Presentation is loading. Please wait.

Presentation is loading. Please wait.

Introduction to the British Crime Survey Keith Bolling, BMRB Social Research.

Similar presentations


Presentation on theme: "Introduction to the British Crime Survey Keith Bolling, BMRB Social Research."— Presentation transcript:

1 Introduction to the British Crime Survey Keith Bolling, BMRB Social Research

2 Outline Brief history of BCS development Design of the survey since 2001 Fieldwork and response Questionnaire development and content Offence classification Data processing and delivery timetable

3 History of the British Crime Survey, E&WE&W&S E&W Coverage

4 History of the British Crime Survey, E&WE&W&S E&W Coverage 14,61710,90510,39211,03010,05914,94716,34819,457 Sample size

5 History of the British Crime Survey, E&W E&W&S E&W Coverage 14,61710,90510,39210,05914,94716,34819,457 Sample size 77%80%77% 1984 E&W 11,030 78%77%79%83%74% Response

6 History of the British Crime Survey, E&W E&W&S E&W Coverage 14,61710,90510,39210,05914,94716,34819,457 Sample size 77%80%77% 1984 E&W 11,030 78%77%79%83%74% Response NoneEMNone EM y.o. EM Boosts EM

7 History of the British Crime Survey, E&W E&W&S E&W Coverage 14,61710,90510,39210,05914,94716,34819,457 Sample size 77%80%77% 1984 E&W 11,030 78%77%79%83%74% Response NoneEMNone EM y.o. EM Boosts EM IC (x3)IC (x1.5) IC (x2) PFA Over sampling IC (x2)

8 History of the British Crime Survey, E&W E&W&S E&W Coverage 14,61710,90510,39210,05914,94716,34819,457 Sample size 77%80%77% 1984 E&W 11,030 78%77%79%83%74% Response NoneEMNone EM y.o. EM Boosts EM IC (x3)IC (x1.5) IC (x2) PFA Over sampling IC (x2) CAPIPAPI Electoral RollPAF

9 Methodological review of the BCS Significant design changes following a methodological review (Lynn and Elliot, 2000) Main recommendations were: Rotation of half the PSUs each year Over sampling in smaller police force areas Application of non-response weighting Better application of low-level geo-identifiers Continuous fieldwork Change of reference period

10 Change to the survey reference period Prior to 2001 fieldwork was conducted in the first few months of the year (January – April) Reference period was fixed - the previous calendar year (although respondents asked about last months) Proposed new reference period was rolling– the last 12 months Experiment carried out in 2001 to test impact of different reference period on victimisation rates

11 Main characteristics of the survey from 2001 Core sample size –approximately 37,000 per year Minimum of core interviews in each PFA Boost samples Ethnic boost (3,000 per year) year old boost (2000 per year) Fieldwork managed on monthly basis, with quarterly delivery of data (moving to fiscal year)

12 Further changes to survey from Increase of core sample size to approximately 46,000 per year Minimum of 1,000 core interviews in each PFA Balancing the precision of national estimates vs. PFA level estimates Largest over samplingLargest under sampling Dyfed Powys (2.26)Metropolitan (0.55) Warwickshire (2.25)West Midlands (0.67) Cumbria (2.18)West Yorkshire (0.68) Bedfordshire (2.10)Greater Manchester (0.69)

13 Current fieldwork and interviewing Sample issued equally across the year – approx. 180 assignments per month Each assignment has 32 addresses selected from across a whole postcode sector Interviewer has to randomly select one adult (16+) from household for interview Interview conducted in-home using CAPI Average interview length of 48 minutes (65 minutes for victims of crime)

14 BCS response rate over time, 1982 – 2005/6

15 BCS response rate by GOR, Base: All respondents (n=1884) s & Ecst9asy Cannabis & Crack Cannabis & Heroin Heroin & Crack Cannabis & Amphmne Cannabis, Heroin & Crack Cocaine & Crack Cannabis, Cocaine & Crack Cocaine & Ecstasy

16 Ethnic minority boost sample on BCS, Done by focused enumeration – screening four adjacent addresses to the main sample address 171,632 addresses screened in Identified 6,195 eligible households (3.6%) Achieved 3,098 interviews – a response rate of 50%

17 Youth (16 – 24 year olds) boost sample on BCS, Done by screening at core sample addresses after main selection 40,440 addresses screened in Identified 3,211 eligible households (7.9%) Achieved 2,301 interviews – a response rate of 72%

18 Questionnaire development Questionnaire development starts about 6 months before start of new survey Home Office consultation to establish areas of interest and produce first draft Draft goes through various stages to produce pilot questionnaire Pilot questionnaire tested using cognitive testing Sometimes alternative piloting used where appropriate New questionnaire introduced in April each year

19 Structure of 2006/7 BCS Questionnaire (part 1) Household Box Screener Questions Victim Form(s) Maximum of 6 Main Questionnaire Fear of crime Perception of ASB Module A Contact with police Attitudes to police Module B Attitudes to CJS and sentencing Module C Crime Prevention/ Witnessing crime Module D Worry about crime Social cohesion Attitudes to CJS Mobile phone theft

20 Structure of 2006/7 BCS Questionnaire (part 1) Household Box Screener Questions Victim Form(s) Maximum of 6 Main Questionnaire Fear of crime Perception of ASB Module A Contact with police Attitudes to police Module B Attitudes to CJS and sentencing Module C Crime Prevention/ Witnessing crime Module D Worry about crime Social cohesion Attitudes to CJS Mobile phone theft 75% report no crimes 18% report 1 crime 5% report 2 crimes 2% report 3 or more crimes

21 Structure of 2006/7 BCS Questionnaire (part 2) Module A Contact with police Attitudes to police Module B Attitudes to CJS and sentencing Module C Crime Prevention/ Witnessing crime Module D Worry about crime Social cohesion Night time economy Technology crime Identity theft Town centre crime Anti social behaviour Teenagers Drunk and rowdy behaviour Anti social behaviour Teenagers Drunk and rowdy behaviour Demographics Drug use Inter-personal violence Domestic violence Sexual assault Stalking

22 Structure of 2006/7 BCS Questionnaire (part 2) Module A Contact with police Attitudes to police Module B Attitudes to CJS and sentencing Module C Crime Prevention/ Witnessing crime Module D Worry about crime Social cohesion Night time economy Technology crime Identity theft Town centre crime Anti social behaviour Teenagers Drunk and rowdy behaviour Anti social behaviour Teenagers Drunk and rowdy behaviour Demographics Drug use Inter-personal violence Domestic violence Sexual assault Stalking Done using CASI 95% acceptance, of which: 88% completed by respondent 12% completed by interviewer

23 Coding of survey data to offence codes Major coding exercise to classify all incidents to particular offence types (or out of scope) Team of office based coders use: Paper-based summary of each incident Modular based computerised questionnaire Not an automated process –needs individual judgement Uncertain codes and 5% quality check sent to Home Office for verification

24 UNCODED DATA Supervisor Check (5%) Home Office Check (5%) Yes No Yes Coder Outcome Supervisor Outcome Reformat data for Sending to Home Office Supervisor Coding Data sent to Home Office Data returned by Home Office Certain Uncertain Automatic referral Automatic referral Certain Uncertain Disagree FINAL OFFENCE CODE No Returned coding checked Agree BCS Offence Coding Data Management

25 Data processing and reporting Quarterly data processing and delivery cycle Delivery of data to Home Office 6 weeks after end of fieldwork period All interviews returned and verified All coding done SPSS data files created and checked Geo-demographic variables attached to data Design weights derived Data files delivered as 12 month rolling annual data

26 Data processing and reporting Home Office apply calibration weights to data Calculation of victimisation rates and other key estimates Publication of results within 12 weeks of end of fieldwork period Annual data files deposited in archive


Download ppt "Introduction to the British Crime Survey Keith Bolling, BMRB Social Research."

Similar presentations


Ads by Google