Introduction
Although research on well-being based on life satisfaction/happiness/quality of life has become a burgeoning area, refugees have only appeared as a target population of well-being research since the 1990s (Colic-Peisker, Reference Colic-Peisker2009). It was also much later in 2015, that the United Nations promoted well-being for all as a key Sustainable Development Goal, and as part of this inclusive call, there is a need for analysis on refugees’ well-being. By the end of 2019, the world saw a record 26.4 million refugees as part of the 82.4 million individuals who were forcibly displaced worldwide as a result of persecution, conflict, violence or human rights violations (UNHCR, 2020). As well-being is said to predict life expectancy related to survival probabilities (Banks et al., Reference Banks, Nazroo and Steptoe2012; Danner et al., Reference Danner, Snowdon and Friesen2001), it is important to focus on how satisfied refugees are with life in the host country. In addition, how well refugees have settled down in the host country has always been a key research focus. The objective of this paper is to examine if (and what) factors associated with life satisfaction (LS) and settlement experience (SE) are different and this is enriched by longitudinal analysis with a focus on gender and age, using Australia as a case study.
Australia is the only country with an established practice of detaining asylum seekers and immigrants without visas including children in detention centres onshore and offshore for an indefinite period of time since 1992 (Mares, Reference Mares2016). This practice has come under great scrutiny and criticism from UNHCR (Hedrick et al., Reference Hedrick, Armstrong, Coffey and Borschmann2019). At the same time, Australia has a long tradition since World War II of providing permanent protection and resettlement to refugees. Australia formally participates in the UNHCR resettlement programme accepting 12 000 to 20 000 refugees a year after their asylum claims have been processed overseas.Footnote 1 The high number of refugees resettling in Australia (180 790 entrants between 2009 and 2018 according to Refugee Council of Australia, 2019) highlights the importance of looking into their well-being and ensuring they are adequately resettled.
Settlement of refugees in the form of how well they have integrated in the society and the factors affecting various aspects of integration has been widely studied using qualitative and quantitative analyses. Mulvey (Reference Mulvey2017) is a qualitative study on the Scottish and UK governments views of social rights and how they apply to refugees. Most of the quantitative studies – with the exception of Beiser (Reference Beiser2006) and Cheung and Phillimore (Reference Cheung and Phillimore2017), Mitton (Reference Mitton2021), and Lichtenstein and Puma (Reference Lichtenstein and Puma2019) – have been static analyses of settlement using cross sectional data. Settlement is, however, an ongoing process of integration, culminating in a sense of belonging that can change over time (Marlowe, Reference Marlowe2019). Thus one-off survey data do not do justice to understanding settlement which is a moving sequence of challenges and adaptive responses to integration, whose importance is conditioned by time (Beiser, Reference Beiser2006). Hence, longitudinal data are especially valuable when studying refugee settlement by means of integration (ibid).
With studies on refugee well-being, here too, most studies such as Beiser and Hou (Reference Beiser and Hou2017) and Çetin (Reference Çetin2019) have examined the effect of pre- and post-migration factors on well-being using cross sectional data. It is important to move away from such snapshot analysis and examine if there has been a change in well-being of refugees over time. To our knowledge, the study of Haj-Younes et al. (Reference Haj-Younes, Stromme, Igland, Kumar, Abildsnes, Hasha and Diaz2020) on Syrian refugees in Norway is the only longitudinal analysis on refugee well-being. Others longitudinal studies such as Ajdukovic and Ajdukovic (Reference Ajdukovic and Ajdukovic1998) focused on the impact of displacement on the well-being of Croatian refugee children while Mitton (Reference Mitton2021) examined the effect of post-migration life experience factors, demographic characteristics and government asylum policy on the housing outcomes of refugees in England.
By using a longitudinal study to trace the same refugees through three time periods from 2013/14 to 2017/18, this paper contributes to the existing literature in four ways. First, it addresses the dearth of studies on intertemporal analysis. Second, the study examines if factors affecting well-being and SE are different from one another and if they have changed over time. In doing so, unlike previous studies which have examined various aspects of integration separately (such as employment, housing etc.) for settlement, here, self-reported SE of refugees is used as a single outcome measure.
With well-being too, unlike previous studies on refugees such as Ikizler and Szymanski (Reference Ikizler and Szymanski2018) and Kim (Reference Kim2016) which consider factors affecting depression often measured by the post-traumatic stress disorder (PTSD) scores, this study’s focus is on self-reported LS. There are far more studies on PTSD than well-being in the literature (Khawaja and Hebbani, Reference Khawaja and Hebbani2019) possibly because refugees are known to be traumatised when they seek refuge and hence depression could be considered a bigger concern. Westerhof and Keyes (Reference Westerhof and Keyes2010) explain that distress (such as PTSD) and well-being measures are two distinct dimensions that are not mutually exclusive or perfectly correlated. Hence factors linked to distress may not necessarily be the same factors linked to flourishing.
The third contribution of the study is that, in addition to using a gender lens to compare the factors that influence male and female adult refugees’ LS and SE, potential non-linear effects of age are also examined for the first time. The focus on gender and age deepens analysis for a more targeted policy regime to provide support to specific refugee communities more effectively. Lastly, the role of psychological capital (if any) – in moderating the effect of discrimination, which is a major concern for refugees (Marlowe, Reference Marlowe2019) – is examined.
The rest of the paper is organised as follows. The next section is a literature review on the studies which have examined SE and LS. Then a description of the data, variables and empirical model is provided. This is followed by a discussion of the results obtained and policy implications are put forward. The conclusion section summarises the key findings of the paper and acknowledges the limitations of this study and suggests future avenues of research.
Literature Review
Settlement Experience
Resettlement refers to the transference of refugees from a state where they have sought protection to a different state in which they have been granted permanent residence status.Footnote 2 It is viewed as an opportunity to rebuild shattered lives and some studies (Pittaway et al., Reference Pittaway, Muli and Shteir2009; Lewis, Reference Lewis2020) note that integration into the host society and having a sense of belonging are core components of resettlement.
Several studies have used the conceptual framework on integration and settlement of Ager and Strang (Reference Ager and Strang2008) in terms of achievement and access to employment, housing, education and health; social connection related to social bonds (family and other members of community), social bridge (with other communities including the host community) and social links (with the structures of the host state); experience of citizenship and rights. Others such as Guo and Guo (Reference Guo and Guo2017) use the broad definition of integration to mean adaptation, adjustment, or acculturation.
Curry et al. (Reference Curry, Smedley and Lenette2018) note that “successful” settlement is generally evaluated by a selection of top-down, predetermined indicators that may miss the challenges and opportunities inherent to refugees’ lived experience. The subjective nature of the term ‘successful’ requires a deeper understanding of what it means to refugees and this can be measured subjectively according to refugees’ self-rated assessment of feelings of integration, belonging and at-homeness (Marlowe, Reference Marlowe2019). In Australia, the discourse surrounding resettlement policies has primarily focused on service performance (Curry et al., Reference Curry, Smedley and Lenette2018) while Fozdar and Banki (Reference Fozdar and Banki2017) focus on the extent to which Australia fulfils its legal obligations for resettled refugees.
Literature (Betts et al., Reference Betts, Sterck, Geervliet, MacPherson, Buith and Armstrong2017; Elliott and Yusuf, Reference Elliott and Yusuf2014; Phillimore, Reference Phillimore2011) has identified that refugees learning to settle in a new country face a range of difficulties and challenges - lingering effects of trauma from fleeing; financial capital lost or left behind and human capital (such as educational credentials) often unrecognised or undervalued; difficulties in getting a job; cultural and language barriers; lack of networks and social ties; discrimination and many other issues. Lamba and Krahn (Reference Lamba and Krahn2003) argue that social capital such as formal and informal social networks including familial and extra-familial ties may be the only useful form of capital available for a refugee on arrival in a new country. The usefulness of social capital revolves around having people to turn to for help, advice, mentoring as well as provide opportunities in various aspects of life including employment.
To date, most studies have used qualitative analysis comprising interviews and focus group discussions to understand how well a particular group of refugees has resettled. For example, Hatoss and Huijser (Reference Hatoss and Huijser2010) examined female Sudanese refugees while Vromans et al. (Reference Vromans, Schweitzer, Farrell, Correa-Velez, Brough, Murray and Lenette2018) interviewed a sample of African and Afghan women in the Australian state of Queensland. Curry et al. (Reference Curry, Smedley and Lenette2018) and Smith et al. (Reference Smith, Hoang, Reynish and Mond2020) focused on refugee resettlement in regional Australia. Slade and Borovnik (Reference Slade and Borovnik2018) focussed on older Bhutanese refugees in New Zealand while Lamba and Krahn (Reference Lamba and Krahn2003) and Guo and Guo (Reference Guo and Guo2017) examined resettlement of refugees in Canada.
Studies using quantitative methods on settlement via integration are very limited (Lichtenstein and Puma, Reference Lichtenstein and Puma2019, Capps et al., Reference Capps, Newland, Fratzke, Groves, Auclair Fix and Mchugh2015) and this is partly due to the need for a large enough sample size on refugee respondents to obtain valid results for interpretation. This exercise needs monetary resources to train interviewers and for data entry, which is costly and time consuming. The limited empirical studies include Lewis (Reference Lewis2020) who used a sample of 244 older Bhutanese refugees in the US to highlight the importance of social capital for their integration. Beiser (Reference Beiser2006) reports on the impact of pre-migration trauma, social resources such as language training, ethnic community and individual coping strategies on temporal reintegration in Canada. Cheung and Phillimore (Reference Cheung and Phillimore2017) investigated factors associated on separate aspects of integration related to health, housing and employment for refugees in the UK comparing cohorts in 2005 and 2007 although they are not necessarily the same people. Lichtenstein and Puma (Reference Lichtenstein and Puma2019) collected data on 467 newly arrived adult refugees from Bhutan and Burma in the US city of Denver from 2011/12 to 2014/15. They found overall integration to increase and noted how some aspects of integration (social capital, language and cultural knowledge, and employment) changed over time.
Well-Being
Similar to studies on non-refugees, different measures of well-being have been used in the research on refugees. For example, Çetin (Reference Çetin2019) defines well-being based on how the individual relates to the self and the environment, measured as a subscale of the Spiritual Well-being scale (see Musa and Pevalin, Reference Musa and Pevalin2012). Haj-Younes et al. (Reference Haj-Younes, Stromme, Igland, Kumar, Abildsnes, Hasha and Diaz2020) examines four separate domains (physical health, psychological health, social relationships and environment) of the World Health Organisation’s quality of life (QoL) assessment while Greene (Reference Greene2019) focused on the psychological QoL as a sum of three items in that same questionnaire. Others such as Colic-Peisker (Reference Colic-Peisker2009), Khawaja and Hebbani (Reference Khawaja and Hebbani2019) and Newman et al. (Reference Newman, Nielsen, Smyth and Hirst2017) used LS as a single measure of well-being. Beiser and Hou (Reference Beiser and Hou2017) used positive mental health (that considered their mental health to be excellent, …. poor) and the meta-analysis by Yoon et al. (Reference Yoon, Chang, Kim and Gomes2013) examined self-esteem, LS, hope and positive affect as positive indicators of mental health.
In terms of analyses of the factors associated with well-being, several studies (Khawaja and Hebbani, Reference Khawaja and Hebbani2019; Marlowe, Reference Marlowe2019) have noted that they are similar to those related to settlement experience and integration. For instance, pre-migration traumatic experiences negatively impact the LS of refugees (Choi et al., Reference Choi, Lim, Jun, Lee, Yoo, Kim, Lee and Kim2017) while Pittaway et al. (Reference Pittaway, Muli and Shteir2009) note that those who spoke English prior to arrival and whose professional qualifications had been recognized in Australia had settled most successfully and had a higher level of LS.
Colic-Peisker (Reference Colic-Peisker2009) found health, employment status, job and financial satisfaction, social support, Australian networks, acculturation (difficulty of understanding Australian way of life) and adaptation to Australian way of life to be correlated with LS. Khawaja and Hebbani (Reference Khawaja and Hebbani2019) found acculturation (measured as a scale of 11 items) was not associated with LS while Beiser and Hou (Reference Beiser and Hou2017) found enculturation (sense of continued belonging to the source country) lowered refugee well-being. Greene (Reference Greene2019) notes that kinship ties and not friendship ties were related to psychological QoL for refugees in Great Lakes district in the USA.
Beiser and Hou (Reference Beiser and Hou2017) found perceived discrimination disadvantaged men but not women’s positive mental health. While the well-being of male refugees was found to be higher than females (Brand et al., Reference Brand, Loh and Guilfoyle2014, Schubert and Punamäki, Reference Schubert and Punamäki2011), Khawaja and Hebbani (Reference Khawaja and Hebbani2019) did not find any significant difference in the LS between the genders.
More recently, psychological capital in the form of resilience emerged as the most significant factor associated with LS of refugees (Khawaja and Hebbani, Reference Khawaja and Hebbani2019). Newman et al. (Reference Newman, Nielsen, Smyth and Hirst2017) constructs psychological capital using the dimensions of hope, resilience, optimism and self-efficacy. The Newman study found that psychological capital fully mediates the relationship between organizational support and well-being but only partially mediates the relationship between perceived family support and well-being for refugees.
Data, variables and model
Data on adult refugees (aged 18 years and above) is drawn from a population-based cohort representative sample of refugees with permanent humanitarian visa from the Building a New Life in Australia (BNLA) (2017) data set (see Edwards et al., Reference Edwards, Smart, De Maio, Silbert and Jenkinson2018 for details). This is the first on-going, largest and most up to date Australia-wide survey data on humanitarian migrants collected by the Australian Institute of Family Studies, a government statutory agency in the Department of Social Services.
The BNLA survey covers about 33% of the humanitarian migrants who arrived in Australia during the sample period and they are followed up annually for a 5 year-period (ibid). This survey has a high retention rate of about 84% at the 12 months’ follow-up (ibid). Another strength of this survey is that it collects data on a wide range of aspects, including mental health and well-being, settlement success and a range of other socio-economic and demographic characteristics. These attributes make BNLA a rich and valuable dataset to investigate the settlement success and well-being of refugee migrants in Australia.
The initial wave was conducted in 2013/2014 (wave 1) with about 75% of the refugees having lived in Australia for less than six months. To date, there are five waves of data. The initial wave and the alternating waves of 2015/16 and 2017/18 took place during home visits where data were collected via face-to-face interviews while waves 2 and 4 were undertaken using telephone interviews. For this study, to be consistent with the mode of data collection, we used data from waves 1, 3 and 5 as they were all collected by the same means of home visits and we obtained a balanced sample of 1058 respondents over the three waves who answered all the questions used in the model for estimation.
Variables
SE is measured by the Likert-scale response to, ‘Overall, your experience of settling in Australia so far has been 1. Very hard; 2. Hard; 3. Good; 4. Very good’. LS is measured by a scale from 0 (completely dissatisfied) to 10 (completely satisfied) for the question of, ‘Thinking about your own life and personal circumstances, how satisfied are you with your life as a whole?’ Apart from socio demographic variables such as age, gender, education, employment status, marital status, various forms of stress (such as housing, financial and language barrier), experience of discrimination, a host of social capital and acculturation variables and psychological capital were used.
Psychological capital was measured as an index using the following four questions with a Likert scale of 1 to 4: i) how much do you agree or disagree that you are optimistic about the future ii) how true is it that if you are in trouble, you can think of a good solution iii) how true is it that you are certain that you can accomplish your goals iv) how true is it that you can handle whatever comes your way. The Bartlett’s test for factor analysis rejected the null hypothesis that the four variables are orthogonal (χ2=3334, p=0.001) and hence data is appropriate factor analysis. The Keiser-Meyer-Olkin value of 0.755 indicated that the sample is adequate and 75.5% variance in the data can be explained by the four factors. The Cronbach’s Alpha statistic of 0.788 also passed the test for internal consistency.
Empirical Models
To compare the factors associated with SE and LS in waves 1 and 5, probit models were first estimated as the independent variables are Likert scale based. Then to exploit the longitudinal nature of the data, panel estimation using the three waves of data was undertaken. For this, we use the random effects ordered probit model rather than the fixed effects model for two reasons. First, although authors have proposed different types of econometric techniques to estimate fixed effects ordered logit models, there is no consensus in the literature on which of these estimators is to be preferred, and on their consistency (Baetschmann et al., Reference Baetschmann, Staub and Winkelmann2011). Second, unlike the fixed effects, the random effects model provides estimates on individual heterogeneity related to socio demographic variables.
To capture the moderating effect of psychological capital, the discrimination variable is interacted with the psychological capital variable to see if this lowers the effect of discrimination. This form of testing in a regression model has been widely used in the literature (Chen et al., Reference Chen, Hall, Ling and Renzaho2017; Pieterse et al., Reference Pieterse, Knippenberg, Schippers and Stam2010). Lastly, two different types of survey weights defined in the dataset were used for the estimation. For the estimations based on wave 1 and 5 (that is, individual years), population survey weights were available to adjust sample estimates to achieve population totals for all participants. For the panel estimation, longitudinal survey weights were used to adjust for attrition between waves and the recruited sample.
Results and discussion
Table 1 provides summary statistics for all the variables used. The ratio of unemployed among refugees fell from 93.47% in wave 1 to about 70% in wave 5 and the ratio of those with language barrier stress was almost halved over this period. The mean age of refugees was between 36 and 40 years for the three waves of data. It should be noted that neither SE or LS shows a statistically significant improvement from wave 1 to 5.
Summary Statistics
Table 2 presents the estimates for the SE and LS in waves 1 and 5 for comparison. Unlike some studies’ finding of the adverse effect of trauma where it has been argued to persist for years even when hostilities have ended (Boor et al., Reference Boor, Amos, Nevitt, Dowrick and White2020), here there is no evidence of trauma having any influence on SE or LS in wave 1 or 5. However, we find support for our result from Porter and Haslam (Reference Porter and Haslam2005) who use a meta-analysis of studies from 1959 to 2002 to highlight that postdisplacement factors (such as life circumstances and what opportunities are present) are more closely related with mental health than predisplacement factors related to previous trauma.
Ordered Probit Estimations for Waves 1 and 5
One reason for our particular result is that the refugees in our sample have permanent humanitarian visas. This reflects a degree of stability which may have enabled them to feel distant from acute refugee trauma. Another reason could be the sharp contrast between their previous conditions when they were fleeing and their present relatively peaceful life in Australia. This may have resulted in an appreciation for being resettled and satisfied with a return to some level of normality.
A number of similarities and differences emerge with factors associated with SE and LF and across time. With SE, males are relatively worse off than females in waves 1 and 5 but with LS, while there was no gender difference in wave 1, males are, however, less satisfied than females in wave 5. There was no difference in age effects on SE in either wave, while there appears to be a non-linear age effect on LS in waves 1 and 5. But these effects are very different when longitudinal data are used (see Tables 3 and 4), cautioning the use of cross sectional data for analysis (based on results in Table 2 per se).
Panel Probit Estimates for Settlement Experience
Note:
* represents the turning point in the quadratic U-shaped relationship of age which was identified to occur at age 38.
Panel Probit Estimates for Life Satisfaction
Relative to those unemployed, refugees with a job did not feel any more settled or satisfied in wave 1 or 5. This result, however, masks important aspects related to job type such as full/part time and unskilled/skilled employment which can be expected to matter. Unfortunately, the BNLA dataset does not have more details on employment to examine these.
Language is not a significant stress factor to SE or LS in either wave. This could be due to the support (in terms of interpretation and English classes) provided to refugees and also potentially due to other more pressing concerns over language competency when they first arrive. For instance, housing and financial stress are concerns in wave 1. But with time, in wave 5, housing is likely to have been sorted out (hence is not statistically significant) but financial stress remains a significant concern. Factors that positively affect SE and LS in both waves are one’s overall good health status, a sense of belonging and the friendliness and trust in the neighbourhood.
Discrimination which is seen not to affect SE or LF in wave 1 is, however, negatively associated with SE and LS in some years to come. This is possibly due to the fact that in the earlier years of arrival, refugees may have not been sufficiently exposed to the Australian society (less interaction in the wider community) to experience discrimination or it could be that there are other major worries such as housing and financial stress that refugees are trying to deal with when they first move to Australia and therefore do not perceive discrimination to be serious.
As Table 2 represents snapshots of wave 1 and 5, we now turn to more robust results using longitudinal data. Table 3 on SE also accounts for heterogeneity in terms of gender and age as these are significant in the full sample seen in the first column of Table 3. Females compared to males settle down better and this could be due to different primary concerns that the genders may have. For instance, females are more vulnerable as they may be subject to more abuse, violence and rape than males in their journey across (Capps et al., Reference Capps, Newland, Fratzke, Groves, Auclair Fix and Mchugh2015). Thus if feeling safe and more secure than the previous conditions is uppermost in the females’ mind when they were fleeing, then being Australia which is relatively safer and secure could provide a feeling of settlement for females compared to males. It could also be that females have more domestic concerns in relation to whether their family members are safe and secure and have food and clothing etc. as home is where family is may be the notion they subscribe to. Males on the other hand are likely to be more concerned about getting a good job, providing for their family and having social recognition and networks.
With regards to age, there is a U-shaped curve depicting the relationship with SE in Table 3 that was not revealed in Table 2. This highlights the importance of using longitudinal data which is more reliable than cross-sectional data analysis. It was further identified that 38 years of age was the turning point – that is, SE declines with age until 38 years after which SE increases with age. It could be that maturity in age and life experience enable the older cohort to settle down better. Pittaway et al. (Reference Pittaway, Muli and Shteir2009), however, warns that it is possible that the feeling of settlement may come at the cost of lowered expectations for the future. This was discussed in terms of losing hope of achieving satisfactory employment, or the realization that they would probably not achieve home ownership, and of losing hope of acceptance by the mainstream Australian community. Unfortunately, there was no further data available to detect if this is true.
Marital status given by ‘never married’ showed a negative association only for the older cohort in terms of SE. As framed by the social capital theory, marriage is one form of bonding social capital (Putnam, Reference Putnam2000) which extends networks via their spouse or their children. In addition, for the older cohort (and not the other subgroups in Table 3), language barrier is not a significant stress and this is the only group which appears to benefit from ethnic support. It is possible that the older cohort is able to retain and benefit from cultural socialisation and thereby draw on social capital with this group. Similar to females, the older cohort benefits from being proficient in English. Acculturation in the form of being able to understand Australian ways is also another significant factor in improving SE for males and the older refugees.
To settle better, it is important for male, female and the younger cohort of refugees to overcome both financial and language barrier stress. With language, around 75% of refugees reported that they did not understand English well or at all before arriving in Australia (Jenkinson et al., Reference Jenkinson, Slibert, Maio and Edwards2016). Evidence shows that settlement is helped by having friends in Australia by way of an informal social network for support. Having a friendly neighbourhood is another significant form of developing a social bond for refugees in making them feel welcome and included in the community. Females in particular settle down better knowing that they can trust the community they are in, possibly reflecting a sense of insecurity and vulnerability they bring with them when they flee their country. SE has, however, not improved for any subgroup over the years while discrimination significantly hinders SE for all subgroups apart from the older cohort.
Table 4 presents results on LS using longitudinal data. The first column in Table 4 shows that there is a difference in gender and hence separate estimations are undertaken for males and females. The first column in Table 4 shows that female compared to male refugees are less satisfied. Brand et al. (Reference Brand, Loh and Guilfoyle2014) explain that women can encounter more traumatic challenges in the migration and resettlement process and thus their life satisfaction may be lower than that of men. Age, however, did not affect LS indicating that there was no difference in the LS experienced by the younger and older cohorts. Females who were without a partner were, however, worse off than those who were in a relationship while unmarried males were less satisfied than their married counterparts.
Housing and financial stress adversely affect male but not female LS. It could be that most females (about 74%) are with their parents or are married and therefore males as bread winners shoulder more of the responsibility in finding appropriate housing and fending for the family. Jenkinson et al. (Reference Jenkinson, Slibert, Maio and Edwards2016) notes that refugees in Australia have to move house frequently due to problems finding secure housing. Their study notes that refugees from wave 1 said it was hard or very hard to find housing, most commonly because of housing costs (57%), language difficulties (50%) and lack of rental references (49%). It could also be the case that those who gain refugee status inevitably settle into areas with low-quality housing (Phillips, Reference Phillips2006) and are therefore not satisfied.
Language barrier adversely affects female and not male LS. Data from BNLA shows that males on average are better educated than females and have higher English language competency based on statistical differences between the genders. Weaker fluency in English relative to males makes females more dependent, less likely to have frequent contact with friends in the wider community and being less self-sufficient could affect LS.
Ethnic support is negatively associated with males and not female LS. Ethnic support can be a double-edged sword and evidence on its effect is mixed. For instance, while there is research showing that ethnolinguistic similarity is important for the formation of social networks and security and is thus beneficial (Algan et al., Reference Algan, Hémet and Laitin2016; Martén et al., Reference Martén, Hainmueller and Hangartner2019), others (Campion, Reference Campion2018; Cooper et al., Reference Cooper, Enticott, Shawyer and Meadows2019) have shown that this is not necessarily true as it can act as a barrier to appropriate assimilation and reinforce expectations of behaviour which may not be conducive.
A friendly and trustworthy neighbourhood is conducive for male but trustworthy neighbourhood has no effect on female LS. This reflects that the surrounding environment around one’s house may be considered a safe and secure place providing some form of protection for one’s family. As mentioned earlier, such aspects related to finding accommodation may be considered more of male’s prerogative in terms of responsibility than the concern of female refugees. Similar factors that positively affect both male and female LS include a sense of belonging and overall health status while discrimination is negatively associated with LS for both genders. According to the Social Identity Theory, being categorised as a group member (in this case, refugees) may trigger some degree of intergroup differentiation and discrimination, which results in decreased psychological well-being (Hogg, Reference Hogg2016).
We now examine if psychological capital can moderate the adverse effects of discrimination. The results obtained are consistent – in that, there is no moderating role for any of the subgroups for SE or LS. Table 5 shows the estimations for just the SE of females and LS for males due to space constraint. It can be seen that psychological capital has a significant effect on females’ SE and males’ LS but the adverse effect of discrimination is not seen to decrease and neither is the interaction term significant. Thus having personal psychological resources and personal resolve does not mean that refugees are able to draw on it to negate the effects of discrimination on SE or LS. In addition, this could depend on the type of discrimination experienced or whether it is systemic discrimination that pervades workplace, job opportunities, access to housing etc. However, readers are cautioned that the result obtained here is suggestive and not conclusive as it is likely that multidirectional causality issues related to the impact of SE and LS on psychological capital can exist or psychological capital can affect the other factors associated with LS and SE. This needs further robust empirical examination and can be undertaken with a lag for psychological capital if longitudinal data on psychological capital was available. This is unfortunately not the case here as the question on psychological capital was only included for wave 5.
Examining Moderating Effect of Psychological Capital in Wave 5
Conclusion
When refugees flee to survive, it is important to examine how well they go beyond survival and settle in a new country and enjoy a sense of well-being in the society they live in. This study provides a deeper understanding of the factors associated with these two aspects using longitudinal data. This analysis is vital to ensure that refugees retake control of their lives to regain their dignity and their freedom and contribute to the richness of the social, cultural, and economic fabric of the host country.
Evidence shows that although females fare better than males in their SE, they are less satisfied with their life than males. To improve SE for females and the older adult cohort, greater emphasis on language competency and understanding Australian ways of life can be effective using special tailored programs for these groups or assigning mentors to help navigate the new culture. With males, ethnic support, housing and financial stress were associated with lower levels of LS while language barrier was a concern for female LS as language skills can help build linguistic and social capital.
Discrimination was negatively associated with both SE and LS for most subgroups and personal psychological resource was not found to be useful in dealing with discrimination. This is an area that requires more focus as one limitation of this study was the inability to distinguish between types of discrimination. For instance, discrimination in the labour market or street discrimination where people are exposed in informal situations in public places by strangers (e.g. Muslim women observed by security personnel in shopping areas) are different. So is ethnic and religious discrimination. Perhaps this is best examined using Phillimore (Reference Phillimore2020)’s notion of the receiving society. For instance, how do the locals perceive refugees and the diversity brought about by refugees in their neighbourhoods.
Several other limitations in this study also need to be acknowledged that may serve to provide direction for future research. As secondary data was used, only factors associated with SE and LS could be identified but the underlying mechanisms for the association could not be examined. Thus there is a need for a mixed methods approach using both quantitative and qualitative evidence based on interviews, and focus group discussions will provide a more holistic analysis. This could also involve other stakeholders. For example, in 2017, the Australian government set up the Community Support Program to enable refugees to be resettled with support from individuals, community groups or business. It may inform policy to devise effective support services, if how these groups view refugees’ integration and problems is understood.
While it is noteworthy that there has been an improvement in the types of questions asked in the BNLA, not all questions could be used for longitudinal analysis. For instance, information on optimism for constructing the psychological capital index was only available in wave 5 but not in wave 1 and 3 while information on how welcome respondents felt was available in waves 1 and 3 but not 5. The question on access to health services was only available for wave 1.
Given the importance of social and psychological capital for well-being, it will be useful to differentiate between the various types of social and psychological capital, and examine whether their effects on well-being change over time reflecting that they are processes rather than static functions. Lastly, the above analysis was not undertaken for different race, religion, ethnicity or country of origin as this is presently beyond the scope of this study given the length of the current paper. Future research accounting for such heterogeneity in the sample may be important in highlighting further differences for policy consideration.
Competing interests
The authors declare none.
Open access
