Presentation is loading. Please wait.

Presentation is loading. Please wait.

21 May 2009 Copyright Kautt Assessments of the CJS: Does respondent ethnicity make a difference? An ESRC-funded Study Dr. Paula M. Kautt.

Similar presentations


Presentation on theme: "21 May 2009 Copyright Kautt Assessments of the CJS: Does respondent ethnicity make a difference? An ESRC-funded Study Dr. Paula M. Kautt."— Presentation transcript:

1 21 May 2009 Copyright Kautt Assessments of the CJS: Does respondent ethnicity make a difference? An ESRC-funded Study Dr. Paula M. Kautt

2 21 May 2009 Copyright Kautt The Research Reported Today  Uses British Crime Survey data Examines: Perceptions of the Justice System (CJS)  Criminal Justices System perceptions index Factor Analysis used to determine components Quantitative Analyses  Multivariate Differences by Ethnicity  Separate analysis for each ethnicity rather than dummy variables in pooled analysis*  Multilevel Respondent/ Household Police Force/Criminal Justice Area (PFA/CJA) levels  Comparative Z-test for equality of coefficients (Clogg et al, 1995; Paternoster et al, 1998;)

3 21 May 2009 Copyright Kautt British Crime Survey Data  Although used to examine various criminological questions Few multivariate analyses examine variation in criminological experiences between BMEs These largely rely on pre-2001 BCS data  Problematic given subsequent events (9/11, CJA 2003, 7/7)  This study addresses the gap by Focusing on differences by BME status  Independent models for each ethnic group Using data from 2001-2007 Accounting for variation by Police Force Area (PFA)

4 21 May 2009 Copyright Kautt Why examine by Ethnicity?  Although comprising a small percentage of the population, adverse differential treatment of BME persons has been well documented  There is every reason to expect their perceptions and experiences will be different both from ‘British Whites’ and each other Differences have implications for policy effectiveness  e.g. Public Service Agreement

5 21 May 2009 Copyright Kautt Background: Perceptions of the CJS  Tension between BMEs and the police may influence BME respondents’ perceptions e.g. Lawrence, Menendez  Clancy et al (2001) find that being a BME group member significantly affects assessment of CJS agencies Yet, other analyses find no such effects (Jansson, 2006) These disparate results suggest that further exploration is necessary

6 21 May 2009 Copyright Kautt A notable anomaly of BCS data  As captured, only ‘Whites’ can be simply ‘British’ ‘White-British’  It is not possible for ‘Black’ or ‘Asian’ respondents to simply be ‘British’ Rather, specific ethnicity is always attached  ‘Black or Black-British– Caribbean’  ‘Asian or Asian-British— Indian’

7 21 May 2009 Copyright Kautt Implications for analysis by ethnicity  Thus, can legitimately partition respondents at two ethnic tiers General Ethnicity  A larger group differentiated mainly by the traditional ‘racial’ groupings White, Black, Asian ‘Other’ is also a category Specific Ethnicity  A more refined comparison of groups falling within a general ethnicity Under ‘Asian’, differences between the sub-groups ‘Indian’, ‘Pakistani’, ‘Bangladeshi’ and ‘Chinese’’ can be compared

8 21 May 2009 Copyright Kautt What about ‘Other’?  Respondents from the ‘Other’ are omitted from the analyses presented here  Rationale: there is no way to meaningfully talk about these groups because : They comprise an unknown number of different ethnicities The proportion of these ethnicities is also unknown

9 21 May 2009 Copyright Kautt And ‘Mixed’?  Due to small numbers ‘Mixed’ cases are merged into the specifically- named BME group Rather than analysed independently or merged into ‘White’ The supposition is that they  Will appear to be a member of that group to a third party (e.g. offender, police)  Perceive themselves as a member of that group At least partially

10 21 May 2009 Copyright Kautt Comparison Problems  BME Sample Size Majority of respondents are White British  Much smaller numbers of the Asian and Black General Ethnicities  Makes meaningful statistical comparisons between ‘White’ and other General Ethnicities problematic Implications for significance tests

11 21 May 2009 Copyright Kautt To understand the complexity, multiply by ten…

12 21 May 2009 Copyright Kautt Solution?: Pooling the Data  Because BMEs comprise a small proportion of the BCS sample, data are combined across years (2001- 2007) to: Meaningfully analyse responses by ethnic group Provide sufficient numbers to support the analyses  Despite this, in some instances, there are still too few cases In such instances, an ethnically pooled model is the only option  NOTE: because of pooling, only indicators available for all years can be used e.g. religion cannot be controlled in these analyses Dummy variables capture year

13 21 May 2009 Copyright Kautt Frequencies: Perceptions of the CJS  Experience of CJS (CJS index) subsample Total: 47,848  Including ‘other’ General Ethnicity Totals  White: 39,933  Black: 2,098  Asian: 3,096  These are not evenly distributed across PFA

14 21 May 2009 Copyright Kautt The 30/30 multilevel “rule”  For continuous outcomes, minimum counts for contextual (variation in predictors by PFA) multilevel analysis (Maas and Hox, 2005) are: Number of groups (PFAs): 30 Number of cases (respondents) per group: 30

15 21 May 2009 Copyright Kautt Capturing Perceptions of the CJS  Criminal Justice Index Composed of seven different ordered (1-5) response items  ‘How confident are you that the CJS is effective in bringing people to justice?‘ the CJS respects the rights of the accused?‘ the CJS meets the needs of crime victims?‘ the CJS deals with cases promptly and efficiently?'  'How effective is the CJS in reducing crime?‘ in dealing with young people accused of a crime?'  'How good a job are the police doing?' Factor Analysis confirms these represent same latent factor  Minimum 7; maximum 35; Normal distribution  Higher scores reflect greater confidence

16 21 May 2009 Copyright Kautt Potential Influences  Respondent Age  Respondent Sex  Marital Status dummies  In a higher professional occupation  Health (ordinal)  Paid Work  Educational Qualifications  Years in Area  Years at Address  Victim of crime? (Ref Per)  Number of Victimisations (Ref Per)  Ever reported victimisation to police?  Ever arrested by police?  Worked in CJS?  Number Pub Visits per week  Number Club Visits per week  Hours away from home daily  Drunk problem in area (ord)  Noisy Neighbour Problem (ord)  Teens loitering problem (ord)  Litter problem in area (ord)  Vandalism problem in area (ord)  Skin colour attack problem (ord)  Drug Problem in area (ord)  Area Index (interviewer perception)  Rented HH  Inner City HH  HH type dummies (e.g. Flat)  Child in HH  Adult male in HH  Single adult HH  Security Index (composite)  Year “Dummy” Variables

17 21 May 2009 Copyright Kautt Criminal Justice System Index: Multilevel  Number of PFAs with sufficient cases General Ethnicity  Black: 9  Asian: 19  Insufficient cases for independent multilevel models of general ethnicity Pooled model only  i.e. Comparison of ‘Black’ and ‘Asian’ to ‘White’ reference category  Pooled Model results Assessment of the CJS was significantly better amongst ‘Asian’ and ‘Black’ respondents as compared to ‘Whites’ The impact of being either ‘Asian’ or ‘Black’ varied significantly by PFA  The higher the percentage of the PFA population that was non-white, the lower the ‘Asian’ assessment of the CJS; no impact for ‘Black’

18 21 May 2009 Copyright Kautt CJS Index: Single-level comparisons Individual Factors  Reference Period Victimisation lowered CJS assessment for White and Asian Respondents with a larger effect for Asians  Time lived in area lowered both Asian and White assessments but had a larger effect for Asians  Having worked in the CJS improved only White assessments  Having been arrested lowered only White assessments  Having been a juror in a criminal case improved Asian assessments but lowered both Black and White assessments  Having been the accused in court lowered assessments for Asian and Black respondents only  Paid work in past week lowered assessments for Asian and White respondents but had a larger effect for Asians

19 21 May 2009 Copyright Kautt CJS Index: Single-level comparisons Area Factors  Teens hanging about problem lowered assessments for Asians and Whites but had a larger effect for Asians  Rubbish, Vandalism/Graffiti, Drug, Public Drunkenness and Noisy Neighbour problems lower assessments for Whites only; no effect Asian or Black  Skin colour attack problem improves assessments for Whites; no effect others

20 21 May 2009 Copyright Kautt CJS Index: Single-level comparisons Routine Activity Factors  The number of Pub Visits in the past month lowered assessments for all three groups but did so more for Asian and Black respondents than White.  Daily hours away from home lowered assessments for Asian respondents only; no effect Black or White

21 21 May 2009 Copyright Kautt CJS Index: Single-level comparisons Household Factors  Presence of Children in HH improved Asian and White assessments but had a larger effect for Asians  Rented accommodation is associated with improved assessments for Asian and Black respondents; no effect White  Household is a flat improved assessments for Whites only; no effect Asian or Black

22 21 May 2009 Copyright Kautt CJS Index: Single-level comparisons Year of Interview  As compared to 2001 Assessments of Asian respondents improved in 2006 Assessments of Black respondents improved in 2003, 2004 and 2005 Assessments of White respondents improved in 2005, 2006 and 2007  Compared to each other 2005 improvement larger for Black respondents than White 2006 improvement larger for Asian respondents than White

23 21 May 2009 Copyright Kautt Implications  Dangerous to treat BME groups as interchangeable Separate analysis for each is best to assess differences  Different factors influence assessments by BME Cultural factors may explain this  Further disaggregation by specific ethnicity is in order  Testing for additional interactions should be explored

24 21 May 2009 Copyright Kautt Limitations and Lingering Questions  Theoretically salient factors are either not available in the data or across all data years Native Language/ESL Immigration status  1 st, 2 nd or 3 rd generation? How long in country? Citizenship status Religion ‘White’ ethnicities?  Sample sizes vary drastically between ethnic groups BCS data may not be fit for purpose of evaluating ethnic differences by location  Qualitative methods or special purpose quantitative surveys (in areas where there are BME populations) may be more appropriate


Download ppt "21 May 2009 Copyright Kautt Assessments of the CJS: Does respondent ethnicity make a difference? An ESRC-funded Study Dr. Paula M. Kautt."

Similar presentations


Ads by Google