35
Immigration Status and Victimisation Evidence from the British Crime Survey George Papadopoulos University of East Anglia December 13, 2012 George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 1 / 35

Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

  • Upload
    lamnhi

  • View
    213

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Immigration Status and VictimisationEvidence from the British Crime Survey

George Papadopoulos

University of East Anglia

December 13, 2012

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 1 / 35

Page 2: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Introduction-Motivation

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 2 / 35

Page 3: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Introduction-Motivation

Introduction - Immigrants as OffendersThe relationship between Immigration and Crime has always been a hot topic among bothscholar & non-scholar communities.

The International Social Survey Programme (ISSP) shows that British citizens believethat immigrants increase crime rates in the UK.

0

5

10

15

20

25

30

35

40

Strongly Disagree Disagree Indifferent Agree Strongly Agree

1995 (n=996)

2003 (n=834)

Most research on ‘Immigration & Crime’ looks at this association from the offendingpoint of view with mixed evidence.The empirical evidence for the UK is limited though.

One comprehensive study of criminal behaviour by Papadopoulos (2011), using theOffending, Crime & Justice Survey of crime self-reports, shows that immigrants are lesslikely to commit crimes compared to natives.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 3 / 35

Page 4: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Introduction-Motivation

Motivation - Immigrants as VictimsCrime has two sides (offenders and victims) which are interconnected, especially forviolent crime.

I For example, being an offender signifies a social lifestyle associated with a higherlikelihood of victimisation compared to a non-offender(see, Deadman & MacDonald, JRSS(A), 2004).

Thus, looking at the association between immigration & victimisation may reveal someinteresting insights on immigrants’ behaviour towards criminal activities.

However, this relationship has been totally neglected, at least in terms of quantitativeresearch.

The present study ‘completes’ the Immigration-Crime picture by comprehensivelyexamining whether immigrants are of higher or lower risk of victimisation compared tonatives.

It uses data from the British Crime Survey (BCS) of 2007/08, a representativevictimisation survey of around 47,000 households, where respondents in England andWales were asked in face-to-face interviews about their victimisation experiences inboth property and personal crime(see, Bolling, Grant & Donovan, 2008, for details).

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 4 / 35

Page 5: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Instrumental Vs Violent Crime

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 5 / 35

Page 6: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Instrumental Vs Violent Crime

Instrumental Vs Violent Crime - An Important Distinction

Property Crimes & Personal Thefts: Target is to acquire victim’s property.I Interactions and interrelations between potential offenders and potential victims

prior to the criminal action are not important.F Personal behaviour (unobserved social lifestyles) is NOT a strong predictor of

instrumental victimisation.

Violent Crimes: Target is to hurt victim.I They are interactive acts that depend upon the actions and relations of both

offenders and victims.F Personal behaviour is a very strong predictor of violent victimisation.

This is recognised by criminological theories of “lifestyle exposure and routineactivities” (see, Meier & Miethe, Crime & Justice, 1993).

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 6 / 35

Page 7: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Immigration and Victimisation

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 7 / 35

Page 8: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Immigration and Victimisation

Immigration & Victimisation

For Property Crime the fact that the household reference person (HRP) is animmigrant should not be related to the risk of victimisation, given that we areable to control for all household/personal characteristics and othersocio-economic constraints associated with risks of victimisation. The sameshould hold for Personal Thefts.

For Violent Crime, however, there might still be unobserved behaviouralcharacteristics specific to immigrants, such as unobserved lifestyle choices,associated with different risks of victimisation.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 8 / 35

Page 9: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Data

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 9 / 35

Page 10: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Data

Data

My final data set consists of 46,827 observations, 16 year olds and older.

The dependent variable (Y) takes value 1 if the individual suffered a crimeduring the 12 months prior to the interview date, and 0 otherwise.Today I present results for the following crime categories:

1 Inside Burglaries (2.32% positives)2 Personal Thefts (1.59% positives)3 Violence (2.54% positives)

I Note that all crimes include attempts.

The the dummy variable IMMIGRANT takes value one if the respondent is bornoutside the UK.

I According to this, we have 4,466 (9.54%) immigrants.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 10 / 35

Page 11: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Data

Native-Immigrant Mean ComparisonsOn average, immigrants:

I are relatively younger and, constrained by their socio-economic status, reside inmore deprived urban areas;

I are relatively more from ethnic minorities⇒ Racially Motivated Crime;I own fewer and less attractive properties.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 11 / 35

Page 12: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Inside Burglaries

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 12 / 35

Page 13: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Inside Burglaries

Inside Burglaries

Indeed, once we control for basic demographics, immigrants and natives are equallylikely to suffer an inside burglary, as expected.

Without controls, ̂Pr(Y = 1|Im = 1) = 3.00%, while ̂Pr(Y = 1|Im = 0) = 2.25% ⇒ Diff= 0.75∗∗∗ p.p., a Relative Effect of 33.3%.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 13 / 35

Page 14: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Personal Thefts

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 14 / 35

Page 15: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Personal Thefts

Personal Theft (roberries, mugging and other thefts)

We observe a positive relationship which, however, vanishes once we control forregion (because personal crime is much higher in areas of London).

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 15 / 35

Page 16: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Violence

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 16 / 35

Page 17: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Violence

Violence: There is a robust negative relationship between immigration statusand the likelihood of violent victimisation. A general explanation could be that:Immigrants set strategies that correspond to unobserved differences in“routine activities and lifestyle exposure” associated with lowervictimisation risks compared to natives.

Page 18: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Types of Violent Crimes

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 18 / 35

Page 19: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Types of Violent Crimes

Types of Violent Crimes: A First Look

However, in the analysis so far we have ignored an important point.

Violence can be divided into three crime types, which may be generated by verydifferent mechanisms:

I Domestic CrimeI Crime by AcquaintancesI Crime by Strangers

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 19 / 35

Page 20: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

A Trivariate Probit Model for Violent VictimisationUsing a trivariate probit that allows for correlated errors among the three crimeequations and controlling for basic demographic characteristics, we find the following:

This implies that, victims of domestic crime are more likely to suffer a crime byacquaintances (and vice versa) due to common unobserved factors, and victims of crimesby acquaintances are more likely to suffer a crime by strangers (and vice versa) due todifferent unobserved common factors.

Page 21: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Types of Violent Crimes

These results are not consistent with the previous hypothesis, since if immigrantspresented the aforementioned social behaviour, we would expect a negativerelationship for crime by strangers as well!

What are the mechanisms generating this result?1 Immigrants actually do not follow the “crime-avoiding" social lifestyles but we still

observe a negative effect for crime by acquaintances and domestic crime because:F immigrants are less willing than natives to report domestic violence and crime

committed by acquaintances;F immigrant have fewer household members or smaller networks of

acquaintances compared to natives.

2 Immigrants actually follow the “crime-avoiding" social lifestyles but we do notobserve a difference for crime committed by strangers because immigrants are morelikely to suffer Racially Motivated Crimes (RMC) compared to natives (RMC ismostly committed by strangers, does not depend on the personal behaviour ofpotential victims and it is traditionally associated with ethnic minorities).

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 21 / 35

Page 22: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Sensitivity Analysis

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 22 / 35

Page 23: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Sensitivity Analysis Self-Completions for Domestic Crime

Under-reporting? Self-Completions for Domestic CrimeRespondents between 16 - 59 were asked to self-complete a computer based questionnairefor domestic crime.

Self-completions are used as a technique to elicit more reliable responses to sensitivequestions, such as questions about domestic crime (see, Turner at al, Science, 1998).

Only 0.51% of the respondents reported a domestic crime in face-to-face, but 3.64% inself-completion interviews.

Given that individuals report truthfully in self-completions, if immigrants under-reportby more than natives in face-to-face interviews, the effect of being an immigrant on domesticvictimisation should be smaller in self-reports compared to face-to-face interviews.

Sample selection issue: Some people refused participation in self-completions.I 32% of immigrants did not participate but 18% of natives.I If this selection is related to unobserved factors that affect victimisation

(if for example, individuals that choose not to participate in self-completions are themost victimized ones), then we will face sample selection bias.

I We can test for it by using sample selection models(Heckman, Econometrica, 1979)

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 23 / 35

Page 24: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Sensitivity Analysis Self-Completions for Domestic Crime

Probit with Sample Selection, Van de Ven & Van Praag JAE (1981)

We first notice that there is not evidence of sample selection bias; selection onobservables seems more possible.

In this case we can also apply the Inverse Probability Weight (IPW) estimator proposedby Wooldridge, JoE (2007).

Most importantly, the difference remains very high and significant!

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 24 / 35

Page 25: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Sensitivity Analysis Network Effects or Assimilation Patterns

Network Effects and/or Assimilation PatternsIs it that immigrants have fewer hhd members or a smaller network ofacquaintances?

For domestic violence this is not the case. Even most recent immigrants’ hhdscontain 0.15 people more and the difference increases by 0.001 persons per year.

However, NO information exists on the number of acquaintances.

But if immigrants’ networks expand as they stay more in the country, then formost recent immigrants we should find a large difference that closes as time inthe country increases.

If network effects exist they are very weak. Instead, other assimilation patternsseem to exist.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 25 / 35

Page 26: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Sensitivity Analysis Racially Motivated Crime

Controlling for Racially Motivated CrimeDo immigrants suffer more racially motivated crime compared to natives?If yes, can this difference explain why immigrants and natives face the samevictimisation risk by strangers?

Victims indicated whether a suffered crime was of racial motive or not.I Here we deal with perceived RMC⇒ misclassification is possible.I If immigrants and natives tend to misclassify similar crimes in opposite directions,

then our estimates will be downward biased.

We can control for this cases, by turning all RMCs to zeroes, both for natives andimmigrants.So, a part of the difference could be explained by RMC(given that there is no misclassification).

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 26 / 35

Page 27: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Sensitivity Analysis Racially Motivated Crime

But why is the estimated difference of lower magnitude than the one found forcrime by acquaintances and domestic crime?

A possible explanation is the following:I Family and acquaintances of an immigrant are likely to include an

important percentage of other immigrants. If immigrants follow acrime-avoiding behaviour, then the low victimisation of an immigrant by hisfamily and acquaintance group happens both because: 1) he/she is acrime-avoider himself/herself; 2) the reference group is composed in largepart of crime-avoiders (enhancement effect).

I On the contrary, the pool of strangers that an immigrant meets is composedby the same proportion of immigrants as the pool of strangers met by anative. Therefore, the enhancement effect does not apply in this case.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 27 / 35

Page 28: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Discussion

Contents1 Introduction-Motivation

2 Instrumental Vs Violent Crime

3 Immigration and Victimisation

4 Data

5 Inside Burglaries

6 Personal Thefts

7 Violence

8 Types of Violent Crimes

9 Sensitivity AnalysisSelf-Completions for Domestic CrimeNetwork Effects or Assimilation PatternsRacially Motivated Crime

10 Discussion

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 28 / 35

Page 29: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Discussion

Conclusion

Immigrants do not under-report by more than natives.

Racially Motivated Crime can explain to some extent why immigrants face thesame probability as natives to suffer a crime by strangers.

If network effects exist, they are quite weak.

Accepting that immigrants are crime-avoiders, so that they also less likely tocommit violent crimes relative to natives, we expect to find a negativerelationship for domestic crime and crime by acquaintances but no relationshipfor crime by strangers, just because immigrants’ family members andacquaintances are mostly immigrants as well, but this is not true for the group ofstrangers.

Thus, to some extent, all evidence shows that, indeed, immigrants follow sociallifestyle choices associated with a lower risk of violent victimisation.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 29 / 35

Page 30: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Discussion

Extra Topics

The count information of victimisation incidents is exploited: This might say somethingabout repeated victimisation.

I There are a few cases where extreme number of victimisation were reported. The count datamodels are very sensitive to these cases, particularly because the positive counts are too few toidentify the parameters of the variables assumed to affect the mean.

I Censored Models (increase robustness), Quantiles for counts (extreme cases do not affect thequantiles).

I Develop a two part-model (hurdle model) for censored counts that recognizes that perhaps theprobability of victimisation and the number of victimisation incidents are generated by twodifferent processes (for instance, probability of violence is higher for men, but once the hurdleis passed, women suffer more crime).

I The results suggest that the effect of immigration on victimisation is fairly constant across thedistribution of crime.

Seriousness of crime: Immigrant victims perceive crime suffered as more serious thannatives.

Interaction Terms: Immigrants not a homogeneous group. However, the results usinginteraction terms show that the effect of immigration status is fairly the same among forexample, among different immigrant ethnic groups or different locations. However, theseresults are inconclusive and further research on the topic is required.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 30 / 35

Page 31: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

ReferencesBolling, K., C. Grant, and J. Donovan (2008), 2007-08 British Crime Survey (England and Wales),

Technical Report I-II, Home Office.

Deadman, D., and Z. MacDonald (1997), “Offenders as Victims of Crime?: an Investigation into theRelationship between Criminal Behaviour and victimisation”, Journal of the Royal Statistical Society,Series A (Statistics in Society), 167, 53-67.

Heckman, J. J. (1979), “Sample Selection Bias as a Specification Error”, Econometrica, 47, 153-161.

Meier, R. F. and T. D. Miethe (1993), “Understanding Theories of Criminal victimisation”, Crime andJustice, 17, 459-499.

Papadopoulos, G. (2011), “The Relationship between Immigration Status and Criminal Behaviour:Evidence from the Offending, Crime, and Justice Survey”, Unpublished PhD manuscript, Department ofEconomics, University of Essex.

Terracol, A. (2002), “Triprobit and the GHK Simulator: A Short Note”, mimeo.

Turner, C. F., L. Ku, S. M. Rogers, L. D. Lindberg, J. H. Pleck, and F. L. Sonenstein (1998),“Adolescent Sexual Behavior, Drug Use, and Violence: Increased Reporting with Computer SurveyTechnology”, Science, 280, 867-873.

Van De Ven, W., and B. Van Praag (1981), “The Demand for Deductibles in Private Health Insurance. AProbit Model with Sample Selection”, Journal of Econometrics, 17, 229-252.

Wooldridge, J. M. (2007), “Inverse Probability Weighted Estimation for General Missing Data Problems”,Journal of Econometrics, 141, 1281-1301.

Page 32: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Discussion

Violent Vs Instrumental Crime - An Important Distinction

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 32 / 35

Page 33: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Discussion

Outside Burglary (garages, outhouses, storehouses, conservatories, gardens)

Marginal Effect (%): 0.65 - 1.05 = -0.40**, Relative Effect (%): 61.5% decreaseImmigrants are less likely to suffer an Outside Burglary⇒ perhaps because they ownfewer ‘outside’ properties.If immigrants become “better-settled” as time in the country increases - so that theirproperties become more similar to natives’ ones -, we would expect that most recentimmigrants are the least victimized with an assimilation pattern as years in the countryincreases⇒ It takes around 56 years to assimilate.George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 33 / 35

Page 34: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Discussion

Partner Present during the Face-to-Face InterviewPresence of partner during face-to-face interviews might result in under-reporting if thequestions refer to very sensitive information (see, Acquilino, Public Opinion Quarterly,1993).

In my data, if the partner is present, the probability to report a domestic crime is only0.19%, but 0.57% if the partner is not present.

Is the effect of being an immigrant on the risk of victimisation different between the‘partner present’ and ‘partner not present’ situations?

Consider the following probit model with an interaction term between ‘being animmigrant’ and ‘partner being present’:

E(yi|X) = Φ(β0 + β1Immig.+ β2Partn.Present + β3Immig.× Partn.Present)

If immigrants under-report by more, the estimated gap for the cases where partner ispresent should be higher.

I β3 should be negative.

But “partner’s presence” is not assigned randomly. We need to control for all those factorsthat are associated with the presence of the partner.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 34 / 35

Page 35: Immigration Status and Victimisation Evidence from the British · PDF file · 2013-08-30Immigration Status and Victimisation Evidence from the British Crime Survey ... there might

Discussion

Partner Present during the Interview

Instead, the opposite pattern is observed in face-to-face Domestic crime!Moreover, in self-completion Domestic crime & Crime by Acquaintances, where thepresence of partner should not make any difference, we actually find no change in bothimmigrants’ and natives’ reporting behaviour.According to this, immigrants’ reporting behaviour seems more reliable.

George Papadopoulos (UEA/ECO) Immigration Status and Victimisation 13-12-12 35 / 35