The diverse geographies of mixed-ethnicity couples

Abstract

The geographical distribution of mixed-ethnicity (or mixed-race) couples is an important indicator of the spatial dimensions of cultural ‘mixing’ in a given society. A small number of studies have mapped the residential geographies of mixed-ethnicity couples, revealing distinctive residential patterns that differ from those of each partner’s respective ethnic group. Most such analyses have adopted broad pan-ethnic or racial categories (e.g. ‘black-white’, ‘white-Asian’). The Australian Census – which eschews broad racial categories – provides an opportunity to explore the unique residential geographies of different ‘types’ of mixed-ethnicity couples. Using customised 2011 Census data, this paper maps the residential geographies of diverse mixed-ethnicity couples in Australia’s most populous cities: Sydney and Melbourne. We focus on couples where one partner nominated ‘Anglo-European’ ancestry and the other nominated one of five ‘minority’ ancestries. Our findings highlight the value in disaggregating the coarse pan-ethnic or racial groupings adopted in some existing studies, and prescribed in some national censuses. The residential geographies of mixed-ethnicity couples challenge established perspectives on urban ethnic landscapes. As these couples grow in number, particularly in immigrant societies, they have the potential to fundamentally reshape established ethnic residential geographies away from patterns of residential segregation. Fine-grained analyses such as ours provide scope to explore the myriad directions in which these shifts will unfurl. They also provide a starting point for better understanding the preferences and pressures that shape the residential geographies of diverse mixed-ethnicity couples.

Keywords

Mixed-ethnicity census diversity residential geography segregation mixed-race households

Introduction

The growth in mixed-ethnicity (or mixed-race) partnerships in multi-ethnic societies in recent decades indicates that social barriers between some ethnic groups are becoming less significant (Kalmijn, 1998; Khoo, 2011). These relationships are, in turn, a powerful catalyst for social and demographic change (Khoo, 2011; Wright et al., 2003). Mixed-ethnicity partnerships facilitate interaction between people of different ethnicities, extending from the partners themselves to family and friendship networks (Kalmijn, 1998). In England and Wales, the number of mixed-ethnicity partnerships grew by over 30% between the 2001 and 2011 Censuses (Office for National Statistics, 2014); while in the US, the number of inter-racial couples grew by 28% between 2000 and 2010 (Lofquist et al., 2012). In Australia, 23% of all marriages formed in 2006 were between an Australian-born and overseas-born partner, a stark increase from 13% in 1990 (Khoo, 2011). These trends have driven a marked rise in populations of mixed-ethnicity individuals in the US, UK and Australia (Johnston et al., 2013; Khoo, 2011; Shih and Sanchez, 2009). Mixed-ethnicity partnerships are fundamentally altering both the ethnic composition of these societies, and their ethnic residential geographies.

We use the term ‘mixed-ethnicity’ to refer to couples in which the two partners have different ethnic backgrounds. While this definition appears straightforward, conceptualisations of ethnically or racially mixed couples are contingent upon the ethno-racial composition of societies, the nature of government population data and contextually specific attitudes toward ethnicity/race. Just as understandings of ethnic categories are fluid and socially constructed, so too are perceptions of which partnerships constitute mixed-ethnicity relationships (Jayasuriya, 2002). Decisions about ‘who counts’ as a mixed-ethnicity couple are spatially and temporally variable and highly complex. In Australia, successive waves of immigration have continually reshaped notions of acceptable and transgressive ethnic mixing. Following World War Two, large numbers of migrants arrived from Southern Europe. These migrants were initially perceived as culturally distant from the Anglo-Celtic Australian population (MacLeod, 2006), thus partnerships with Anglo-Celtic Australians were quite rare and contentious (Price and Zubrzycki, 1962). Yet the social and cultural distance between those groups has faded. Partnerships involving an Anglo-Celtic Australian and Italian partner, for example, are now both common and normalised. Today, more salient cultural barriers are perceived to exist between Australia’s Anglo-European ethnic majority and migrant communities from Asia, Africa and the Middle East (Blair et al., 2017).

Cognisant of this fluidity, our study does not include all couples with partners of different ethnic backgrounds. Instead, we focus on those couples where perceived cultural barriers remain salient, and whose partnerships have been shown to raise concern amongst some Australians (Blair et al., 2017). As discussed by Luke and Carrington (2000), such perceptions are often based on ‘visible differences’ between partners. Further detail on these definitions is provided in the methods section of the paper. Below, we review the established literature on the geographies of mixed-ethnicity couples. We use the term ‘mixed-race’ only when referring to other studies that have used that terminology in their own analyses.

The residential geographies of mixed-ethnicity couples

Geographers have established that mixed-ethnicity couples inhabit space unevenly (Holloway et al., 2005; Smith et al., 2011). The residential decisions of mixed-ethnicity couples – within financial and other constraints – thus have the potential to reshape established ethnic residential geographies away from patterns of segregation. Existing studies – focussed predominantly on large cities in the US and UK – have made four key observations regarding the unique residential geographies of mixed-ethnicity couples (Ellis et al., 2006, 2012; Holloway et al., 2005; Smith et al., 2011; White and Sassler, 2000). Mixed-ethnicity couples: (i) are more residentially dispersed than their broader ethnic groups; (ii) tend to live in moderately diverse neighbourhoods not dominated by single ethnic groups; (iii) are less likely to live in socio-economically disadvantaged neighbourhoods than broader ethnic minority populations; and (iv) gravitate towards areas known for progressive social attitudes. These observations are discussed further below. The empirical portion of this paper considers the salience of these findings in Australia, and the extent to which they hold true when broad ethnic categories are disaggregated.

Residential dispersal and neighbourhood diversity

When compared with co-ethnic couples (partners sharing the same ethnicity), mixed-ethnicity couples have lower levels of residential segregation (Ellis et al., 2012; Iceland and Nelson, 2010). Ethnic minority persons in mixed-ethnicity relationships are less likely to live in neighbourhoods with high concentrations of their own ethnic group, and accordingly are more dispersed than those in co-ethnic relationships (Ellis et al., 2006; Holloway et al., 2005). In Australia, Walker and Heard (2014) and Biddle (2013) found that Indigenous persons with non-Indigenous partners tend to live in urban areas, which typically have lower proportions of Indigenous residents. Our research in Sydney (Tindale et al., 2014) demonstrated that ethnic minority persons with ethnic majority partners were considerably less likely (than their co-ethnically partnered peers) to live in neighbourhoods with a high presence of same-ethnicity residents. Dispersal was also evident amongst ethnic majority persons: those with an ethnic minority partner were much less likely to live in predominantly ‘white’ parts of Sydney.

Mixed-ethnicity couples often live in diverse neighbourhoods with moderate proportions of each partner’s ethnic group. In a study of 12 US metropolitan areas, Holloway et al. (2005) found that most types of mixed-race households (combinations of white, black, Asian, Latino and American Indian) lived in more diverse neighbourhoods than white same-race couples, but less diverse neighbourhoods than non-white same-race couples. These couples occupied an ‘in-between’ space in the ethno-racial landscape of cities, not aligned with existing racial geographies (Holloway et al., 2005; Wright et al., 2011). Lending support to Holloway et al.’s (2005) ‘in-between’ thesis, Tindale et al. (2014) showed that mixed-ethnicity couples in Sydney live in more diverse areas than co-ethnic majority couples, but less diverse areas than co-ethnic minority couples.

‘In-betweenness’ may be a function of mixed-race couples’ desire to avoid neighbourhoods dominated by single racial groups, where their identities may be constrained (Holloway et al., 2005: 321). Diverse places may provide accepting environments (Dalmage, 2000). Of course, individuals who already live in a diverse neighbourhood may be more likely to meet a partner from another ethnic background. However, neighbourhoods are declining in their significance as meeting points for couples (Bozon and Héran, 1989; Kalmijn and Flap, 2001), hence recent studies argue that the residential geographies of mixed-ethnicity couples reflect the decisions made by these couples, rather than their in-situ formation (Ellis et al., 2012; Holloway et al., 2005).

Ethnicity intersects with other demographic characteristics in shaping the residential geographies of mixed-ethnicity couples. Neighbourhood diversity may be particularly important to mixed-ethnicity couples with dependent children (Dalmage, 2000; Twine, 1999). They may seek neighbourhoods where their children will not be “hyper-visible”, to minimise exposure to racism (Twine, 1999: 737). Geographers have shown that, in England and Wales, mixed-ethnicity couples with children cluster in multicultural areas more than those without children (Caballero et al., 2008). Education, income, housing tenure, religious affiliation, nativity and age also intertwine to produce the spatial differentiations observed in the geographic literature on mixed-ethnicity couples (Smith et al., 2011). For instance, Holloway et al. (2005) noted that higher-income, home-owning mixed-race couples tended to live in ‘whiter’ neighbourhoods than other mixed-race couples. Meanwhile Wright et al. (2013) revealed gender asymmetries: mixed-race couples’ settlement patterns tend to be more aligned with the male partner’s racial group. However, regression analyses have shown that the ethnicity of one’s partner exerts an independent effect on residential location after controlling for other demographic characteristics (Ellis et al., 2006; White and Sassler, 2000).

Neighbourhood socio-economic status and social attitudes

Existing studies have also shown that mixed-ethnicity couples are less likely to live in socio-economically disadvantaged neighbourhoods than broader ethnic minority populations (Tindale et al., 2014; Feng et al., 2014), even after controlling for their own personal socio-economic characteristics (White and Sassler, 2000). Relatedly, Tindale et al. (2014) identified mixed-ethnicity couples’ propensity to reside in expensive inner city areas of Sydney, a pattern that distinguishes them from broader ethnic minority populations. The latter exhibit highly suburbanised residential geographies. Interestingly, Tindale et al. (2014) found that mixed-ethnicity couples have a propensity to live in areas of Sydney with high concentrations of same-sex couples (Gorman-Murray and Brennan-Horley, 2010). With a reputation for progressive social attitudes, low levels of intolerance and highly diverse populations across a variety of axes, inner cities may provide residential environments where mixed-ethnicity households (and indeed same-sex households) can feel safe from discrimination (Tindale et al., 2014; Gorman-Murray and Brennan-Horley, 2010; Twine, 1999). We are unaware of other studies that have identified the inner city as an attractive residential environment for mixed-ethnicity couples. Here, we build upon our earlier analyses by investigating whether the propensity to live in inner city areas varies for diverse (disaggregated) mixed-ethnicity couples.

Disaggregation: The benefits and challenges of pulling apart broad ethnic categories

Geographers have almost universally adopted broad pan-ethnic or racial categorisations of mixed-ethnicity couples. US-based studies typically use ‘white’, ‘black’, ‘Asian’ and ‘Latino’, while UK-based studies use ‘white British’, ‘black Caribbean’, ‘black African’, ‘South Asian’ and ‘Other Asian’. Yet critical whiteness scholars have challenged the unproblematic acceptance of ‘white’ as a fixed and homogenous category, because it hides a multitude of diverse ethnicities and is temporally variable (Mateos, 2014; Pavlovskaya and Bier, 2012). In the US Census, people of Arab ethnicity are subsumed within the ‘white’ category, despite widespread evidence of racism against this group (Salaita, 2005). Equally, pan-ethnic ‘black’ and ‘Asian’ categories have been criticised for conflating several heterogeneous ethnic and religious groups with distinctive migration histories and residential geographies (Aspinall, 2003; Stillwell, 2010).

The propensity to adopt broad pan-ethnic or racial categories has important implications for studies of mixed-ethnicity couples, because these categories conceal potentially important differences in their geographies. Ellis et al. (2006) offered an exception, studying the effect of partnership on residential choice for eight foreign-born groups in Los Angeles, including four Asian sub-groups. For all groups, partnership with someone of a different national origin decreased the likelihood of living in a clustered own ethnic group neighbourhood. Holloway et al. (2005) called for more studies investigating the geographies of select types of mixed-race households; as did Smith et al. (2011) in the UK. This paper responds to this gap.

We use customised data from the 2011 Australian Census to explore the geographies of select types of mixed-ethnicity couples in Sydney and Melbourne. Our focus is on mixed-ethnicity couples in which one partner belongs to the Anglo-European ethnic majority, and the other is from one of five ethnic minority groups with a sizeable presence in Australia: Lebanese, Vietnamese, Filipino, Chinese and Indian. This study deconstructs the ‘white’ and ‘Asian’ categories commonly deployed in research on the residential geographies of mixed-ethnicity couples, shedding light on the diverse geographies of different mixed-ethnicity couple types. Such disaggregation enables greater responsiveness to the changing social salience of ethnic categories in different places, and at different times. As noted earlier, ‘who counts’ as a mixed-ethnicity couple has changed over time – pulling apart categories that are usually considered in aggregate offers a means of responding to such changes.

However, there is often a necessary trade-off between the granularity of ethnic groups and geographic scale. Finer-grained groupings are available within the US and UK census classifications, but these are usually aggregated in analyses of small geographical areas, to preserve confidentiality in the data (Mateos et al., 2009). Equally, trade-offs need to be made between the granularity of ethnic categories, and the range of socio-demographic variables that can be considered. These issues are discussed further below. Given these trade-offs, we argue that different types of studies are needed to shed light on the geographies of mixed-ethnicity couples – those that adopt a fine-grained approach to geography but use broad ethnic/racial categories (as has typically been the case); alongside those that adopt fine-grained ethnic groupings but at coarser geographical scales (as in the present study).

Methods

Our analysis deploys data from the 2011 Australian Census. Information on partners’ ethnic backgrounds is not openly available, so a customised dataset was purchased from the Australian Bureau of Statistics (ABS). The data tables, which provide 100% coverage of census responses, consisted of counts of mixed-ethnicity and co-ethnic couples across 86 areas within Sydney and Melbourne. We included all co-habiting couples (de facto and formally married; same-sex and opposite-sex).

The Australian Census includes three variables commonly used as indicators of ethnicity: country of birth, language spoken at home and ancestry. We used the ancestry question as it most readily facilitates the inclusion of second and later generation migrants. The census form asks ‘What is the person’s ancestry?’ and advises respondents to consider the origins of their parents/grandparents. Our conceptualisation of ethnicity is based on self-identified ancestry, but does not distinguish between immigrant generations. Adding country of birth to our customised dataset would have resulted in a preponderance of small cell counts, impacting reliability.

As discussed earlier, the social salience of different types of mixed-ethnicity couples varies temporally and spatially, necessitating nuanced analysis and flexible categories. Accordingly, only some couple-types were included in our analysis. When there is a “visible phenotypical difference” between partners, they are at greater risk of experiencing discrimination in their everyday lives, which may affect residential decisions (Luke and Carrington, 2000: 9; Wright et al., 2003). We restricted our analysis to a subset of mixed-ethnicity couples – those in which one partner is part of the ethnic majority (defined below), and the other partner is from one of five fine-grained ‘visible’ ethnic minority groups with sufficiently large populations to enable local scale geographical analyses (see also Tindale et al., 2014).

The ‘ethnic majority’ category was constructed within the framework of the Australian Standard Classification of Cultural and Ethnic Groups (ASCCEG) (ABS, 2011a). It includes any ancestry responses from the following categories: Australian, New Zealander (excluding Maori), North-West European and Caucasian. We refer to this group as ‘Anglo-European’. It is worth noting that this group excludes Southern and Eastern Europeans. As noted by Farquharson (2007), Southern and Eastern Europeans occupy an ‘intermediate’ position in Australia’s contemporary ethnic landscape. They have high levels of integration, but are not yet definitively accepted as part of the dominant (white) Anglo-Celtic and Northern European culture (see also Tindale et al., 2014). Due to small population numbers, it was only possible to include five ‘visible’ ethnic minority groups: Lebanese, Vietnamese, Filipino, Chinese and Indian. This paper thus focuses on mixed-ethnicity couples in which one partner is ‘Anglo-European’ (as defined above), and the other identifies with one of those five ethnicities. We refer to these couples as ‘mixed Lebanese couples’, ‘mixed Vietnamese couples’, and so on. To illustrate the importance of using fine-grained ethnic categories, we also briefly explore the geographies of more broadly defined ‘mixed Asian’ couples. The ‘Asian’ group aggregates the South-East Asian, North-East Asian and Southern and Central Asian categories in the ASCCEG, and closely approximates the ‘Asian’ category adopted in international studies (e.g. Holloway et al., 2005; Smith et al., 2011).

Dual ancestry responses complicated our custom data specifications but enriched our data. In 2011, 32% of Census respondents nominated two ancestries (ABS, 2012). While previous studies of mixed-ethnicity couples have excluded dual ancestry individuals, we considered this too large a population to ignore. To avoid ‘overlapping’ ethnicities between partners in mixed-ethnicity couples, we only included dual ancestry persons who nominated both ancestries within the same category. For example, a person who stated the ancestries English and Australian (the most common combination) would be included in the Anglo-European category. However an individual of dual English (in the Anglo-European group) and Chinese ancestry would be excluded from both of those categories. Persons who stated two ancestries in different fine-grained ethnic minority categories (e.g. Chinese and Lebanese) were also excluded from the analysis; as were those who stated a second ancestry not included in our analysis (e.g. South African or Brazilian).¹ This approach excluded a small minority of the population. Of all partnered persons in Sydney and Melbourne who stated at least one ancestry in the six included categories, only seven per cent were excluded because they stated a second ancestry in a different category.

Data were obtained at Statistical Area Level 3 (SA3) within the Greater Sydney and Greater Melbourne metropolitan areas (referred to henceforth as Sydney and Melbourne). Forty-six SA3s were included covering Sydney, and 40 covering Melbourne, with a mean population of around 98,000. Within metropolitan areas, SA3s consist of economic hubs or “groups of related suburbs” (ABS, 2011b: 26). Research in the US and UK usually uses smaller census tracts/wards that resemble neighbourhoods. SA3s were the smallest level of geography at which it was possible to use the detailed ethnic minority categories noted above.

The geographies of different types of mixed-ethnicity couples are analysed through location quotients² (LQs), which depict levels of concentration in each area relative to the metropolitan average. An LQ of 1.0 means the proportion of mixed-ethnicity couples in an SA3 is identical to the broader metropolitan area. LQs greater and less than one indicate above- and below-average concentrations respectively. The geographies of co-ethnic couples (where partners share the same ethnicity) are used for comparison. Maps were developed using QGIS to depict the spatial patterning of high LQ scores.

Our results also show the percentage distributions of mixed-ethnicity couples across SA3s based on: each ethnic minority group’s percentage share of the SA3 population; the Anglo-European share of the SA3 population; and overall ethnic diversity. Census data on the ethnic composition of SA3s were extracted using ABS TableBuilder Pro. SA3-level ethnic diversity was calculated using the standardised entropy index³ on the basis of 11 ethnic groups.⁴ SA3s across both cities were ranked and divided into diversity quintiles. Relationships between the prevalence of mixed-ethnicity couples and SA3 characteristics (including percentage Anglo-European, percentage identifying with the relevant ethnic minority group, percentage with a bachelor degree, and population density) were modelled using negative binomial regression, as discussed in the “The residential geographies of mixed-ethnicity couples in Sydney and Melbourne” section.

In later sections of the analysis, we disaggregate each mixed-ethnicity couple type by two additional variables: the ethnicity of the male partner (for opposite-sex couples) and the presence of dependent children. These attributes were selected based on evidence that gender (Wright et al., 2013) and family composition (Caballero et al., 2008; Twine, 1999) impact mixed-ethnicity couples’ decisions about where to live. As our data consist of counts of couples in geographic areas, disaggregation according to other axes of difference was not feasible – it would have resulted in a proliferation of small numbers. The core aim of the paper, however, is to highlight the diverse geographies of different types of mixed-ethnicity couples which are often obfuscated within broader ethnic categorisations.

The residential geographies of mixed-ethnicity couples in Sydney and Melbourne

The 2011 Census recorded 35,636 mixed-ethnicity couples in Sydney and Melbourne where one partner fit our ethnic majority (Anglo-European) definition, and the other was Lebanese, Vietnamese, Filipino, Chinese or Indian (Table 1). Together, these couples comprise just over two per cent of all co-habiting couples in Sydney and Melbourne combined. Mixed Chinese couples are by far the most numerous, and mixed Vietnamese couples the least numerous of these five couple types, in each city (see Table 1).

Table 1.

Sizes of mixed-ethnicity couple and co-ethnic couple populations, Sydney and Melbourne.

Couple type	Sydney		Melbourne		Total
Couple type	Count	% of all couples	Count	% of all couples	Count	% of all couples
Mixed Lebanese	3349	0.37	923	0.11	4272	0.24
Co-ethnic Lebanese	20,939	2.30	5383	0.64	26,322	1.51
Mixed Vietnamese	1093	0.12	1134	0.14	2227	0.13
Co-ethnic Vietnamese	11,450	1.26	11,463	1.37	22,913	1.31
Mixed Filipino	5097	0.56	3154	0.38	8251	0.47
Co-ethnic Filipino	11,818	1.30	5842	0.70	17,660	1.01
Mixed Chinese	9339	1.03	6389	0.76	15,728	0.90
Co-ethnic Chinese	63,420	6.98	39,534	4.72	102,954	5.90
Mixed Indian	2538	0.28	2620	0.31	5158	0.30
Co-ethnic Indian	27,658	3.04	25,921	3.10	53,579	3.07
Total of five mixed-ethnicity	21,416	2.36	14,220	1.70	35,636	2.04
Total of five co-ethnic minority	135,285	14.89	88,143	10.53	223,428	12.79
Co-ethnic Anglo-European	381,866	42.02	369,840	44.16	751,706	43.05

Source: generated by the authors using data supplied by the ABS.

Figures 1 and 2 demonstrate the importance of accounting for diverse ethnic minority backgrounds. These couples are not homogenous – and at times have divergent residential geographies. Using LQs, Figure 1 depicts the residential distribution of ‘mixed Asian’ couples – those in which one partner is Anglo-European and the other belongs to the aggregated ‘Asian’ group. Darker shading represents higher LQ values. For ease of interpretation, those SA3s with ‘well below average’ LQs (<0.75) have dotted patterning, those with ‘around average’ LQs (0.75 to <1.25) have lined patterning, and those with ‘well above average’ LQs (≥1.25) have no additional patterning. For comparative purposes, any SA3s where co-ethnic Asian couples registered LQs greater than 2.0 (hotspots) are indicated using bold outlines. Using the same symbology, Figure 2 depicts the residential geographies of three ‘subsets’ of mixed Asian couples – mixed Filipino, mixed Chinese and mixed Indian – and highlights hotspots for co-ethnic Filipino, co-ethnic Chinese and co-ethnic Indian couples, respectively.

Aggregated mixed Asian couples are strikingly concentrated in inner city Sydney and Melbourne (Figure 1; inner city SA3s labelled), contrasting with the suburban hotspots of co-ethnic Asian couples. However, Figure 2 reveals multiple geographies for finer-grained Asian sub-groups. Not all types of mixed Asian couples reside at their greatest concentrations in inner city areas. Inner city SA3s have the highest concentrations of mixed Chinese and mixed Indian couples, but suburban SA3s are focal points for mixed Filipino couples. Further, the highest concentrations of mixed Chinese and mixed Indian couples do not coincide with hotspots for their comparative co-ethnic minority couples. Yet the highest concentrations of mixed Filipino couples largely overlap with hotspots of co-ethnic Filipino couples, suggesting lesser residential dispersal. The following sections explore the diverse geographies of mixed-ethnicity couples in Sydney and Melbourne in greater detail; comparing them against their co-ethnically partnered peers. We consider whether the key findings foregrounded in existing studies of the geographies of mixed-ethnicity couples (based on broad pan-ethnic and racial groups) hold true when ethnicities are disaggregated.

A preference for inner city living?

Tindale et al. (2014) identified a propensity for inner city living amongst Sydney’s mixed-ethnicity couples. Figure 2 shows that, when disaggregated, diverse mixed-ethnicity couple types concentrate in different parts of Sydney and Melbourne – whether by choice or due to financial or other constraints. Mixed Chinese, mixed Indian and mixed Vietnamese couples (not shown) reside at high concentrations in inner city and nearby SA3s, contrasting with the suburban hubs of their co-ethnic minority peers. In Sydney, over 20% of mixed Chinese, mixed Indian and mixed Vietnamese couples live in either Sydney Inner City or an adjacent SA3, compared to just 13.3% of all couples, and even lower percentages of co-ethnic Chinese (10.4%), co-ethnic Indian (4.5%) and co-ethnic Vietnamese (5.4%) couples. Melbourne City and adjacent SA3s are home to 11.2% of all couples, but 20.9% of mixed Chinese couples, 25.2% of mixed Vietnamese couples, and 16.7% of mixed Indian couples. These three mixed-ethnicity couple types are more likely to reside in inner city areas than both their co-ethnic minority counterparts and co-ethnic Anglo-European couples (13.4% in Sydney; 10.4% in Melbourne). However, the proportions of mixed Filipino and mixed Lebanese couples living in the inner cities and adjacent SA3s are around half those of the other mixed-ethnicity couple types. The highest concentrations of these mixed-ethnicity couples are in the suburban SA3 hotspots for their respective co-ethnic minority couples.

Residential dispersal

International literature suggests that individuals in mixed-ethnicity partnerships are more geographically dispersed than those in co-ethnic partnerships. They are less likely to live in residential areas with high concentrations of their own ethnic groups (Ellis et al., 2006; Feng et al., 2014; Iceland and Nelson, 2010). This is apparent for all five mixed-ethnicity couple types in Sydney and Melbourne. Table 2 shows the extent to which couples congregate in SA3s with high concentrations of either partner’s ethnic group. ‘High concentrations’ are SA3s where the particular ethnic minority group’s share of the local population is over twice their share of the metropolitan area population. High concentrations of Anglo-Europeans are SA3s where they constitute greater than 75% of the population.

Table 2.

Percentage of mixed-ethnicity and co-ethnic couples living in SA3s with (i) high concentration of own minority group^a (LQ > 2) and (ii) high concentration of Anglo-Europeans (>75% of local population), Sydney and Melbourne.

Couple type	% living in SA3s with high concentration of own minority group		% living in SA3s with high concentration of Anglo-Europeans
Couple type	Sydney	Melbourne	Sydney	Melbourne
Mixed Lebanese	28.3	29.7	12.2	12.0
Co-ethnic Lebanese	62.8	62.1	1.6	2.8
Mixed Vietnamese	29.6	26.5	9.7	7.7
Co-ethnic Vietnamese	78.8	62.6	1.0	0.9
Mixed Filipino	16.4	23.3	16.3	19.2
Co-ethnic Filipino	43.1	51.8	2.3	4.6
Mixed Chinese	18.8	23.9	11.4	11.8
Co-ethnic Chinese	35.7	38.6	2.4	3.1
Mixed Indian	11.1	9.9	14.5	14.1
Co-ethnic Indian	39.0	20.7	2.3	3.9
Co-ethnic Anglo-Euro.	–	–	33.8	29.9

SA3: Statistical Area Level 3.

Source: generated by the authors using data supplied by the ABS.

For example, first row, first column shows the percentage of mixed Lebanese couples in Sydney who live in SA3s with high concentrations of the Lebanese population.

Mixed-ethnicity couples in Sydney and Melbourne are much less likely than co-ethnic minority couples to live in SA3s with high own ethnic minority group concentrations. For instance, 78.8% and 62.6% of co-ethnic Vietnamese couples in Sydney and Melbourne (respectively) reside in a high-concentration Vietnamese SA3. The comparable figures for mixed Vietnamese couples are just 29.6% and 26.5%. The proportion of mixed-ethnicity couples residing in SA3s with high own-minority group concentrations is typically less than half that of co-ethnic couples (Table 2). Yet important differences are apparent. Mixed Vietnamese and mixed Lebanese couples have the strongest tendency to live in areas with high own-minority group populations, despite being far more dispersed than their co-ethnically partnered peers.

All mixed-ethnicity couple types considered in this study are far more likely than their respective co-ethnic minority couples to live in SA3s with high concentrations (>75%) of Anglo-Europeans (Table 2). In each city, these areas are home to less than 5% of each co-ethnic minority couple type, but around one-third of co-ethnic Anglo-European couples. Mixed-ethnicity couples fall in-between those extremes. In both cities, mixed Filipino and mixed Indian couples are most likely to live in SA3s with high Anglo-European concentrations. Mixed Vietnamese couples have the lowest propensity to live in these locations, but remain far more likely to do so than co-ethnic Vietnamese couples. These patterns may arise from distinctive cultural preferences, or occur due to pragmatic concerns (e.g. housing affordability) and require further unpacking through qualitative research.

Ethnically mixed couples in ethnically diverse areas?

Several studies have shown that mixed-ethnicity couples reside in ethnically diverse neighbourhoods (Holloway et al., 2005; Smith et al., 2011). All SA3s (Sydney and Melbourne together) were grouped into quintiles based on entropy diversity scores. Table 3 shows the distribution of each couple type according to SA3-level ethnic diversity. The results support Holloway et al.’s (2005) in-between thesis, regarding the distribution of mixed-ethnicity couples. All types of mixed-ethnicity couples are less likely to live in the most diverse SA3s (Quintile 5) than their co-ethnic minority peers, but are more likely to do so than co-ethnic Anglo-European couples. Conversely, mixed-ethnicity couples live in the least diverse SA3s (Quintile 1) at higher levels than their co-ethnic minority counterparts, but are less inclined to do so than co-ethnic Anglo-European couples.

Table 3.

Percentage distribution of mixed-ethnicity and co-ethnic couples across SA3s by diversity quintiles, Sydney and Melbourne combined.

Couple type	Diversity quintile
Couple type	1 (Very low)	2 (Low)	3 (Moderate)	4 (High)	5 (Very high)
Mixed Lebanese	10.3	13.2	16.2	20.6	39.8
Co-ethnic Lebanese	1.5	3.4	7.6	16.1	71.4
Mixed Vietnamese	6.9	15.0	23.4	20.9	33.8
Co-ethnic Vietnamese	0.5	2.8	9.8	12.1	74.8
Mixed Filipino	15.1	16.6	20.6	19.8	27.9
Co-ethnic Filipino	2.5	7.5	18.0	19.6	52.5
Mixed Chinese	9.3	21.5	28.1	24.8	16.4
Co-ethnic Chinese	1.5	12.2	20.9	33.6	31.9
Mixed Indian	11.8	20.6	26.1	22.3	19.2
Co-ethnic Indian	2.2	10.0	19.0	27.6	41.2
Co-ethnic Anglo-European	25.7	26.9	23.4	14.8	9.2
All couples	14.7	18.9	22.8	21.1	22.4

Source: generated by the authors using data supplied by the ABS.

Notwithstanding these broad trends, important differences exist. The distributions of mixed Vietnamese and mixed Lebanese couples peak at the most diverse SA3s (Quintile 5). These SA3s are also places with high concentrations of Vietnamese and Lebanese persons, so it is difficult to untangle the attraction of diversity versus own-group presence. Mixed Filipino, mixed Chinese and mixed Indian couples’ distributions peak in SA3s of moderate or moderately high diversity. Their residential focal points are locations where neither partner’s ethnic group is present at particularly high levels – echoing trends among black-white couples in the US (Wright et al., 2011). These insights disrupt expectations about the ‘types’ of places in which mixed-ethnicity couples typically reside. Finer-grained ethnic groupings make plain that there is no single geography of mixed-ethnicity couples. Rather, there are multiple geographies that appear to be contingent upon the ethnic minority groups involved.

Variations by gender and family composition

The spatial patterns identified thus far are based solely on combinations of partners’ ethnicities. The following analysis disaggregates each mixed-ethnicity couple type according to two attributes likely to intersect with ethnicity to shape residential outcomes: the ethnicity of the male partner, and the presence of dependent children.

Research in the US suggests that mixed-race couples’ residential geographies more closely reflect the male partner’s race (Wright et al., 2013). We find no substantive evidence of gender imbalances in Australian mixed-ethnicity couples’ propensity to live in areas with either high ethnic minority or high Anglo-European concentrations.⁵ The largest disparity (of just 5.6 percentage points) is among mixed Filipino couples in Sydney: 21.8% of those with a male Filipino partner lived in SA3s with high Filipino concentrations, compared to 16.2% of those with a male Anglo-European partner. Distributions across diversity quintiles are also barely affected by the intersection of gender and ethnicity. The largest difference (6.4 percentage points) is among mixed Lebanese couples in the most diverse SA3s (Quintile 5), which were home to 42.5% of those with a male Lebanese partner, compared to 36.1% of those with a male Anglo-European partner. Mixed Lebanese, mixed Chinese and mixed Indian couples with a male Anglo-European partner were very slightly more likely to live in inner city or adjacent SA3s in both cities, but the opposite was true for mixed Filipino couples. The most pronounced difference was among mixed Lebanese couples in Sydney: 15.2% of those with a male Anglo-European partner lived in Sydney Inner City or an adjacent SA3, compared to 9.1% of those with a male Lebanese partner. Thus in this study, gender had minimal impact on the residential geographies of mixed-ethnicity couples.

The presence of dependent children can motivate mixed-ethnicity couples to live in diverse neighbourhoods (Caballero et al., 2008; Dalmage, 2000; Twine, 1999). Yet we find no substantial differences by parenting status in the propensity for mixed-ethnicity couples to reside in highly diverse SA3s. Across almost all mixed-ethnicity couple types in both cities, those with dependent children are more likely to live in SA3s with high own-minority group concentrations, although the differences are very small (Figure 3). The sole exception is mixed Indian couples in Melbourne: those with dependent children are slightly less likely than their childless counterparts to reside in an SA3 with a high Indian concentration. Co-ethnic minority couples are considerably more likely (than all mixed-ethnicity couple types) to live in SA3s with high concentrations of their respective minority ethnic group – whether they have dependent children, or not (Figure 3). Conversely, all types of mixed-ethnicity couples are more likely to live in SA3s with high Anglo-European concentrations than their comparative co-ethnic minority couples, irrespective of children. As expected, mixed-ethnicity couples without dependent children more commonly live in the inner cities (or adjacent SA3s) of Sydney or Melbourne. But, in both cities, all types of mixed-ethnicity couples with dependent children more commonly live in inner city areas than co-ethnic minority couples with dependent children. The differences between mixed-ethnicity and co-ethnic couples are much larger than differences resulting from parenting status (see Figure 3). Thus, accounting for the presence or absence of children does not erase differences in the residential geographies of mixed-ethnicity and co-ethnic couples.

Negative binomial regression

The final component of our analysis ran a set of multivariate regression models to explore whether relationships between concentrations of mixed-ethnicity couples and those of their respective ethnic groups are statistically significant after controlling for other SA3 characteristics. We follow Wright et al.’s (2011) approach, and build on their work by using finer-grained ethnic groups, and by further disaggregating couples by gender configurations and the presence/absence of children. We focus this analysis on mixed Chinese and mixed Indian couples – the only couple types with sufficient total counts, once disaggregated according to the ethnicity of the male partner and the presence/absence of dependent children. Our dependent variables consisted of overdispersed⁶ count data, so we adopted negative binomial regression with a log link function (Wright et al., 2011).

Our models estimated SA3 counts of mixed Chinese and mixed Indian couples, both overall and when disaggregated according to the gender and parenting status. We also modelled counts of co-ethnic Chinese, Indian and Anglo-European couples, in total and disaggregated by parenting status. The independent variables consisted of SA3-level characteristics. Ethnicity-based SA3-level variables included the SA3 percentage Anglo-European and percentage Chinese or Indian. The entropy measure of SA3 ethnic diversity was excluded due to high correlation with the SA3 percentage Anglo-European. The percentage with a bachelor degree provided a measure of residents’ socio-economic status in each SA3. Population density (persons per square kilometre) captured the position of each SA3 within the urban landscape (higher population densities tend to be found closer to the inner city). Finally, the total number of couples in each SA3 was included to control for population size, and a dummy variable controlled for potential metropolitan-area variations between Sydney and Melbourne.

The results of the negative binomial models are shown in Table 4. Following Wright et al. (2011), the parameter estimates for each independent variable were multiplied by its standard deviation, and then exponentiated. The resulting numbers are interpreted as the estimated factor change in predicted counts of the dependent variable with a one standard deviation increase in the independent variable (Wright et al., 2011). Thus, in the model for mixed Chinese couples, the value of 1.311 for ‘SA3 percentage Chinese’ indicates that a one standard deviation increase in the SA3 percentage Chinese generates a 31.1% increase in the count of mixed Chinese couples. Similarly, counts of mixed Indian couples increase by 20.2% in response to a one standard deviation increase in ‘SA3 percentage Indian’. The corresponding values are substantially larger for co-ethnic Chinese and Indian couples respectively (96.3% and 117.5%). The models thus confirm the results presented earlier: mixed-ethnicity couples are drawn to areas with high proportions of their own ethnic minority groups, but to a far lesser extent than co-ethnic minority couples. Increases in ‘SA3 percentage Anglo-European’ produce the opposite effect – moderate increases in counts of mixed Chinese and mixed Indian couples (16.3% and 17.4%, respectively), but sizeable decreases in counts of their co-ethnic peers (−39.9% and −31.7%). Lending support to the ‘in-between’ thesis (Holloway et al., 2005), the responsiveness of mixed Chinese and mixed Indian couples to increases in ‘SA3 percentage Anglo-European’ fall in-between their co-ethnic minority and co-ethnic Anglo-European reference groups.

Table 4.

Negative binomial regression results – parameter estimates multiplied by standard deviation of predictor variable, then exponentiated.

Dependent variable (couple type)	SA3 % Chinese	SA3 % Anglo-European	SA3 % with bachelor degree	SA3 pop. density
Co-ethnic Chinese	1.963	0.601	1.443	0.939
Mixed Chinese	1.311	1.163	1.494	1.125
Co-ethnic Anglo-European	1.017	1.573	1.078	0.985
Mixed Chinese (male Anglo-Euro.)	1.332	1.161	1.476	1.107
Mixed Chinese (male Chinese)	1.341	1.236	1.489	1.014
Co-ethnic Chinese (DC)	2.035	0.601	1.385	0.840
Mixed Chinese (DC)	1.389	1.146	1.586	0.938
Co-ethnic Anglo-European (DC)	0.967	1.590	1.167	0.882
Co-ethnic Chinese (NDC)	1.914	0.613	1.475	1.041
Mixed Chinese (NDC)	1.260	1.181	1.406	1.282
Co-ethnic Anglo-European (NDC)	1.021	1.564	1.052	1.062
	SA3 % Indian	SA3 % Anglo-European	SA3 % with bachelor degree	SA3 pop. density
Co-ethnic Indian	2.175	0.683	1.059	1.030
Mixed Indian	1.202	1.174	1.146	1.223
Co-ethnic Anglo-European	1.050	1.610	1.071	1.003
Mixed Indian (male Anglo-Euro.)	1.192	1.167	1.184	1.185
Mixed Indian (male Indian)	1.178	1.161	1.108	1.226
Co-ethnic Indian (DC)	2.292	0.689	1.015	0.902
Mixed Indian (DC)	1.247	1.208	1.190	1.053
Co-ethnic Anglo-European (DC)	1.088	1.720	1.113	0.923
Co-ethnic Indian (NDC)	1.930	0.659	1.130	1.164
Mixed Indian (NDC)	1.150	1.141	1.091	1.407
Co-ethnic Anglo-European (NDC)	1.027	1.542	1.045	1.061

DC: dependent children in family; NDC: no dependent children in family.

Source: generated by the authors using data supplied by the ABS.

Bold numbers indicate significant predictor variables at p < 0.05. Parameter estimates for SA3 population size and metropolitan area dummy variables are excluded to save space.

The additional predictor variables also provide valuable insights into mixed-ethnicity couples’ residential patterns. Mixed Chinese and mixed Indian couples were both drawn to SA3s with higher percentages of residents with a bachelor degree. In terms of population density, a one standard deviation increase significantly predicts increases in both mixed Chinese (12.5%) and mixed Indian (22.3%) couples, but was not significant in the models for co-ethnic couples. This supports our contention that higher density, cosmopolitan inner city areas are attractive residential locations for mixed-ethnicity couples.

When the ethnicity of the male partner was accounted for, there were few notable changes in the results (Table 4). SA3 ‘percentage own minority group’ (Chinese or Indian) and SA3 ‘percentage Anglo-European’ generate increases in counts of mixed Chinese and mixed Indian couples of both gender configurations. Population density significantly predicts increases in counts of mixed Indian couples (irrespective of gender configurations) and mixed Chinese couples with a male Anglo-European partner.

Finally, the draw of own-minority group presence in the SA3 is higher for mixed Chinese and mixed Indian couples who have dependent children, than for those who do not. This potentially points to the importance of raising mixed-ethnicity children in residential environs that are supportive of their ethnic minority identities. However, counts of co-ethnic Chinese and co-ethnic Indian couples with dependent children grow more sharply in response to increases in SA3 percentage Chinese and Indian respectively (Table 4). Thus, there are clear differences in residential outcomes for mixed-ethnicity and co-ethnic couples even after accounting for parenting status.

Conclusions

Literature on the geographies of mixed-ethnicity couples has predominantly focussed on broad pan-ethnic or racial groups. Researchers have generally deployed large categories such as ‘white’ and ‘Asian’ that enable analyses at very fine spatial scales. While the results of these analyses provide rigorous statistical insight into the neighbourhood dynamics of mixed-ethnicity partnering, they obfuscate the distinctive experiences of the groups subsumed within those broader categories. Findings thus tend to be skewed towards the geographies of numerically dominant mixed-ethnicity couples (e.g. mixed Chinese couples in Australia).

We used customised data from the 2011 Australian Census to reveal the diverse residential geographies of five types of mixed-ethnicity couples in Australia’s two largest and most ethnically diverse cities. When considered in aggregate, mixed Asian couples in Sydney and Melbourne are heavily concentrated in inner city areas. However, when finer-grained Asian ancestry groups are considered separately, we find that the inner city is the primary residential location for mixed Indian and mixed Chinese couples only. Mixed Filipino couples evince suburban residential concentrations more closely aligned with those of co-ethnic Filipino couples. Mixed Vietnamese couples have high concentrations in both the inner city and outer suburban areas. Further, we find that mixed Lebanese couples – which are typically subsumed by the ‘white’ racial category – tend to concentrate most heavily in similar suburban regions to co-ethnic Lebanese couples.

Through its focus on fine-grained ethnic minority groups, this paper adds nuance to established theories regarding the residential locations of mixed-ethnicity couples and how these connect to cities’ broader ethnic geographies. We have shown that mixed-ethnicity couples vary in their degree of residential concentration and in their responsiveness to SA3 ethnic diversity. Mixed Lebanese and mixed Vietnamese couples are most spatially concentrated, and are more likely to live in SA3s of high ethnic diversity, with higher proportions of their own ethnic minority groups. Mixed Chinese and mixed Indian couples more frequently reside in parts of the city with lower ethnic diversity and larger ethnic majority populations. Mixed Filipino couples are most likely to live in SA3s in the highest category of ethnic majority concentration.

These findings offer important insights into the implications of mixed-ethnicity partnerships for urban ethnic geographies. While dispersal occurs across all groups, not all mixed-ethnicity couple types display an equal tendency to shift away from patterns of ethnic segregation. Our results support Holloway et al.’s (2005) ‘in-between’ thesis regarding the residential geographies of mixed-ethnicity couples in relation to each partner's respective ethnic group. This pattern is evoked in the geographical distributions of all of our mixed-ethnicity couple types. Yet our results show that there are residential spaces where mixed-ethnicity couple concentrations do not neatly fall in-between those of their corresponding co-ethnic minority and co-ethnic majority couples. Namely, mixed-ethnicity couples in Sydney and Melbourne tend to live in inner city areas at higher proportions than either of their comparative co-ethnic couple types. The spatial patterns of mixed-ethnicity couples in Australia are both distinctive and reflective of the geographies of their constituent ethnic groups.

There are many possible explanations for the distinctive residential geographies of different types of mixed-ethnicity couples. Differences in socio-economic profiles may afford some mixed-ethnicity couples the residential mobility to move into a larger range of areas. Socio-economic status may also affect the geographical location of partners’ workplaces (e.g. corporate careers in the inner city) – with implications for residential decision-making. Variations in cultural norms may lead some mixed-ethnicity couple types to prefer to stay close to extended family. Further research is required to determine which attributes of inner cities are attractive to some mixed-ethnicity couples, while others maintain suburbanised concentrations (whether by choice or constraint). Future research needs to better understand what factors beyond neighbourhood ethnic diversity and socio-economic status shape the residential geographies of mixed-ethnicity couples. The varied geographies of the mixed-ethnicity couple types considered in our study may reflect distinctive cultural preferences, socio-economic limitations, unique migration histories and durations of residence in Australia, family ties, industries of employment, proximity to workplaces and gendered power dynamics. We encourage future quantitative studies of the geographies of mixed-ethnicity couples to focus on fine-grained ethnic groupings, whenever possible. However, in order to unpack these complexities it will be necessary to expand the use of qualitative methods in this field of study. Such efforts are an integral part of our own ongoing research on the geographies of mixed-ethnicity couples in Australia.

This article has responded to calls for greater attention to the residential outcomes of ethnic mixing within couples and households (Holloway et al., 2005; Smith et al., 2011). As mixed-ethnicity partnerships become increasingly common in multi-ethnic societies, traditional understandings of ethnic diversity and segregation across urban space will be challenged. The preponderance of evidence indicates that mixed-ethnicity couples will, over time, contribute to more diffuse urban ethnic geographies and a concomitant decrease in ethnic residential segregation. This has important implications at the societal level – but also at the household scale. We have argued that a focus on fine-grained ethnic categories facilitates insights into diverse residential geographies. However, this approach also supports fluid analyses – that adapt to take into account the social salience of different mixed-ethnicity couples at different times and in different places.

To the extent that mixed-ethnicity couples and their children challenge existing ethnic residential geographies, they may find themselves living in neighbourhoods with minimal previous exposure to ethnic diversity, and in which they may be confronted by social expectations around the cultural (in)compatibility of particular ethnic groups. These experiences are likely to raise particular challenges for couples in which the two partners are ‘visibly different’ from one another, such as those included in our analysis. As pioneers, in predominantly white neighbourhoods, these couples and their children may face discrimination. Such experiences demand research attention and – if found to be pervasive – intervention.

Footnotes

Acknowledgements

The authors would like to thank the staff at the Australian Bureau of Statistics (ABS) Information Consultancy Service for providing the customised data tables, and Dr Chris Brennan-Horley for his advice and assistance in developing the maps. We would also like to thank the reviewers for their helpful feedback on the original manuscript. This research has been conducted with the support of the Australian Government Research Training Program Scholarship. All errors remain our own.

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: This project was supported financially by a URC Small Grant (University of Wollongong (AU)) awarded to Natascha Klocker.

Notes

References

ABS (2011a) Australian Standard Classification of Cultural and Ethnic Groups (ASCCEG) Second Edition, Cat. No. 1249.0, Canberra: Australian Bureau of Statistics.

ABS (2011b) Australian Statistical Geography Standard (ASGS): Volume 1 – Main Structure and Capital City Areas, Cat. No. 1270.0.55.001, Canberra: Australian Bureau of Statistics.

ABS (2012) Reflecting a Nation: Stories from the 2011 Census, 2012–2013, Cat. No. 2071.0, Canberra: Australian Bureau of Statistics.

Aspinall

(2003) Who is Asian? A category that remains contested in population and health research. Journal of Public Health 25(2): 91–97.

Biddle N (2013) Indigenous and Non-Indigenous Marriage Partnerships. Paper 15, CAEPR Indigenous Population Project: 2011 Census Papers. Canberra: Centre for Aboriginal Economic Policy Research, Australian National University. Available at: http://caepr.anu.edu.au/sites/default/files/cck_indigenous_outcomes/2013/11/2011CensusPaper_15.pdf.

Blair

Dunn

Kamp

et al. (2017) Challenging Racism Project 2015–16 National Survey Report, Sydney: Western Sydney University.

Bozon

Héran

(1989) Finding a spouse: A survey of how French couples meet. Population 44(1): 91–121.

Caballero

Edwards

Smith

(2008) Cultures of mixing: Understanding partnerships across ethnicity. Twenty-First Century Society: Journal of the Academy of Social Sciences 3(1): 49–63.

Dalmage

(2000) Tripping on the Color Line: Black-White Multiracial Families in a Racially Divided World, New Brunswick, NJ: Rutgers University Press.

10.

Ellis

Holloway

Wright

et al. (2012) Agents of change: Mixed-race households and the dynamics of neighborhood segregation in the United States. Annals of the Association of American Geographers 102(3): 549–570.

11.

Ellis

Wright

Parks

(2006) The immigrant household and spatial assimilation: Partnership, nativity and neighborhood location. Urban Geography 27(1): 1–19.

12.

Farquharson K (2007) Racial categories in three nations: Australia, South Africa and the United States. In: Proceedings of ‘Public sociologies: Lessons and trans-Tasman comparisons', the annual conference of The Australian Sociological Association (TASA), Auckland, New Zealand, 4–7 December.

13.

Feng

van Ham

Boyle

et al. (2014) A longitudinal study of migration propensities for mixed-ethnic unions in England and Wales. Journal of Ethnic and Migration Studies 40(3): 384–403.

14.

Gorman-Murray

Brennan-Horley

(2010) The geography of same-sex families in Australia: Implications for regulatory regimes. Law in Context 28(1): 43–64.

15.

Holloway

Ellis

Wright

et al. (2005) Partnering ‘out’ and fitting in: Residential segregation and the neighbourhood contexts of mixed-race households. Population, Space and Place 11(4): 299–324.

16.

Iceland

Nelson

(2010) The residential segregation of mixed-nativity married couples. Demography 47(4): 869–893.

17.

Jayasuriya L (2002) Understanding Australian Racism. Australian Universities Review 45(1): 40–44.

18.

Johnston

Poulsen

Forrest

(2013) Multiethnic residential areas in a multiethnic country? A decade of major change in England and Wales. Environment and Planning A 45(4): 753–759.

19.

Kalmijn

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

20.

Kalmijn

Flap

(2001) Assortative meeting and mating: Unintended consequences of organized settings for partner choices. Social Forces 79: 1289–1312.

21.

Khoo S (2011) Chapter 6: Intermarriage, integration and multiculturalism: A demographic perspective. In: Clyne M and Jupp J (eds) Multiculturalism and Integration: A Harmonious Relationship. Canberra: ANU E Press, pp.101–119.

22.

Lofquist D, Lugaila T, O’Connell M, et al. (2012) Households and Families: 2010. 2010 Census Briefs. Washington, DC: U.S. Census Bureau. Available at: www.census.gov/prod/cen2010/briefs/c2010br-14.pdf.

23.

Luke

Carrington

(2000) Race matters. Journal of Intercultural Studies 21(1): 5–24.

24.

MacLeod

(2006) Multiethnic Australia: Its History and Future, Jefferson, NC: McFarland & Co.

25.

Mateos

(2014) The international comparability of ethnicity and collective identity: Implications for segregation studies. In: Lloyd

Shuttleworth

Wong

(eds) Socio-Spatial Segregation: Concepts, Processes and Outcomes, Bristol: Policy Press, pp. 163–193.

26.

Mateos

Singleton

Longley

(2009) Uncertainty in the analysis of ethnicity classifications: Issues of extent and aggregation of ethnic groups. Journal of Ethnic and Migration Studies 35(9): 1437–1460.

27.

Office for National Statistics (2014) What Does the 2011 Census Tell Us About Inter-ethnic Relationships? Newport: Office for National Statistics. Available at: www.ons.gov.uk/ons/rel/census/2011-census-analysis/what-does-the-2011-census-tell-us-about-inter-ethnic-relationships/rpt—inter-ethnic-relationships.html.

28.

Pavlovskaya

Bier

(2012) Mapping census data for difference: Towards the heterogeneous geographies of Arab American communities of the New York metropolitan area. Geoforum 43(3): 483–496.

29.

Price

Zubrzycki

(1962) Intermarriage patterns in Australia. Population Studies 16(2): 123–133.

30.

Salaita

(2005) Ethnic identity and imperative patriotism: Arab Americans before and after 9/11. College Literature 32(2): 146–168.

31.

Shih

Sanchez

(2009) When race becomes even more complex: Toward understanding the landscape of multiracial identity and experiences. Journal of Social Issues 65(1): 1–11.

32.

Smith

Edwards

Caballero

(2011) The geographies of mixed-ethnicity families. Environment and Planning A 43(6): 1455–1476.

33.

Stillwell

(2010) Ethnic population concentration and Net Migration in London. Environment and Planning A 42(6): 1439–1456.

34.

Tindale A, Klocker N and Gibson C (2014) Inter-ethnic Partnerships: Remaking Urban Ethnic Diversity. Urban Policy and Research 32(4): 393–415.

35.

Twine

(1999) Transracial mothering and antiracism: The case of white birth mothers of “black” children in Britain. Feminist Studies 25(3): 729–746.

36.

Walker

Heard

(2014) Interethnic partnering: Patterns by birthplace, ancestry and indigenous status. In: Heard

Arunachalam

(eds) Family Formation in 21st Century Australia, Dordrecht: Springer, pp. 53–75.

37.

White

Sassler

(2000) Judging not only by color: Ethnicity, nativity and neighbourhood attainment. Social Science Quarterly 81(4): 997–1013.

38.

Wright

Ellis

Holloway

(2011) Where black-white couples live. Urban Geography 32(1): 1–22.

39.

Wright

Holloway

Ellis

(2013) Gender and the neighborhood location of mixed-race couples. Demography 50(2): 393–420.

40.

Wright

Houston

Ellis

et al. (2003) Crossing racial lines: Geographies of mixed-race partnering and multiraciality in the United States. Progress in Human Geography 27(4): 457–474.