in Nordic Human Geograpy. Aldershot: Ashgate, pp.141–155.

45.

Hanson

(1993) The geography of everyday life. In: Gärling

and Golledge

(eds) Behaviour and Environment: Psychological and Geographical Approaches. Amsterdam: Elsevier, pp.249–269.

46.

Hedberg

Haandrikman

(2014) Repopulation of the Swedish countryside: Globalisation by international migration. Journal of Rural Studies 34: 128–138.

47.

Huschek

De Valk

(2012) Partner choice patterns among the descendants of Turkish immigrants in Europe. European Journal of Population = Revue Europeenne De Demographie 28: 241–268.

48.

Jennings

Axinn

Ghimirea

(2012) The effect of parents’ attitudes on sons’ marriage timing. American Sociological Review 77(6): 923–945.

49.

Kalmijn

(1991) Shifting boundaries. Trends in religious and educational homogamy. American Sociological Review 56: 786–800.

50.

Kalmijn

(1998) Intermarriage and homogamy: Causes, patterns, trends. Annual Review of Sociology 24: 395–421.

51.

Kalmijn

Van Tubergen

(2010) A comparative perspective on intermarriage: Explaining differences among national-origin groups in the United States. Demography 47: 459–479.

52.

Knippenberg

De Pater

(1988) De Eenwording Van Nederland. Schaalvergroting En Integratie Sinds 1800. Nijmegen: SUN.

53.

Kok

Mandemakers

(2008) Free choice from a limited supply. The marriage market in two Dutch provinces, 1840-1940. Romanian Journal of Population Studies 2(1): 82–104.

54.

Kolk

(2017) A life-course analysis of geographical distance to siblings, parents, and grandparents in Sweden. Population, Space and Place 23: (3). doi: 10.1002/psp.2020.

55.

Kupiszewski

Borgegård

L, E

Fransson

et al . (2001) Internal migration and regional population dynamics in Europe: Sweden case study. Working paper 01/01, School of Geography, University of Leeds.

56.

Kvillner

(1969) En studie i giftermålsfält. Fataburen. Nordiska Museets Och Skansens Årbok 1969: 61–72.

57.

Liefbroer

Billari

(2010) Bringing norms back in: A theoretical and empirical discussion of their importance for demographic behaviour. Population, Space and Place 16(4): 287–305.

58.

Løken

Lommerud

Lundberg

(2013) Your place or mine? On the residence choice of young couples in Norway. Demography 50 (1): 285–310.

59.

Maas

Zijdeman

(2010) Beyond the local marriage market: The influence of modernization on geographical heterogamy. Demographic Research 23(33): 933–962.

60.

Malmberg

Pettersson

(2007) Distance to elderly parents. Demographic Research 17(23): 679–704.

61.

Marble

Nystuen

(1963) An approach to the direct measurement of community mean information fields. Papers in Regional Science 11: 99–109.

62.

Mayfield

(1972) The spatial structure of a selected interpersonal contact: A regional comparison of marriage distances in India. In: English

Mayfield

(eds) Man, Space, and Environment. Concepts in Contemporary Human Geography. New York: Oxford University Press, pp. 385–401.

63.

Morrill

Pitts

(1967) Marriage, migration, and the mean information field: A study in uniqueness and generality. Annals of the Association of American Geographers 57: 401–422.

64.

Mulder

(2007) The family context and residential choice: A challenge for new research. Population, Space and Place 13: 265–278.

65.

Mulder

(2009) Leaving the parental home in young adulthood. In: Furlong

(ed) Handbook of Youth and Young Adulthood. New Perspectives and Agendas. Routledge, Milton Park, Oxon, pp.203–210.

66.

Mulder

Malmberg

(2012) Local ties and family migration. Paper presented at the European Population Conference, 13-16 June, Stockholm. Available at: http://epc2012.princeton.edu/papers/120184 (accessed 19 June 2018).

67.

Mulder

Wagner

(1993) Migration and marriage in the life course: A method for studying synchronized events. European Journal of Population = Revue Europeenne De Demographie 9: 55–76.

68.

Niedomysl

Östh

Van Ham

(2009) The globalisation of marriage fields: The Swedish case. Journal of Ethnic and Migration Studies 36: 1119–1138.

69.

Nilsson

Strandh

(1999) Nest leaving in Sweden: The importance of early educational and labor market careers. Journal of Marriage and Family 61(4): 1068–1079.

70.

Nowok

Van Ham

Findlay

et al . (2013) Does migration make you happy? A longitudinal study of internal migration and subjective well-being. Environment and Planning A 45: 986–1002.

71.

Öberg

Springfeldt

(eds) (1991) The Population. The National Atlas of Sweden. Stockholm: SNA.

72.

OECD (2010) OECD regional typology. Directorate for public governance and territorial development. Available at: www.oecd.org/dataoecd/35/62/42392595.pdf (accessed 18 June 2018).

73.

Ohlsson-Wijk

(2011) Sweden’s marriage revival: An analysis of the new-millennium switch from long-term decline to increasing popularity. Population Studies 65(2): 183–200.

74.

Östh

(2014) Introducing the EquiPop software: An application for the calculation of k-nearest neighbour contexts/neighbourhoods. Available at: http://equipop.kultgeog.uu.se/Tutorial/Introducing%20EquiPop.pdf (accessed 18 June 2018).

75.

Östh

Malmberg

Andersson

(2014) Analysing segregation using individualized neighbourhoods. In: Lloyd

Shuttleworth

Wong

(eds) Social-Spatial Segregation: Concepts, Processes and Outcomes. Bristol: Policy Press, pp.135–162.]

76.

Ouyang

Zhong

Xiao

et al . (2009) Residents’ intermarriage behavior in Guangzhou, China in last 60 years: Social geographic report. Espace Populations Sociétés 3: 485–495.

77.

Pred

(1997) Somebody else, somewhere else: Racisms, racialized spaces and the popular geographical imagination in Sweden. Antipode 29(4): 383–416.

78.

Pullum

Peri

(1999) A multivariate analysis of homogamy in Montevideo, Uruguay. Population Studies 53(3): 361–377.

79.

Relethford

Mielke

(1994) Marital exogamy in the Åland islands, Finland, 1750–1949. Annals of Human Biology 21(1): 13–21.

80.

Samers

(2010) Migration. London: Routledge.

81.

Sandström G (2011) Time-space trends in Swedish divorce behavior 1911–1974. Scandinavian Journal of History 36(1): 65–90.

82.

Scholten

Friberg

Sandén

(2012) Re-reading time-geography from a gender perspective: Examples from gendered mobility. Tijdschrift Voor Economische En Sociale Geografie 103(5): 584–600.

83.

Schwarz

Hassebrauck

(2012) Sex and age differences in mate-selection preferences. Human Nature 23: 447–466.

84.

Shorter

(1975) The Making of the Modern Family. Basic Books: New York.

85.

Sivaram

Richard

Rao

PSS

(1995) Early marriage among rural and urban females of South India. Journal of Biosocial Science 27(3): 325–331.

86.

Smaldino

Schank

(2012) Human mate choice is a complex system. Complexity 17(5): 11–21.

87.

Smits

Ultee

Lammers

(1998) Educational homogamy in 65 countries: An explanation of differences in openness using country-level explanatory variables. American Sociological Review 63: 264–285.

88.

Statistics Sweden (2008) Ungdomars Flytt Hemifrån. Demographic Reports 2008:5, Örebro: Statistics Sweden.

89.

Statistics Sweden (2010) Construction and Use of Labour Market Areas in Sweden. Örebro: SCB.

90.

Statistics Sweden (2018) Total fertility rates by region, gender and year. Statistical database. Available at: www.scb.se (accessed 14 February 2018).

91.

Stjernström

Strömgren

(2012) Geographical distance between children and absent parents in separated families. Geografiska Annaler B 94(3): 239–253.

92.

Stockdale

(2017) From ‘trailing wives’ to the emergence of a ‘trailing husband’ phenomenon: Retirement migration to rural areas. Population, Space and Place 23: ▪.

93.

Thulin

Vilhelmson

(2010) The internet, mobile phones and the geographies of everyday life. In: Hermelin

Jansson

(eds) Placing Human Geography: Sweden through Time and Place. Visby: YMER, Swedish Society for Anthropology and Geography, pp.277–309.

94.

Van de Putte

(2003) Homogamy by geographical origin: Segregation in nineteenth century Flemish cities (Gent, Leuven and Aalst). Journal of Family History 28(3): 364–390.

95.

Van Leeuwen

MHD

Maas

(2002) Partner choice and homogamy in the nineteenth century: Was there a sexual revolution in Europe? Journal of Social History 36(1): 101–123.

96.

Van Poppel

Ekamper

(2005) De Goudse horizon verruimd. Veranderingen in de herkomst van Goudse bruiden en bruidegoms. In: Kok

van Leeuwen

(eds) Genegenheid En Gelegenheid. Twee Eeuwen Partnerkeuze En Huwelijk. Amsterdam: Aksant, pp.181–212.

97.

Wright

Ellis

(2016) Perspectives on migration theory: Geography. In: White

(ed) International Handbook of Migration and Population Distribution. Dordrecht: Springer, pp.11–30.