Getting There? The Effects of Functional Factors, Time and Place on the Social Integration of Migrants

Abstract

A survey of 1,400 migrants, including many asylum seekers and refugees, living in deprived areas in Glasgow UK is used to test hypotheses in the literature about the effects of functional factors (educational qualifications, ability to speak English, employment), time and place upon the social integration of migrants. Three aspects are considered: trust, reliance and safety; social relations; sense of community. Overall, social integration indicators were worse for migrants than for British citizens living in the same places. Functional factors were positively associated with different aspects of social integration: higher education with more neighbourly behaviours; employment with better social relations and belonging; and English language with greater reliance on others and available social support. Time was positively associated with most social integration indicators; time in the local area more so than time in the UK. Living in a regeneration area was negatively associated with many aspects of social integration. The findings raise questions about the doubly negative effects of the use of dispersal policy for asylum seekers to regeneration areas, necessitating secondary relocation of migrants through further, forced onward migration.

Introduction

There has been much recent interest in the notion of integration, prompted by ‘new immigration’ (Robinson 2010). The UK Home Office has published several documents on what integration could mean, for refugees particularly (Home Office 2005). Integration is a policy goal but also a ‘chaotic’ and ‘contested’ concept (Ager and Strang 2008; Robinson 1998); one that can be defined according to different actors’ interests and values (Castles et al. 2002). Here, we review academic perspectives on the issue to outline some key distinctions within, and influences upon, integration.

Commentators have seen integration as either personal or public, involving structures or processes, and as a means or an end; all have related integration to the question of how migrants become part of the society in which they find themselves. For Da Lomba (2010) integration has both public and private dimensions, the former comprising the legal and social environments, and the latter concerned with the personal experiences and social connections of migrants. Legal status is seen as something which shapes the public dimension, though in our view it may also influence the private dimension as well. Similar distinctions are made by Ager and Strang (2004) in describing the state of being integrated as involving three things: ‘public outcomes’ related to employment, housing, education and health; social connections with members of their own and other communities; and personal competencies in language, cultural knowledge and security/stability. They also make a reference to the issue of status in saying that integration involves ‘shared notions of nationhood and citizenship’, which implies both membership and identification with the country in which one is residing. The end state of integration has been described by the Home Office as involving migrants being participative in three ways: achieving their potential; contributing to the community; and accessing services (Home Office 2005). The contribution principle has been further emphasised by the UK Government in the idea that refugees can help earn their citizenship through volunteering (Phillimore 2012).

Cheung and Phillimore (2013) describe two kinds of integration – social and structural – but the relationship between the two is unclear and possibly two-way. They highlight ‘the importance of social capital and social networks…in integration theory’ (p.2), and that we do not know enough about how social networks help or hinder access to employment for migrants (refugees in their case). Employment is also said to be the most important aspect of structural integration because it ‘may facilitate access to new social networks, increase prospects for learning English, and provide opportunities to regain confidence and economic independence’; this is the official UK Government approach of ‘paid work as the path to social inclusion’ (p.4). Thus, functional factors, e.g. employment, education, health and housing, are considered as ends in themselves (or ‘markers of integration’), and as means towards integration (Ager and Strang 2004), or as the precursors to integration (Fyvie et al 2003).

Phillimore (2012) defines three key themes: the development of a sense of belonging to the host community; the development of social relationships and social networks; and, the development of the means and confidence to exercise rights to resources such as education, work and housing. These key elements have echoes of the three definitions of the word ‘integration’ offered by the OED, namely, ‘making something whole, by the harmonious combination of different elements, and the bringing of groups into equal membership of a society’ (www.oed.com).

Social integration may differ between migrant groups and places - regions and neighbourhoods. Migrants of different legal status may vary in the degree to which they feel legitimate, accepted, or in their confidence to engage with other people to develop familiarity and social relationships. The development of social connections also depends on the services and social environment in a neighbourhood: people need places to meet and a safe public realm: individuals need a sense of security and stability to be integrated (Ager and Strang 2004). We hypothesise that this will depend on the individual’s legal status and place of residence (temporary or permanent), with this combination reflected in the role of choices and entitlements. For example, refugee integration is said to be different to that of economic migrants ‘because they are forced to migrate, and are less able to choose where to live on the basis of the availability of social networks’ (Cheung and Phillimore 2013, p.5). Harmonious, integrated communities are said to rely upon refugees having the same rights and entitlements as others, because this affects how other people view them, and the respect afforded them (Ager and Strang 2008).

There is also debate about the role of time, wherein integration is a form of acculturation. Rather than comprising assimilation (migrants acquiring the host culture), acculturation is a learning process of second-culture acquisition by both groups (Berry 1997; Rudmin 2009). Integration is most often assumed to be a linear process that progresses over time, but this notion is contested by those who see integration as a negotiation between old and new identities and locations (Phillimore 2012; Bhatia and Ram 2009), or as interrupted or impeded by events (Atfield et al 2007).

This paper focuses on social integration within communities (places of residence), and the role of functional factors such as employment, education and language as potential facilitators. We consider three components of social integration: trust, reliance and safety; social relations; sense of community. These elements are akin to Phillimore’s (2012) first two components of integration, belonging and social relationships, and cover three of the five domains of neighbourhood social cohesion: social order and social control; social networks and social capital; and place attachment and identity (Forrest and Kearns (2001). We also have a particular interest in the effects of time and place upon social integration, for reasons related to the study area. Like other UK cities, Glasgow has received many ‘new migrants’ over the past decade alongside more traditional migrant groups, and we have explored whether the process of integration over time differs between migrant groups.

The Glasgow Context: Rapid Social Change in a Deprived City

Glasgow is a very deprived city that had until recently been losing population. By 2012, the city’s population stood at 595,080, having regained half the losses of the previous decade (NRS 2013). Concurrently, in 2012, 42% of Glasgow’s 694 spatial data zones (statistical units) fell within the 15 percent most deprived in Scotland, a drop from 54% a decade earlier (Scottish Government 2012), although this was still the largest share of any district in the country. One of the main drivers of these changes has been migration. Over the inter-census period 2001 to 2011, the ethnic minority population in Glasgow more than doubled, from 42,000 (7.2% of the total city population) to 92,000 (15.4%). Of the city’s 56 planning neighbourhoods, eleven had an ethnic minority population of 12 percent or more by 2010 (Freeke 2012a). Thus, a predominantly white city has changed, due to three policy thrusts at European, national and local levels. After EU enlargement in 2004/2007, migrants from Central and Eastern Europe came to Glasgow, as to other UK cities, mostly living in the private rented sector which doubled in size over this period, to 18.7% of dwellings in 2011 (Freeke 2012b).

At national level, The Scottish Government has, pursued what has been termed a ‘pro-migration strategy’ to attract high-skilled and entrepreneurial individuals to Scotland, either when applying for UK work permits, or after studying at a Scottish University (Scottish Executive 2004). The aims have been to counter population decline and boost economic growth (Mooney and Williams 2006). Migrants were generally more skilled than out-migrants, but self-employment and entrepreneurial activity were the main benefits of the ‘Fresh Talent’ initiative in the long-term (Houston et al 2008). However, two key challenges were identified. The initiative only applied to a select group of high-status migrants and left other groups untouched, including Africans already in the country and EU migrants in low skilled jobs (Williams and De Lima 2006). The second challenge was that Scotland’s ‘history of welcoming migrants is not a happy one’, and a call by the First Minister that ‘Scotland must be seen as a safe, welcoming place, wherever new residents come from’ (CRE 2005 and McConnell 2003, both cited in Williams and De Lima 2006).

At the city level, Glasgow City Council has, since 2000, been receiving UK asylum seekers and is the largest ‘dispersal site’ for asylum seekers outside London: the first contract from 2000-06 involved 2,500 units of social housing accommodation, and the second contract from 2006-11 was for 5,800 bed spaces. It was recorded that 6,000 asylum seekers were present in the city in 2003, and 2,800 in 2010 with the annual rate of arrival dropping to 1,300 in 2008/9 and again to 700 in 2009/10 (Freeke 2012b). The majority of asylum seekers are from African and Middle Eastern countries, with 40% coming from five countries: Iran, Pakistan, Iraq, Somalia and Democratic Republic of Congo (ODS 2007). It is hard to estimate how many refugees granted leave to remain live in the city. The rate of acceptance of asylum cases has fluctuated over time, estimated at around a third in 2005, but possibly higher in Glasgow’s case due to high numbers (>80%) of families with children (ODS 2007). Second, upon receiving leave to remain, migrants can choose to live anywhere in the UK, although social landlords have agreed to treat refugees as unintentionally homeless and to offer them permanent accommodation, thus encouraging many to stay in the same locations (Binns 2002; Mignard 2011).

The impacts of asylum seekers and refugees upon receiving communities can be great, with 90 percent of asylum seekers accommodated in just half a dozen locations around the city. The available accommodation tended to be in low demand, high-rise estates so as not to impact upon the housing waiting list for locals (Crawford et al 2012). Several of these areas were later included as regeneration areas in a strategy running from 2005 onwards, involving wholesale demolition and redevelopment of the estates over a period of a decade (GHA 2006; GCC 2007). These regeneration plans have over-run due to the economic downturn so that by the start of 2014, the regeneration programmes are nowhere near completion.

Accommodating large numbers of asylum seekers in weakening, unprepared, communities was not easy. In the first year of the resettlement programme, City Council housing services recorded 107 incidents of racial harassment, 96 involving an asylum seeker, including 35 in a single estate (Binns 2002). There were responses to assist the integration of asylum seekers and refugees who remained in the city. The City Council established a Refugee Support Team, the Scottish Government developed a Refugee Integration Action Plan (Scottish Government 2003), and the Scottish Refugee Council helped to set up Refugee Integration Networks in key parts of the city. The Networks developed activities over time as needs changed, including: reception services and humanitarian support; orientation towards public services; intercultural contact activities; emergency legal and other support; and ongoing support for the destitute. Funding for integration activities totalled nearly £1m per annum from 2002/3 to 2007/8 (ODS 2007), but was severely cut around 2010/11.

Research Aims and Objectives

We explore the effects of time and place upon self-reported indicators of social integration for migrants in Glasgow, taking into account other influential functional factors. We address the following four sets of questions:

The Effects of Functional Factors:

• Are different forms of functional integration, namely educational qualifications, ability to speak English, and employment, positively associated with self-reported indicators of social integration for migrants?

The Effects of Time:

• Is duration of stay in the UK and/or the area of residence associated with improvements in self-reported indicators of social integration?

The Effects of Place:

• What effect does living in a regeneration area have upon indicators of social integration for migrants?

• Are any effects of duration of stay different between migrants living in regeneration areas versus those living in other places?

The Effects of Migrant Group:

• Are there any differences in the effects of functional factors, time or place between migrant groups?

• For refugees, do the effects of time differ for the period before and after the granting of leave to remain?

Methods

Data Source

The data are from two surveys carried out across fifteen communities in Glasgow in summer 2008 and 2011 as part of the GoWell study of the impacts of housing investment and regeneration (Egan et al. 2010). The communities all fall within the 15% most deprived in Scotland, the target group for many area-based policy interventions (Walsh 2008). A random sample of addresses was selected in nine of the areas; in the other six, where redevelopment was underway, all addresses were selected for the survey. The survey asked householders about their home, neighbourhood, community, and physical and mental health. All respondents were asked to describe their current legal status as residents; the current analyses are based on those who were not UK-born British Citizens. Results from the two surveys were similar and thus we present results from both samples combined. Dual respondents were included only once; sensitivity analyses confirmed that the choice of survey for these cases did not affect the results.

Social integration Variables

We selected variables from the survey related to three aspects of social integration: trust, reliance and safety; social relations; and sense of community. The questions used pre-set response categories, typically with four or five responses; these variables were dichotomised for analyses - see Table 3 for the specific classifications used.

Table 3.

A Comparison of Social Integration Indicators for Migrants and British-Born Citizens

Positive vs. negative responses		Migrants
					N=1326-13581
British Born Citizens N=5720-57831	P2
			% giving positive responses
Trust, Reliance & Safety: Neighbours look out for each other		Great deal/fair amount vs. Not much/at all	42.9	72.3	<0.001
Neighbours are honest		Agree vs. Disagree/don’t know	11.3	21.1	<0.001
Neighbours relied on for control		Agree vs. Disagree/don’t know	20.3	41.9	<0.001
Feel safe after dark		Safe vs. Unsafe/don’t walk alone/don’t know	35.3	46.5	<0.001
Antisocial behaviours		None vs. Any slight/serious problems	53.5	59.0	<0.001
Social Relations: Visit neighbours in their homes		Great deal/fair amount vs. Not much/at all	26.3	44.3	<0.001
Exchange with neighbours		Great deal/fair amount vs. Not much/at all	18.2	32.1	<0.001
Stop and talk to people in nhd.		Great deal/fair amount vs. Not much/at all	47.2	74.6	<0.001
Know people in the nhd.		Most/many vs. Some/few/none	12.3	45.4	<0.001
Meet up with relatives		Weekly or more vs. Less than weekly	37.1	67.1	<0.001
Meet up with friends		Weekly or more vs. Less than weekly	65.2	70.6	<0.001
Meet up with neighbours		Weekly or more vs. Less than weekly	56.9	79.4	<0.001
Available practical support		One or more vs. None/wouldn’t ask/don’t know	64.8	78.2	<0.001
Available financial support		One or more vs. None/wouldn’t ask/don’t know	47.2	56.3	<0.001
Available emotional support		One or more vs. None/wouldn’t ask/don’t know	62.6	74.2	<0.001
Use social amenities in last week		One or more vs. None	51.9	47.6	0.004
Use local social amenities		One or more vs. None	43.5	40.0	0.02
Sense of Community: Enjoy living here		Great deal/fair amount vs. Not much/at all	73.7	85.5	<0.001
Belong to the neighbourhood		Great deal/fair amount vs. Not much/at all	56.3	80.9	<0.001
Feel part of the community		Great deal/fair amount vs. Not much/at all	48.9	76.7	<0.001
Satisfied with Neighbourhood		Very/fairly satisfied vs. Dissatisfied/neither	66.9	78.3	<0.001

¹N varies due to missing values in the social integration variables

²p for heterogeneity.

In relation to trust and reliance, respondents were asked to what extent they felt that ‘My neighbourhood is a place where neighbours look out for each other’, and to what extent they agreed with two statements: ‘It is likely that someone would intervene if a group of youths were harassing someone in the local area’; and, ‘Someone who lost a purse or wallet around here would be likely to have it returned without anything missing’. We also included two measures of perceived safety: ‘How safe would you feel walking alone in this neighbourhood after dark?’; and whether any of the following three things were problems in the neighbourhood: ‘violence, including assaults and muggings’, ‘people being insulted, pestered or intimidated in the street’, and ‘people being attacked or harassed because of their skin colour or ethnic origin’.

Four aspects of social relations were examined. For neighbourliness, respondents were asked to what extent they did three things: ‘visit neighbours in their home’; ‘borrow things and exchange favours with my neighbours’; and, ‘stop and talk to people in my neighbourhood’. Respondents were also asked to select ‘how many of the people in [their] neighbourhood’ they knew. For social contacts, respondents were asked ‘not counting the people you live with, how often do you do any of the following’: ‘meet up with relatives’; meet up with friends’; and ‘speak to neighbours’. For social support, respondents were asked ‘not counting those you live with, can you tell me around how many people could you ask for the following kinds of help?’: ‘to go to the shops for you if you are unwell’; ‘to lend you money to see you through the next few days’; and ‘to give you advice and support in a crisis’. Lastly, respondents were asked whether they had used the following amenities in the last seven days: ‘sports facilities, swimming pool or gym’; ‘social venues (e.g. pub, bingo, bowling, dancing, social club)’; ‘park or play area (e.g. multi-purpose games courts, including basketball and football)’; ‘library’; and ‘community centre’, and also whether use was ‘within’ or ‘outside my local area’. We derived two measures: use of any of the social amenities, irrespective of location; and, use of any of the amenities locally.

Four measures of sense of community were used. Respondents were asked ‘to what extent do the following apply to you: I enjoy living here; I feel I belong to this neighbourhood; and, I feel part of the community’. Lastly, respondents were asked ‘how satisfied or dissatisfied are you with this neighbourhood as a place to live’.

Potential Explanatory Variables

We investigated the effects of three functional factors, or facilitators or markers of integration (Ager and Strang 2004): educational qualifications (any versus none); employment status (divided into those with a full- or part-time job versus those without); and difficulty speaking English (any versus none, based on an interviewer recorded measure).

Two measures of time were considered. Respondents were asked for the month and year of their arrival in the UK, from which we calculated the duration of stay in the UK in years up to the date of interview. Respondents were also asked how long they had lived in their current area in year bands (e.g. under 1, 1-2, 3-5 years etc.). In addition, for migrants who had received leave to remain, the month and year of decision was recorded, from which we calculated time in the UK before and after achieving refugee status for this group. For the effects of place, we used a location variable that divided migrants into those living in one of the city’s regeneration areas, and those living elsewhere.

Finally respondents were classified into four migrant groups: British citizens born outside the UK; Asylum Seekers (those had applied for asylum and were either awaiting a decision or appealing a decision); Refugees (those who had been given leave to remain); and Other Migrants (including EU passport holders and those who described their status as ‘other’).

Analysis

We initially compared rates of positive responses to the 21 social integration indicators between migrants and British-born citizens in the survey, to establish whether self-reported social integration is lower for migrants. We then explored predictors of positive social integration indicators in the four migrant groups. Preliminary analyses considered each migrant group separately and formal tests of effect modification/statistical interaction were conducted. There was no consistent evidence to suggest that associations differed between the migrant groups and, for simplicity, we present results based on all four groups combined. Similarly, there was no evidence to suggest that the impact of time varied according to residence in a regeneration or other area and, again, we present results for both areas combined.

Logistic regression was used to explore the impact of functional factors on social integration, adjusted for migrant group. Analyses also controlled for three socio-demographic measures: gender; age group (16-24, 25-9, or 40+); and whether or not there were dependent children aged under 16 in the household. We present odds ratios (OR) and 95% confidence intervals (CI) for positive social integration responses in respondents with any (versus no) educational qualification, in employment (versus not), and with no (versus any) difficulties speaking English. The impacts of time and place were also explored using logistic regression, adjusting for migrant group, age, sex, dependent children, education, employment and difficulties with English. Having established that associations with time were approximately linear, we present ORs (95% CI) for the impact on positive social integration of each additional year in the UK, each additional year in the area, and of living in a regeneration area. Finally, among refugees, we examined the effects of time waiting for a leave to remain decision and time spent after that decision, again having confirmed that associations were approximately linear. Results are presented unadjusted and adjusted for time in the UK, time in the area, and living in a regeneration area. Additional adjustment for demographic and functional factors had no impact.

Results

The survey achieved response rates of 47.5% (2008) and 45.4% (2011), yielding 4,709 and 4,063 completed questionnaires. From these, we identified 1,493 unique respondents who were not born in the UK. Of these, 1,358 (91%) had complete data and are included in the analyses.

Migrant Characteristics

Migrants in the sample had a variety of ethnicities and came from a wide range of places of origin (Table 1). The majority of migrants were of Asian or Black ethnicity, although the balance between these two varied across the four migrant groupings, with Asians being the larger group amongst British citizens born abroad, whilst Blacks were the larger among Refugees. One-in-six asylum seekers were of Chinese ethnicity, and two-in-five Other Migrants were of white ethnicity. The dominant continent of origin differs between the migrant groups: the largest number of British citizens born abroad and of asylum seekers came from Asia, whilst for refugees it was Africa, and in the case of Other Migrants, nearly a third came from Europe.These findings appear to reflect what is known about the city’s changing population (Freeke 2013c). A total of 73 countries of origin were reported by respondents; a survey only of refugees in 2011 found a total of 37 countries of origin (Mulvey 2011).

Table 1. Ethnic and Geographical Origin of Migrants.

	British Non-UK born (N=425)	Asylum seeker (N=304)	Refugee (N=368)	Other migrant (N=261)
Ethnicity
White	65 (15.3)	14 (  4.6)	38 (10.3)	101 (38.7)
Mixed	11 (  2.6)	6 (  2.0)	5 (  1.4)	3 (  1.1)
Asian	129 (30.4)	88 (28.9)	95 (25.8)	30 (11.5)
Black	118 (27.8)	89 (29.3)	143 (38.9)	70 (26.8)
Chinese	31 (  7.3)	47 (15.5)	32 (  8.7)	25 (  9.6)
Other	64 (15.1)	58 (19.0)	54 (14.7)	31 (11.9)
Unknown	7 (  1.6)	2 (  0.7)	1 (  0.3)	1 (  0.4)
Country of birth
Unknown	224 (52.7)	118 (38.8)	232 (63.0)	125 (47.9)
Africa	65 (15.3)	57 (18.8)	68 (18.5)	29 (11.1)
Asia	113 (26.6)	126 (41.4)	58 (15.8)	29 (11.1)
Europe	20 (  4.7)	3 (  1.0)	8 (  2.2)	78 (29.9)
North America	1 (  0.2)	0 (  0.0)	2 (  0.5)	0 (  0.0)
Oceania	2 (  0.5)	0 (  0.0)	0 (  0.0)	0 (  0.0)

Table 2 presents the characteristics of the four migrant groups. The largest group of migrants in the study was British citizens born abroad (N=425, 31.3%). The next largest groups were refugees (368, 27.1%, and asylum seekers (304, 22.4%), and the smallest group was Other Migrants (261, 19.2%), which includes EU citizens, other economic migrants, and students. British citizens born abroad had been in the UK the longest (on average 12.4 years), followed by refugees (6.3 years), Other Migrants (5.4) and Asylum Seekers (4.3). British citizens born abroad tended to be older than the other migrant groups. Asylum seekers were the most likely to be female, to have no qualifications and to be not-working. Refugees were more likely to have qualifications and to be in employment than asylum seekers, but slightly less so than British citizens born abroad. Other Migrants were most likely to be male, to have educational qualifications, to be in employment, and to be without dependent children. Around a third of all migrant groups had difficulty speaking English. Finally, 80 to 90% of refugees and asylum seekers lived in regeneration areas, compared with around 70% of the other two migrant groups.

Table 2.

Duration of Stay, Socio-demographic and Other Control Measures, by Migrant Group

		Measure	British Non- UK Born N=425
Asylum Seeker N=304
Refugee
N=3681
Other Migrant N=261	p2
			Mean (SD)
Years in UK		12.4 (13.3)	4.3 (3.1)	6.3 (4.1)	5.4 (6.7)	<0.001
Years in area		4.8 (  5.1)	1.5 (1.5)	3.1 (2.4)	2.2 (2.7)	<0.001
Years in UK with Leave to Remain				3.2 (4.2)
Years in UK without Leave to Remain				3.4 (2.9)
			N (Col. %)
Age Group
16-24
25-39
40+ Gender
Male
Female
Dependent children
None
One or more
Educational qualification
None
Any
Employment status Not working Working
Difficulty with English No difficulty Difficulty
Location
49 (11.5)
197 (46.4)
179 (42.1)
187 (44.0)
238 (56.0)
170 (40.0)
255 (60.0)
191 (44.9)
234 (55.1)
283 (66.6)
142 (33.4)
286 (67.3)
139 (32.7)
54 (17.8)
198 (65.1)
52 (17.1)
107 (35.2)
197 (64.8)
110 (36.2)
194 (63.8)
233 (76.6)
71 (23.4)
296 (97.4)
8 (  2.6)
197 (64.8)
107 (35.2)
47 (12.8)
198 (53.8)
123 (33.4)
159 (43.2)
209 (56.8)
136 (37.0)
232 (63.0)
193 (52.5)
175 (47.6)
268 (72.8)
100 (27.2)
244 (66.3)
124 (33.7)
46 (17.6)
154 (59.0)
61 (23.4)
133 (51.0)
128 (49.0)
156 (59.8)
105 (40.2)
105 (40.2)
156 (59.8)
126 (48.3)
135 (51.7)
182 (69.7)
79 (30.3)
<0.001
0.002
<0.001
<0.001
<0.001
0.65
Regeneration Area		307 (72.2)	283 (93.1)	305 (82.9)	178 (68.2)	<0.001
Other Areas		118 (27.8)	21 (6.9)	63 (17.1)	83 (31.8)

¹N=320 for time since LTR and time in UK without LTR

²p for heterogeneity

Social Integration for Migrants Compared with British-Born Citizens

Migrants as a whole were less positive than British-born citizens about social integration (Table 3). With regard to trust, reliance and safety, the differences were greatest in relation to issues of trust in others, with British-born citizens twice as likely as migrants to feel their neighbours were honest (21% versus 11%) and could be relied upon to exercise informal control (42% versus 20%).

With respect to social relations, the differences were greatest in relation to indicators of neighbourliness, where British-born citizens were at least one-and-a-half times more likely than migrants to have social relations with their neighbours, and over three times more likely to know people in the neighbourhood. British-born citizens were also more likely to meet up regularly with relatives and neighbours, although there was very little difference in rates of contact with friends, possibly reflecting contacts made amongst migrants themselves as a group. British-born citizens were around a fifth more likely to have available means of social support. Migrants were somewhat more likely to use social amenities than British-born citizens.

With respect to sense of community, British-born citizens were one-and-a-half times more likely than migrants to feel they belonged to the neighbourhood (81% versus 56%) and were part of the community (77% versus 49%). British-born citizens were also more likely than migrants to enjoy and be satisfied with their area.

Differences in Social Integration between Migrant Groups

From a cohesion perspective, Other Migrants, and to a slightly lesser extent British citizens born abroad, tended to have the more positive views of their neighbours and of safety in the neighbourhood (Table 4). Asylum seekers were both the least likely to feel safe after dark, and the least likely to identify antisocial behaviour problems.

Table 4.

A Comparison of Social Integration Indicators by Migrant Group

	British Non-UK Born N=414-4251	Asylum Seeker N=293-3041	Refugee N=364-3681	Other Migrant N=255-2611	P2
	% giving positive responses
Trust, Reliance & Safety:
Neighbours look out for each other	45.9	39.9	37.7	47.5	0.06
Neighbours are honest	10.6	8.2	10.3	17.6	0.003
Neighbours relied on for control	24.0	17.8	17.1	21.5	0.06
Feel safe after dark	35.1	32.3	33.7	40.6	0.19
No antisocial behaviours	46.4	65.8	47.6	59.4	<0.001
Social Relations:
Visit neighbours in their homes	29.3	23.4	27.5	24.9	0.43
Exchange with neighbours	20.1	13.5	16.1	23.4	0.01
Stop and talk to people in nhd.	50.0	39.5	47.3	51.7	0.01
Know people in the nhd.	15.1	10.9	9.8	12.7	0.12
Meet up with relatives weekly	42.0	31.7	34.8	38.9	0.03
Meet up with friends weekly	62.1	61.7	66.3	72.8	0.02
Meet up with neighbours weekly	56.6	54.3	57.9	59.0	0.69
Available practical support	68.2	63.5	62.0	64.8	0.29
Available financial support	45.9	45.5	47.4	51.2	0.52
Available emotional support	65.6	62.1	59.0	63.6	0.28
Use social amenities in last week	49.2	42.4	59.0	57.5	<0.001
Use local social amenities	41.9	36.8	49.7	45.2	0.01
Sense of Community:
Enjoy living here	68.6	79.0	67.4	84.6	<0.001
Belong to the neighbourhood	59.1	55.8	50.8	60.0	0.07
Feel part of the community	52.6	43.6	47.3	51.3	0.08
Satisfied with Neighbourhood	66.6	69.1	63.9	69.0	0.44

¹N varies due to missing values in the social integration variables

²p for heterogeneity.

There were no notable differences between migrant groups in their frequency of visits to neighbours’ homes, or in their levels of acquaintance with people in the area. However, British citizens born-abroad and Other Migrants were somewhat more likely than asylum seekers and refugees to exchange things with neighbours and to stop and talk to people in the neighbourhood. Other Migrants and Refugees were also more likely than others to meet up with friends on a regular (weekly) basis, and also the most likely to use social amenities. There were no marked differences between migrants groups on indicators of available social support.

Other Migrants and asylum seekers were more likely to say they enjoyed living in the area than the other two groups. Feelings of inclusion were perhaps more common among Other Migrants and British citizens born abroad, than among asylum seekers and refugees. There were no marked differences between migrant groups in rates of neighbourhood satisfaction.

The Effects of Functional Factors

The majority of associations with functional factors were consistent with better social integration for migrants who had educational qualifications, were in employment, and who had no difficulties speaking English, although some associations were more marked than others, and there were also a few exceptions to this pattern (Table 5).

Table 5.

Associations of Functional Factors with Social Integration Indicators for All Migrants¹

Trust, Reliance & Safety:
Any educational qualifications N=1326-13582
In employment
N=1326-13582
OR (95% CI)
No difficulties speaking English N=1326-13582
Neighbours look out	1.16 (0.93, 1.46)	1.13 (0.87, 1.46)	0.86 (0.69, 1.08)
Neighbours are honest	1.28 (0.90, 1.82)	1.41 (0.97, 2.06)	1.24 (0.86, 1.78)
Neighbours relied on for control	1.26 (0.96, 1.66)	1.12 (0.82, 1.52)	1.42 (1.07, 1.90)*
Feel safe after dark	0.98 (0.78, 1.24)	1.53 (1.18, 1.99)**	1.12 (0.88, 1.41)
No antisocial behaviours	0.53 (0.42, 0.67)***	1.00 (0.78, 1.29)	0.72 (0.58, 0.91)**
Social Relations:
Visit neighbours	1.37 (1.07, 1.76)*	1.18 (0.89, 1.56)	0.91 (0.71, 1.17)
Exchange with neighbours	1.47 (1.10, 1.96)**	1.26 (0.92, 1.72)	0.92 (0.69, 1.22)
Stop and talk to people in nhd.	1.35 (1.08, 1.69)**	1.20 (0.93, 1.54)	1.01 (0.80, 1.26)
Know people in the nhd.	1.07 (0.76, 1.50)	1.04 (0.71, 1.52)	1.09 (0.77, 1.54)
Meet up with relatives	0.82 (0.65, 1.03)	1.50 (1.15, 1.94)**	0.90 (0.71, 1.13)
Meet up with friends	1.35 (1.07, 1.71)*	1.36 (1.04, 1.79)*	1.00 (0.79, 1.26)
Meet up with neighbours	0.98 (0.79, 1.23)	1.19 (0.92, 1.54)	1.15 (0.92, 1.44)
Available practical support	1.36 (1.08, 1.72)*	1.15 (0.88, 1.51)	1.33 (1.06, 1.68)*
Available financial support	1.20 (0.96, 1.50)	1.27 (0.98, 1.63)	1.20 (0.96, 1.51)
Available emotional support	1.10 (0.88, 1.39)	1.29 (0.99, 1.68)	1.03 (0.82, 1.30)
Use of social amenities	2.02 (1.61, 2.54)***	1.13 (0.88, 1.46)	1.01 (0.81, 1.27)
Use of local social amenities	1.63 (1.30, 2.04)***	1.24 (0.96, 1.60)	0.84 (0.67, 1.06)
Sense of Community:
Enjoy living here	0.78 (0.61, 1.01)	1.11 (0.83, 1.48)	0.81 (0.63, 1.06)
Belong to the neighbourhood	0.91 (0.73, 1.14)	1.30 (1.00, 1.69)*	1.05 (0.84, 1.32)
Feel part of the community	1.01 (0.81, 1.27)	1.30 (1.01, 1.68)*	1.09 (0.87, 1.36)
Neighbourhood satisfaction	0.86 (0.68, 1.09)	1.04 (0.79, 1.36)	0.68 (0.54, 0.89)**

¹Adjusted for migrant group

²N varies due to missing values in the social integration variables

*p<0.05; **p<0.01; ***p<0.001

With regard to indicators of trust, reliance and safety, migrants who could speak English without difficulty were forty percent more likely than others to consider that their neighbours could be relied upon to exercise informal social control (p<0.05) and those in employment were 50% more likely to feel safe after dark (p<0.01). Conversely, having educational qualifications and the ability to speak English were both associated with a higher likelihood of identifying antisocial behaviour problems (p<0.01), which might reflect a greater awareness of those issues within the community. The remaining associations were weaker although generally consistent with more positive outcomes in those with qualifications, who were working and who had better English; the exception being that those with better English were less likely to report that their neighbours looked out for each other. There were no associations between education and feeling safe after dark, or with employment and reporting anti-social behaviours.

For social relations, migrants with educational qualifications were around a third or more likely than others to engage in neighbourly behaviours (p<0.05), to meet up with friends regularly (p<0.05), to have available practical support (p<0.05), and up to twice as likely to make use of social amenities (p<0.001). Those in employment were a third to a half times more likely to meet up regularly with relatives and friends than those without jobs (p<0.05). Those who could speak English were more likely to report having available social support, in particular a third more likely to have practical support (p<0.05). The remaining associations were weaker. Associations with employment were consistent with a positive impact on social relations, other than knowing people in the neighbourhood where there was no association. Associations with education were more mixed: individuals with an education qualification tended to be more positive overall, although they were less likely to meet up with relatives and there was little evidence of any association with knowing people in the neighbourhood or meeting up with neighbours. The majority of remaining associations with English language suggested no substantial impact on social relations other than a slight increase in meeting up with neighbours and financial support, and a small decrease in use of local social amenities.

The strongest associations found with indicators of sense of community were that migrants in employment were thirty percent more likely than others to feel belonging in the neighbourhood and inclusion in the community (p<0.05), and, to a lesser extent, to enjoy living in their neighbourhood. Individuals with no difficulty speaking English were less satisfied with their neighbourhood (p<0.01) and were less likely to report enjoying living there. Associations with education were weak and consistent with either no impact or a slightly negative effect on sense of community.

The Effects of Time

Table 6 (first 2 columns) shows that positive social integration in all three aspects (trust, reliance and safety; social relations; sense of community) generally increased with increasing time in the UK and, more markedly, with time in the area.

Table 6.

Associations of Duration of Stay and Place of Residence with Social Integration

	Indicators for All Migrants1
Trust, Reliance & Safety:
Additional year in UK N=1326-13582
Additional year in area N=1326-13582
OR (95% CI)
Regeneration area N=1326-13582
Neighbours look out	1.03 (1.02, 1.05)***	1.05 (1.02, 1.09)**	0.61 (0.46, 0.80)***
Neighbours are honest	1.01 (0.99, 1.03)	1.02 (0.97, 1.07)	0.51 (0.35, 0.74)***
Neighbours relied on for control	1.02 (1.01, 1.04)**	1.05 (1.01, 1.09)**	0.57 (0.42, 0.78)***
Feel safe after dark	1.00 (0.99, 1.02)	0.99 (0.96, 1.03)	0.51 (0.38, 0.68)***
No antisocial behaviours	0.99 (0.98, 1.01)	0.97 (0.94, 1.01)	0.50 (0.38, 0.67)***
Social Relations:
Visit neighbours	1.01 (0.99, 1.02)	1.02 (0.99, 1.06)	1.09 (0.80, 1.48)
Exchange with neighbours	1.01 (1.00, 1.03)	1.03 (0.99, 1.07)	0.68 (0.49, 0.94)*
Stop and talk to people in nhd.	1.03 (1.02, 1.05)***	1.07 (1.03, 1.11)***	0.83 (0.63, 1.09)
Know people in the nhd.	1.05 (1.03, 1.06)***	1.13 (1.08, 1.17)***	0.71 (0.48, 1.04)
Meet up with relatives	1.01 (1.00, 1.02)	1.06 (1.02, 1.09)**	1.04 (0.78, 1.37)
Meet up with friends	1.00 (0.99, 1.01)	1.03 (0.99, 1.06)	0.91 (0.68, 1.22)
Meet up with neighbours	1.02 (1.00, 1.03)*	1.05 (1.01, 1.08)**	0.77 (0.58, 1.01)
Available practical support	1.01 (1.00, 1.03)	1.08 (1.03, 1.12)***	0.98 (0.74, 1.31)
Available financial support	0.99 (0.97, 1.00)*	1.04 (1.00, 1.07)*	1.48 (1.12, 1.95)**
Available emotional support	1.00 (0.99, 1.02)	1.04 (1.00, 1.08)*	0.99 (0.75, 1.31)
Use of social amenities	1.01 (0.99, 1.02)	1.02 (0.98, 1.05)	1.08 (0.82, 1.43)
Use of local social amenities	1.01 (1.00, 1.02)	1.02 (0.99, 1.06)	0.99 (0.75, 1.31)
Sense of Community:
Enjoy living here	1.02 (1.00, 1.04)*	1.03 (0.99, 1.07)	0.47 (0.33, 0.67)***
Belong to the neighbourhood	1.02 (1.01, 1.04)**	1.06 (1.02, 1.10)**	0.78 (0.59, 1.03)
Feel part of the community	1.03 (1.01, 1.04)**	1.06 (1.02, 1.10)***	0.78 (0.59, 1.02)
Neighbourhood satisfaction	1.01 (1.00, 1.03)	1.01 (0.97, 1.04)	0.46 (0.33, 0.63)***

¹Adjusted for migrant group, age, sex, dependent children, education, employment and difficulties with English

²N varies due to missing values in the social integration variables

*p<0.05; **p<0.01; ***p<0.001

For each additional year in the UK, migrants were 3% more likely to feel that their neighbours looked out for each other (p<0.001) and 2% more likely to feel that their neighbours could be relied upon to intervene to prevent harassment (p<0.01). The equivalent impact for each additional year living in the area was 5% (p<0.01) in both cases. Thus, the likelihood of a positive response on both cohesion measures increases by over a quarter (28%) over five years. Time in the UK and in the area had no marked impact on perceptions of neighbours’ honesty or feelings of safety walking after dark.There was a weak suggestion that those who had lived in the area for longer were more likely to report anti-social behaviours.

The strongest associations of time with aspects of social relations were with neighbourliness, social contacts, and social support. For each additional year in the area, respondents were 7% and 13% more likely to talk to and know their neighbours respectively and similar increases of 3% and 5% were observed for each additional year in the UK (p<0.001 for all). The likelihood of meeting up regularly with relatives and neighbours increased with increasing years in the area (by 6% and 5% for each additional year respectively (p<0.01)) and, to a lesser extent, in the UK (1% and 2% respectively (p<0.05). The impact of time in the area on aspects of social support was positive, with each additional year associated with an increase in the availability of support of between 4 to 8% (p<0.05). On the contrary, for each additional year in the UK, the availability of financial support fell slightly, by 1% (p<0.005). Other associations with time in the UK were negligible while those for time in the area were generally positive, albeit weakly so.

Time in UK and area had a similarly positive impact upon feelings of belonging, inclusion, and, to a lesser extent, enjoyment, but little impact upon neighbourhood satisfaction. Each additional year spent in the UK increased the likelihood of migrants feeling a sense of belonging and inclusion by 2 and 3 percent respectively (p<0.01). Each additional year spent living in the area increased the likelihood of both positive indicators by 6 percent (p<0.01). Thus, the likelihood of migrants having a sense of community increased by a third (34%) after five years in the area.

Finally, we examined the effects of time spent before and after a leave-to-remain decision for refugees. Only the social support outcomes were associated with the two refugee-time measures, and thus we present ORs (95% CI) for these variables (Table 7). Each additional year spent waiting for a decision reduced the likelihood that a migrant would have available social support by 12-14 percent (p<0.01), whilst each year spent after obtaining leave-to-remain increased the likelihood of having each of the three kinds of social support by 14-17 percent (p<0.01). This suggests that, once refugees have permission to stay and some certainty of status, all three kinds of social support rapidly improve for them.

Table 7.

Associations of Time with Social Support Indicators for Refugees

Additional Year Prior to Leave to Remain
Additional Year Since Leave to Remain
Unadjusted
N=1326-13582
Adjusted1 N=1326-13582
Unadjusted N=1326-13582
Adjusted1 N=1326-13582
	OR (95% CI)		OR (95% CI)
Practical support	0.93 (0.86,1.00)	0.88 (0.80,0.97)**	1.10 (1.02,1.19)*	1.14 (1.03,1.26)**
Financial support	0.88 (0.82,0.96)**	0.87 (0.80,0.95)**	1.02 (0.97,1.08)	1.15 (1.05,1.25)**
Emotional support	0.90 (0.83,0.97)**	0.86 (0.78,0.94)**	1.10 (1.02,1.19)*	1.17 (1.06,1.29)**

¹Adjusted for years in UK, years in area, and regeneration area

²N varies due to missing values in the social integration variables

*p<0.05; **p<0.01; ***p<0.001

The Effects of Place

Living in a regeneration area, as many migrants did, had predominantly negative effects upon reports of social integration (Table 6, last column). Three main areas of impact were apparent. First, migrants in regeneration areas were half as likely as those living elsewhere to give positive responses to items concerned with safety and reliance on others (p<0.001 for all associations). Adjustment for time in the UK and in the area had no impact on these associations (not shown but available on request). Second, migrants living in regeneration areas were a third less likely than others to say that they engaged in neighbourly exchanges (p<0.05); again, this association was independent of time spent in the UK or area. Third, migrants living in regeneration areas were around half as likely as those elsewhere to report enjoyment and satisfaction with their area (p<0.001). Other associations suggested no impacts or weak negative effects of living in a regeneration area, the latter for example in relation to talking to, knowing and meeting up with neighbours, and belonging to the neighbourhood/feeling part of the comminuty. The only apparent benefit for migrants of living in regeneration areas was a higher likelihood of having available financial social support, by around 50% (p<0.01). This may reflect mutual support between the many migrants in these areas as well as the operation of support projects for migrants, as reported earlier.

Discussion

We have shown that, even after having lived in the UK for between four and twelve years on average (depending on the migrant group concerned), migrants exhibit much lower levels of social integration than British born citizens. Migrants are less likely to be familiar with and trusting of their neighbours, less likely to engage in neighbour relations, and less likely to feel a part of the neighbourhood and community. The extent to which asylum seekers and refugees, feel part of their local community (at 44% and 47% respectively) is very similar to the 41% found by Mulvey (2011) in a survey of the same groups. These findings temper reports that migrants receive a much more positive reception in Scotland than elsewhere in the UK (Pillai et al 2007), but are in accord with other findings that perceptions of community cohesion in circumstances of diversity are lower in more deprived areas (Aspinall and Watters 2010).

For supposedly migrant-friendly destinations (McCollum et al 2014), the findings suggest caution in concluding too much from past integration success, such that ‘Glasgow is increasingly being recognised as a cluster area where reception and resettlement has worked relatively well, despite initial teething problems’ (Wren 2007, p.409). Our later findings indicate that although integration projects have been running for over a decade, initially in response to racial tensions and migrant destitution, the challenges to be addressed may have changed but not disappeared. Integration projects are still needed, perhaps more for community cohesion and development reasons, rather than primarily to ease tensions and provide humanitarian support.

In accord with integration models (Ager and Strang 2008), our analysis confirms that all three functional factors of education, English language and employment have a generally positive association with social integration. Educational qualifications were particularly associated with greater use of local amenities and higher rates of neighbourly behaviours by migrants. Employment enabled migrants to meet up with friends and relatives (possibly due to available resources), helped migrants feel a sense of belonging and enhanced their sense of safety in the local area. English language results were mixed but suggested that greater competency was associated with higher trust in neighbours, and greater availability of practical forms of social support. This confirms what refugees have said about the positive effect of language on other aspects of their integration, and supports concerns expressed about delays accessing language classes (Mulvey 2013). The findings support those from a much smaller qualitative study among refugees in Glasgow, which found that employment provided the opportunity to learn English, and the absence of employment, and especially the exclusion from paid employment for asylum seekers, was isolating (Sim 2009).

Improvements in integration over time have been described for refugees in the UK in respect of language and employment (Cebulla et al 2010). We have gone further and demonstrated the generally positive effects of time on a wider range of integration measures, using multivariable models controlling for socio-demographic characteristics and the effects of functional factors, said to be necessary to better understand integration indicators (Phillimore and Goodson 2008). We confirm that social integration is positively associated with time in the UK and time in the area of residence, at least in the short to medium term, and that this was true for most aspects of trust reliance and safety, social relations, and sense of community. The positive impacts of functional factors may be a contributory factor, for UK Census 2011 findings indicate that some functional factors improve with time in the UK (ONS 2014). English language proficiency is higher for migrants who have been in the UK longer, particularly European and Chinese migrants, but not those from South Asia (e.g. India, Bangladesh). Employment also rises with length of residence in the UK for non-EU migrants. Educational qualification may play less of a role though, as the Census findings indicate a negative association between time in the UK and educational attainment.

We also found that the effects of time on social integration were consistent across migrant groups. Although ‘the liminal times of migration’ involve indecision, confusion and uncertainty (Cwerner 2001, p.27), neither this liminal reality nor the ‘cultural determination of temporal experience’ (ibid. p.31) prevent the majority of migrants from enjoying better social integration over time. This improvement with time reflects not only the attitudes, efforts, abilities and achievements of migrants, but also the attitudes, behaviours and response of the receiving communities, since integration is a two-way process (Strang and Ager 2010; ECRE 1999). The only status-specific effects of time were in relation to refugees. Our findings that the availability of social support to refugees decreases with time spent awaiting a decision on their asylum application and then increases with time spent after receiving leave to remain provide another reason for ensuring that asylum decisions are made as quickly as possible. However, around a fifth of male asylum seekers in Scotland and half of women wait over two years for their status (Mulvey 2013).

Our findings also contribute to the debate about dispersal policy in the UK, where ‘there have been only a few efforts to assess the impact of the dispersal process on integration and cohesion’ (Aspinall and Watters 2010, p.91). Compulsory dispersal for asylum seekers has been criticised for several ‘failures’ resulting from the use of very deprived communities as housing locations (Stewart 2011) including: local hostility and prejudice (Zetter et al 2002); poor community relations (Dawson 2002); exclusion and isolation of migrants (Spicer 2008); pressures on local voluntary sector services (Griffiths et al 2005); and fears of resentment of specialist services for migrants where local needs have been neglected and services cut (Wren 2007). We have shown that the use of regeneration areas in Glasgow to house asylum seekers and refugees has a large, negative effect upon social integration, including migrants’ feelings of trust reliance and safety, neighbourly behaviours, and neighbourhood satisfaction. This substantive evidence demonstrates what service providers in Glasgow said early in the dispersal programme (2001-2), namely ‘the deprivation of the communities where asylum seekers had been placed…made the processes of integration more difficult’ (Bowes et al 2009). Our findings also concur with Anie et al (2005) that hostility towards migrants is greater in areas with a higher proportion of lower social groups, a higher proportion of asylum seekers in the local population, and a high proportion of vacant housing stock, all of which are true in our study areas.

Despite the negative effects of residence in a regeneration area, we have shown that social integration improves with time in the local area, often more so than with time in the UK, and this positive association with time holds true just as much in the regeneration or dispersal neighbourhoods as elsewhere. These are new findings, supporting arguments that, although it might have been better not to place asylum seekers in very deprived, ‘already fractured communities [which are] not multi-ethnic’ (Wren 2007, p.407), there are grounds to expect that integration may advance as refugees make an economic contribution to deprived communities (Phillimore and Goodson 2006) and as communities become multicultural (Sim and Bowes 2007). The findings exemplify Mulvey’s finding that refugees ‘felt that the unit into which they were integrating was a small spatial one. Integration was seen to happen in neighbourhoods rather than in a nation state…’ (2013, p.12).

Place is important to integration as both destination and intention. It is argued that once migrants (refugees) reach the place they consider to be their destination, and where they want to stay, then they are ‘strongly motivated to contribute’ and ‘integration begins’ (Strang and Ager 2010; Losi and Strang 2008). Our findings on the positive associations of time in the local area upon social integration question the wisdom of requiring migrants to relocate as the regeneration process proceeds across the estates. This could amount to a double disadvantage for migrants, who initially had to cope with difficult social and physical conditions on the estates, and now face moving away from the multi-ethnic communities that have developed and are being broken up by regeneration, which requires a further, forced, relocation. Community connections and the growth of ethnic and refugee groups would, other things being equal, influence onward migration decisions (Stewart 2011). Settlement for migrants requires both certainty of status as well as certainty of location; while migration and asylum policy in the UK makes the former difficult, dispersal combined with regeneration policy makes the latter less attainable.

Limitations and Future Research Needs

Whilst we have used two measures of time, the data are cross-sectional, and thus we cannot draw any direct conclusions about causation. Because of difficulties in tracing migrants, our analyses are based on relatively small numbers (though larger than many other studies of the same groups) and therefore have restricted statistical power. While many of the associations fail to reach conventional levels of “statistical significance”, they are generally consistent across many different aspects of social integration, which adds considerable confidence in our results.

The response rate to the survey is low in absolute terms, but good for surveys in deprived areas, where response is typically lower than elsewhere (Tipping et al 2010), due to hard-to-reach groups and survey fatigue. Our survey in deprived areas may limit the findings’ relevance to migrants with greater resources living in other places. However, a large number of ‘new migrants’ are reported to cluster together in ‘less popular, inner-city neighbourhoods’ (Robinson and Reeve 2006), so that our findings are likely to be generalizable to this diverse group.

There are limitations to some of the indicators used. The survey contains good measures of legal status and time, including time in the area of residence which many other studies do not use. But some of the functional factors used are crudely measured: our measures of education and language lack range to investigate how increasing ability impacts upon social integration. We have identified those places with the worst overall conditions, but we have not been able to measure some important characteristics of these and other places. The influence of neighbourhood physical and institutional characteristics upon migrant social integration merit closer examination in the future.

We have a wide range of outcome variables, but we do not know the true scale and extent of respondents’ social networks, nor the extent to which migrants’ social connections are with other migrants or indigenous residents. The effects of the ethnic and status mixture within migrants’ places of residence and social networks upon their social capital and integration are key issues for further research (Wickes et al 2013).

The sample is not large enough to allow analysis according to geographical or national origin. We have adopted a compromise by grouping migrants according to their status, which also partly overlaps with their route of entry. However, we cannot examine the effects of different migrant origins upon their capability or preparedness for social integration. Migrants from countries which are culturally very different to the UK/Western Europe, or from places of conflict and trauma, may experience more psychological distress and have greater needs for social support (Schweitzer et al 2006), and find it harder to socially integrate in their new situation, but we have not been able to take this into account. Expanding the sample size of studies such as this in future work would enable the effects of geographical origin and prior circumstances upon social integration to be investigated.

Conclusion

Our final reflections are on the relevance of our findings for theoretical understandings of migrant integration. The fact that time was found to have independent effects upon social integration outcomes, controlling for functional factors, raises questions about the relationships between the components of integration contained in conceptual frameworks. It is plausible to think that over time, the elements of integration identified by Phillimore (2012) will improve, including migrants’ familiarity with the people and place around them, their tacit knowledge of cultural norms, and, partly as a result, also their confidence for social engagement. These ‘direct’ effects of time may further lead to indirect effects upon other functional factors, or what Ager and Strang call ‘public outcomes’, such as employment and education. Looking at the role of time may illuminate relationships between the private/personal and public dimensions to integration (Da Lomba 2010). Moreover, these links may operate through the participative elements of integration, particularly accessing services (Home Office 2005), requiring the distinction between the ‘means and the ‘marker’ functions of public outcomes to be better explained (Ager and Strang 2008).

Integration studies anticipate differences according to legal status (e.g. Cheung and Phillimore 2013), but we did not find such differences in how time influences social integration, raising a number of issues. First, the meaning and effect of legal status will vary between contexts. In the UK, asylum seekers are mostly placed amidst an existing community, which brings its own problems, but is unlike other countries where asylum seekers are routinely detained, affecting their mental health and social engagement (Robjant et al 2009). Second, given that legal status can affect structural outcomes, our finding suggests that the relationship between structural and social outcomes may not merely be that the former acts as an enabler for the second, but also that structural outcomes may enhance the personal competencies of migrants and thus affect social outcomes. Third, the status-independent effect of time on social integration may indicate that differences in the legal status of migrants are not as identifiable to co-residents as is often assumed.

The implications of our findings for the role of status differences, along with the finding that time in the local area had a greater effect than time in the UK, are redolent of the fact that integration is a two-way process of adaptation, akin to acculturation (Berry 1997). However, further understanding is required as to the balance of direction involved: whose adaptation and engagement to achieve integration matters most, and in what contexts? Our findings also suggest that theories of integration must consider the balance and relations between societal or national integration on the one hand, and local or community integration on the other.

Overall, we conclude that theories of the integration of migrants need to be more dynamic in order to understand the operational relations between key elements of integration, and be more spatially differentiated and specific.