Australian Institute of Criminology

Skip to content

Experiences of crime in two selected migrant communities

Trends & issues in crime and criminal justice no. 302

Holly Johnson
ISBN 0 642 53888 3 ISSN 0817-8542
Canberra: Australian Institute of Criminology, August 2005

Foreword | Australia's immigration rate is among the highest in the world. Migrants face special challenges integrating into a new country, especially if their language, skin colour, religion or cultural practices set them apart from mainstream society. To assess the experiences of crime among migrants, the Australian component of the 2004 International Crime Victimisation Survey oversampled migrants who were born or whose parents were born in Vietnam or the Middle East. The selected migrant and main community samples reported comparable rates of victimisation overall, but lower rates of personal crime were reported by the migrant sample. However, the migrant samples were more likely to feel that assaults and threats perpetrated against them were racially-motivated, and were more likely to be worried about experiencing a racially-based attack in the future. Higher proportions of migrants, particularly women, also feel unsafe walking alone in the local area after dark.

Toni Makkai
Director

Australia is an immigrant society, yet little is known about migrants' experiences of crime. In 2001, 23 per cent of the Australian population was born overseas (Australian Bureau of Statistics 2004). There have been high rates of migration, particularly from non-English speaking backgrounds, since the Second World War. There is even greater diversity in the ethnic and racial composition of recent migrants. Since 1982/83, migrants from North Africa and the Middle East increased from three to 11 per cent of all migrants, those from Southern and Central Asia increased from three to 11 per cent, while migrants from northwest Europe declined from 39 to 16 per cent of all arrivals. The Australian component of the International Crime Victimisation Survey (ICVS) included some additional questions and a sampling of people with Vietnamese or Middle Eastern backgrounds. The purpose of the sample was to enable a robust analysis of crime victimisation and perceptions of safety amongst two groups of Australians who have migrated in large numbers since the 1970s. For the purposes of this analysis, responses of the two migrant groups are combined.

The ICVS sample

The key results from the Australian component of the 2004 ICVS have already been released (Johnson 2004). Further analysis revealed significant differences in the socio-demographic profiles of the Middle Eastern/Vietnamese sample and the main sample. Middle Eastern/Vietnamese migrants were more likely to be:

  • younger (under 35 years of age);
  • involved in home duties as a main occupation;
  • living in a capital city (94 per cent compared with two-thirds of the main sample);
  • born overseas (83 per cent compared with one in five of the main sample);
  • recent arrivals (18 per cent compared with five per cent of the main sample arrived within the last 10 years); and
  • speaking a language other than English at home (96 per cent compared with 13 per cent of the main sample).

The Middle Eastern/Vietnamese migrants were less likely to report weekly household income in the top two categories. However almost one-third refused to state their income.

Sample and method

The ICVS is coordinated through the United Nations and involves the participation of about 60 countries. Australia has participated in four previous cycles of the survey. In the 2004 ICVS, a random sample of 6000 Australians was selected. As expected from a random community survey, it included 118 migrants with Vietnamese or Middle Eastern background. This community sample was generated through random digit dialling (RDD).

An additional oversample of 1001 persons with Vietnamese or Middle Eastern backgrounds was drawn. However, the RDD method was not cost-effective for oversampling the two migrant groups because of the relatively small number of Vietnamese and Middle Eastern people living in Australia, even in areas with high concentrations. A surname-based approach was therefore used (see Challice & Johnson 2005 for more detail). Respondents self-selected into the survey through the following question:

We are particularly interested in speaking with people who were born overseas. Were you or your parents born in any of the following regions?

  1. Vietnam
  2. Middle East
  3. None of these

If required, the definition of 'Middle East' was given as including: Bahrain, Gaza Strip and West Bank, Israel, Iran, Iraq, Jordan, Kuwait, Lebanon, Oman, Qatar, Saudi Arabia, Syria, Turkey, United Arab Emirates and Yemen (consistent with the ABS classification). However, if someone from North Africa self-identified as Middle Eastern, they were included. Similarly, if someone identified with a relevant culture but were born elsewhere (for example, ethnic Vietnamese born in China) they were screened in. Migrants from Vietnam and the Middle East who were selected by chance in the main community sample were added to the Middle Eastern/Vietnamese sample which raised the total to 1119.

There was a slight under-representation of males, single people and employed people in the main sample when matched with the ABS benchmarks. However, the sample matched well in terms of age, Indigenous status and language spoken at home. There was an under-representation of young people and larger households from the two selected migrant groups. To correct for this, weights were applied to ensure that both samples represented the age, gender and place of birth of the Australian population according to the 2001 census of population and housing. For the purpose of this analysis, the Middle Eastern/Vietnamese sample was re-weighted to the effective sample size.

The response rate for the main sample was 53 per cent. With respect to the migrant sample, the response rate was 46 per cent overall, 75 per cent for Vietnamese persons and 36 per cent for persons from the Middle East. The difference in response rates for the two migrant groups was partly due to the fact that Middle Easterners were more likely to be suspicious about being contacted, having had very little experience with survey research (Challice & Johnson 2005).

Data are not presented where cell sizes were less than five or the relative standard errors greater than 25 percent.

Victimisation in migrant communities

Overseas research presents a conflicting picture of the experience of crime victimisation for migrants. In Canada, migrants who were also 'visible minorities' reported significantly lower rates of personal victimisation compared with the rest of the population, while controlling for known risk factors (Brzozowski & Mihorean 2002). 'Visible minority' is a broad term defined as persons, other than Indigenous people, who are non-Caucasian in race or non-white in colour. The British Crime Survey finds ethnic minorities (Blacks, Indians, Pakistanis and Bangladeshis) at lower risk of personal crime overall, but at higher risk of mugging, and at higher risk of household crimes such as burglary and motor vehicle theft (Clancy et al. 2001).

In the ICVS, slightly fewer of the Middle Eastern/Vietnamese sample (48%) reported experiencing any of the crimes included in this survey over the previous five years than the main sample (52%). This was not a statistically significant difference (see Table 1 for crimes included in this survey).

Five-year rates of victimisation were similar in both groups for the crimes of robbery, burglary and attempted burglary, theft of property from motor vehicles, and bicycle theft. A very small number of incidents reported by the Middle Eastern/Vietnamese group (1%) and the main sample (2%) occurred outside Australia.

The main sample had significantly higher:

  • overall rates of victimisation in the previous 12 months (17 per cent compared with 13 per cent for Middle Eastern/Vietnamese migrants);
  • one-year and five-year rates of personal crime; and
  • five-year rates of assault/threat and personal theft.

The Middle Eastern/Vietnamese sample reported significantly higher rates of motor vehicle theft over the previous five years compared with the main sample.

Ideally, one-year rates of victimisation should be calculated to assess risk because some personal characteristics, such as age and income, change over time. However, it is necessary to use five-year rates in this analysis due to the small sample of migrants.

Research has found that differences in rates of victimisation can be accounted for by certain socio-demographic characteristics such as age, gender, marital status and recreational activities. Among ethnic groups, victimisation risks may reflect variations in the demographic profile of groups, including socio-economic status and living in high crime neighbourhoods (Clancy et al. 2001). Lower rates of personal victimisation among migrants in Australia may be linked, in part, to the fact that they have some socio-demographic characteristics associated with lower risk. Equal proportions of the main sample and the Middle Eastern/Vietnamese migrant group lived in areas where they often see evidence of drug use (high social disorder areas). Although the Middle Eastern/Vietnamese migrants tended to be younger than the general population and more likely to live in large cities (factors which increase the risk of victimisation), they are more likely to be married and occupied with home duties, which predicts lower rates. Women in the Middle Eastern/Vietnamese sample (70%) had higher rates of marriage than their male counterparts (62%). In addition, the selected migrant group was more likely than the main community sample to say they never go out in the evening or go out less than once per month (22 per cent compared with 17 per cent). Although the ICVS does not specify type of night-time activities, these may be different for the two groups (e.g. attendance at meetings compared with going to pubs or nightclubs or doing evening work) which could lead to differences in risk of personal crime.

Logistic regression was conducted to identify the most important risk factors for personal victimisation for the two sample groups, while holding constant the effects of the other variables (Table 2). Being unmarried and living in an area where drug use was common increased the risk of personal victimisation for both groups. One factor was significant for the Middle Eastern/Vietnamese sample but not the main sample - being born overseas reduced the likelihood of personal victimisation. In other words, risk was higher for second generation migrants who were born in Australia.

The three factors that significantly increased the likelihood of personal victimisation in the main sample but not in the Middle Eastern/Vietnamese sample were:

  • being young (under 25 years of age);
  • household income over $400 per week; and
  • spending most evenings outside the home.
Table 1 : One-year and five-year rates of victimisation (per cent)
Main sample Middle Eastern/Vietnamese sample
Five-year One-year Five-year One-year
% 95% CI RSE % 95% CI RSE % 95% CI RSE % 95% CI RSE
Rates of vehicle, motorcycle and bicycle theft and theft from vehicle are based on the number of vehicle and bicycle owners.
CI is confidence interval. There are 19 chances in 20 that the true figure lies within this range.
RSE is relative standard error, the percentage of the estimate accounted for by the standard error.
- relative standard error is greater than 25 per cent
* Differences between the main and migrant samples are statistically significant p
Source: Australian Institute of Criminology, International Crime Victimisation Survey, 2004 [computer file]
Total 52 51 - 54 1 17* 16 - 19 3 48 45 - 52 4 13* 11 - 15 9
Total personal crime 30* 28 - 31 2 9* 8 - 10 5 17* 14 - 19 8 3* 2 - 4 21
Assault/threat 19 17 - 20 3 5 4 - 5 6 9* 7 - 11 11 - - -
Robbery 4 3 - 4 7 1 1 - 1 16 3 2 - 5 23 - - -
Personal theft 15 14 - 16 4 4 3 - 4 7 8* 6 - 10 14 - - -
Total household crime 39 38 - 40 2 11 10 - 11 4 40 37 - 43 4 10 8 - 12 10
Burglary 13 12 - 14 3 3 2 - 3 8 16 13 - 18 7 3 2 - 4 19
Attempted burglary 10 9 - 11 4 2 2 - 3 8 8 6 - 10 11 - - -
Motor vehicle theft 7* 6 - 8 5 1 1 - 2 13 13* 10 - 15 9 2 2 - 3 21
Theft from motor vehicle 21 20 - 22 3 5 4 - 6 6 19 16 - 21 7 4 3 - 6 15
Motorcycle theft 5 3 - 6 17 - - - - - - - - -
Bicycle theft 10 9 - 11 5 2 2 - 3 11 12 9 - 23 14 - - -

Risk of household victimisation

Within the selected migrant group, patterns in household victimisation were similar to those shown for personal victimisation. Rates were higher for those who often see evidence of drug use in the local area and lower for those who were born overseas.

Table 2 : Risk factors for personal victimisation, logistic regression
Main sample Middle Eastern/Vietnamese sample
Adjusted odds ratios RSE 95% CI Adjusted odds ratios RSE 95% CI
CI is confidence interval. There are 19 chances in 20 that the true figure lies within this range.
RSE is relative standard error, the percentage of the estimate accounted for by the standard error.
* p
Source: Australian Institute of Criminology, International Crime Victimisation Survey, 2004 [computer file]
Gender (male) 0.90 0.06 0.79 - 1.0 1.24 0.17 0.88 - 1.74
Age (16-24) 1.30* 0.10 0.10 - 1.54 1.02 0.24 0.63 - 1.64
Marital status (unmarried) 2.10* 0.07 1.84 - 2.39 2.21* 0.22 1.45 - 3.39
Income (<$400 per="" week)=""> 0.66* 0.09 0.55 - 0.78 0.93 0.23 0.58 - 1.47
Unemployed 1.30 0.19 0.88 - 1.87 1.82 0.33 0.95 - 3.49
Evenings out almost every day 1.27* 0.10 1.05 - 1.54 1.23 0.23 0.78 - 1.94
Born overseas 0.99 0.05 0.90 - 1.09 0.55* 0.22 0.36 - 0.86
Drug use in local area 3.18* 0.09 2.69 - 3.77 2.80* 0.22 1.81 - 4.31
-2 log likelihood 6651.1
Model chi square 505.5* (8 df)
-2 log likelihood 920.49
Model chi square 83.7* (8 df)

For the main group, rates of household victimisation were higher for those reporting household income of $400 per week or more, perhaps due to the greater availability of valuable household goods belonging to higher income households. Those who were new to the neighbourhood (living at their current postcode for less than one year) and those living in high drug use areas also reported higher rates of household victimisation (Table 3). The one risk factor common to both sample groups was living in a high drug use area.

Table 3 : Risk factors for household victimisation, logistic regression
Main sample Middle Eastern/Vietnamese sample
Adjusted odds ratios RSE 95% CI Adjusted odds ratios RSE 95% CI
CI is confidence interval. There are 19 chances in 20 that the true figure lies within this range.
RSE is relative standard error, the percentage of the estimate accounted for by the standard error.
* p
Source: Australian Institute of Criminology, International Crime Victimisation Survey, 2004 [computer file]
Income (<$400 per="" week)=""> 0.54* 0.08 0.47 - 0.63 1.03 0.16 0.76 - 1.41
Time at postcode (< one year) 1.41* 0.10 1.18 - 1.68 0.93 0.29 0.53 - 1.63
Born overseas 0.93 0.07 0.82 - 1.06 0.73* 0.16 0.53 - 1.00
Drug use in local area 2.27* 0.08 1.93 - 2.66 2.19* 0.20 1.47 - 3.26
-2 log likelihood 7640.6
Model chi square 188.7* (4 df)
-2 log likelihood 1481.2
Model chi square 19.3* (4 df)

Racially-motivated assault/threats

Qualitative interviews with migrants have found that racially-based threats and attacks can have a negative impact on victims and their communities, resulting in an increase in fear, a growing sense of alienation, a distrust of authority and an eventual reluctance to engage with police (Human Rights and Equal Opportunity Commission 2004). The British Crime Survey estimates that 40 per cent of assaults and 72 per cent of threats against some racial minorities are racially-motivated (FitzGerald & Hale 1996). Four per cent of all crimes and 11 per cent of assaults in the most recent Canadian victimisation survey were considered by victims to be hate crimes, with the majority motivated by hatred of the victim's race or ethnicity (other possible responses were gender, sexual orientation, language and disability). Rates were higher for visible minorities compared with the rest of the population (Janhevich 2001; Silver, Mihorean & Taylor-Butts 2004).

Victims of assault/threat in the ICVS were asked the following question to determine whether the incident was perceived to be racially motivated.

Do you feel you were assaulted or threatened because of your skin colour, ethnicity, race or religion?

Forty-two per cent of incidents of assault against the Middle Eastern/Vietnamese sample were perceived to be racially-motivated, including 53 per cent of threats and 38 per cent of attacks. Ten per cent of incidents against the main sample were considered by victims to be racially-motivated (12 per cent of attacks and eight per cent of threats). Those in the main sample who perceived these incidents to be racially-motivated were disproportionately born in Asia or Africa (14%), elsewhere overseas (30%), and non-English-speaking (20%).

Other differences in racially-motivated incidents reported by the Middle Eastern/Vietnamese migrant group and the main community sample were:

  • Middle Eastern/Vietnamese respondents were more likely to perceive racial assault/threats to be very serious (60 per cent compared with 45 per cent of non-migrants), but less likely to consider racial incidents to be crimes (58 per cent compared with 75 per cent). Migrants were more likely to say they didn't know whether it was a crime or not (12 per cent compared with none of the main sample);
  • racially-motivated incidents against the Middle Eastern/Vietnamese sample were slightly less likely to involve multiple offenders (49 per cent compared with 55 per cent of the main sample); and
  • strangers were the perpetrators in a higher proportion of racially-motivated assault/threats against the selected migrant sample (79 per cent compared with 55 per cent of the main sample).

The rate at which these incidents were reported to police was similar for the two sample groups. Migrants reported 38 per cent of racially-motivated incidents compared with 42 per cent of the main sample (a non-significant difference). The reporting rate for other types of assaults was 38 per cent for both groups. Reasons for not reporting could not be examined in detail due to small sample counts in the selected migrant sample. Equal proportions of both samples were satisfied with the way police responded to a report of assault (about two thirds of both groups).

In some countries, including Australia, agencies have been set up to help victims of assault by providing information or practical or emotional support. Small proportions of the Middle Eastern/Vietnamese migrant group (10%) and the main sample (7%) used the services of a specialised agency following an assault or threat. Use of a specialised agency rose to 16 per cent of Middle Eastern/Vietnamese victims of racially-motivated incidents.

Fear for personal safety

Research in Canada and the UK has found that minority groups are more likely to express concerns about their personal safety (Brzozowski & Mihorean 2002; FitzGerald & Hale 1996). In the ICVS, the majority of both samples feel safe while walking alone in the local area after dark, however a slightly smaller proportion of the Middle Eastern/Vietnamese sample feel this way (65 per cent compared with 72 per cent of others). Even after controlling for other possible confounding factors, such as age and living in an area with high public drug use, those from Middle Eastern/Vietnamese backgrounds report higher levels of fear.

A significantly higher proportion of the main sample feel very safe (32%) compared with those from the Middle Eastern/Vietnamese migrant sample (24%). Conversely, those from the selected migrant groups (12%) are more likely to feel very unsafe than the main sample (9%). There were differences between the two migrant groups. Although sample sizes are small, a higher proportion of those with Middle Eastern backgrounds (35%) report feeling very safe compared with Vietnamese migrants (9%). Vietnamese migrants were more likely to say they feel fairly safe (53% compared with 31%).

The differences are exacerbated when gender is added (Figure 1). Those who feel safest walking alone in the local area after dark are males from the main sample, followed by selected migrant men, then women from the main sample. Women with Middle Eastern/Vietnamese backgrounds are least likely to feel safe. In fact, almost equal proportions of these women feel very unsafe (18%) as very safe (16%). Middle Eastern/Vietnamese women were about one-third as likely as men from the main sample to feel very safe, and nine times as likely to feel very unsafe. However, the differences between the two sample groups were greater for men. The proportion of men who feel very safe walking alone in the local area after dark was higher for the main sample (44%) than for the Middle Eastern/Vietnamese sample (31%), whereas the differences for women were small (19 per cent compared with 16 per cent).

Part of the explanation for higher fear levels among those with Middle Eastern/Vietnamese background may be their higher rates of racially-based incidents. This group was more likely to express concern about racially-motivated attacks than the main sample. Seven per cent were very worried and 19 per cent were somewhat worried about being assaulted or threatened because of their skin colour, ethnicity, race or religion (Figure 2). The comparative figures for the main sample were one and eight per cent respectively.

Amongst those who had experienced racially-motivated assaults or threats within the previous five years, the Middle Eastern/Vietnamese sample in particular, were far more likely to be worried about a similar assault or threat in the future. Twenty-eight per cent of migrant victims of a racially-motivated incident and five per cent those in the main sample were very worried about another attack. Those in the main sample who were worried about a racially-motivated attack in the future were disproportionately born in Asia or Africa and are non-English-speaking.

Figure 1: Feelings of safety walking alone in the local area after dark, by migrant status and gender

Differences for males and females and for the main and migrant samples are statistically significant p

Source: Australian Institute of Criminology, International Crime Victimisation Survey, 2004 [computer file]

Figure 2: Concerns about racially-motivated assault/threats, by migrant status and prior racially-based victimisation

Differences between the main and migrants samples are statistically significant p

Source: Australian Institute of Criminology, International Crime Victimisation Survey, 2004 [computer file]

A logistic regression was run to identify significant risk factors for feeling very unsafe walking alone in the local area after dark (Table 4). Even after controlling for a range of factors, including being in the selected migrant group, being female was the most important risk factor of feeling very unsafe. Belonging to the Middle Eastern/Vietnamese migrant group also remained a significant risk factor. Other risk factors were:

  • being 60 years of age or older;
  • having income under $400 per week;
  • having been a victim of crime in the previous five years;
  • living in area were there is visible drug use; and
  • perceiving the police to be doing a poor job at controlling crime.
Table 4 : Risk factors for feeling very unsafe, logistic regression
Adjusted odds ratios SE 95% CI
* p
Source: Australian Institute of Criminology, International Crime Victimisation Survey, 2004 [computer file]
Gender (female) 6.08* 0.11 4.87 - 7.59
Age (60 or over) 1.83* 0.11 1.46 - 2.30
Income (<$400 per="" week)=""> 1.71* 0.11 1.37 - 2.12
Capital city 1.17 0.10 0.96 - 1.43
Victim of crime 1.59* 0.10 1.32 - 1.91
Drug use in local area 3.20* 0.11 2.56 - 4.00
Police performance poor 2.29* 0.12 1.82 - 2.88
Migrant 1.69* 0.11 1.35 - 2.12
-2 log likelihood 3670.2
Model chi square 615.3* (8 df)

Conclusions

Victimisation surveys are important tools for understanding the experiences of crime in migrant communities. The results of this survey suggest that these selected migrants have similar rates of victimisation overall compared with a general community sample, and lower rates of assault and personal theft. One factor that increases risk of personal and household crime for both sample groups was living in an area with high levels of social disorder (as indicated by visible signs of drug use). The selected migrant group reported significantly higher rates of racially-motivated assaults and threats and are more likely to fear a racially-motivated incident in the future. Migrant victims of racially-motivated incidents tended to feel the situation was very serious, but many are unsure whether to label it a crime. Ongoing dialogue between migrant groups, police and local communities may help to reduce the vulnerability of certain groups to this form of victimisation and improve feelings of safety.

Future investigations into the experiences of crime among migrant communities require larger samples that permit deeper analyses of diverse groups separately; 'migrant' is not a homogeneous category. Future work in this area also needs to go beyond the crimes included in this survey to examine verbal harassment, vandalism, graffiti and damage to community agencies or buildings, all of which can have a widespread negative impact on minority communities.

Acknowledgment

The ICVS was funded by the Australian Government Attorney-General's Department and the Department of Immigration and Multicultural and Indigenous Affairs.

References

All URLs correct at August 2005

  • Australian Bureau of Statistics 2004. Migration. Catalogue no. 3142.0. Canberra: Australian Bureau of Statistics
  • Brzozowski JA and Mihorean K 2002. Technical report on the analysis of small groups in the 1999 General Social Survey. Catalogue no 85F0036XIE. Ottawa: Statistics Canada.
  • Challice G and Johnson H 2005. The Australian component of the International Crime Victimisation Survey (ICVS). Technical and background paper series no. 16. Canberra: Australian Institute of Criminology.
  • Clancy A, Hough M, Aust R and Kershaw C 2001. Crime, policing and justice: the experience of ethnic minorities. Home Office Research Study no 223. London: Home Office.
  • FitzGerald M and Hale C 1996. Ethnic minorities, victimisation and racial harassment. Research findings no 39. London: Home Office.
  • Human Rights and Equal Opportunity Commission 2004. IsmaU - Listen: national consultations on eliminating prejudice against Arab and Muslim Australians. Sydney: Human Rights and Equal Opportunity Commission.
  • Janhevich D 2001. Hate crime in Canada: an overview of issues and data sources. Ottawa: Statistics Canada.
  • Johnson H 2005. Crime victimisation in Australia: key results of the 2004 International Crime Victimisation Survey. Research and public policy series no 64. Canberra: Australian Institute of Criminology.
  • Silver W, Mihorean K and Taylor-Butts A 2004. Hate crime in Canada. Juristat vol 24 no 4.

Dr Holly Johnson was a visiting research fellow at the Australian Institute of Criminology from 2003 to 2005.