The Scottish Government

« Previous | Contents | Next »

Listen

ANNEX 2 METHODOLOGY

A3.1 Full details of the methodology used in this survey can be found in the Technical Report.

Sampling

A3.2 The 2006 SCVS, like previous sweeps of the survey, employed a probability sample with a multi-stage stratified design and addresses randomly selected from the Postcode Address File ( PAF). The stratifiers used in 2006 were population density and the proportion of adults aged 16-74 in non-manual occupations.

A3.3 Sample points were clustered within postcode sectors but were spread across the whole of Scotland (including the Highlands and larger islands).

A3.4 Since not all PAF addresses yield eligible addresses the first task of the interviewer at each issued address was to identify whether it was an eligible residential address. Examples of ineligible addresses include vacant properties, second homes, non-residential addresses where no-one is living such as shops or factories, and establishments where people were living communally.

A3.5 In the relatively infrequent cases where a PAF address generated more than one household, the interviewers in the field used a random selection method to select one.

A3.6 Once households had been selected, one respondent was randomly selected from all the adults aged 16+ who were resident at the address (using the Kish grid method ²²). No substitution of households or individual respondents was permitted.

Survey Development

A3.7 The SCVS questionnaire was based on that used in previous sweeps of the survey. Two pilots were carried out in May 2006 before fieldwork started in June. These pilots allowed the new questionnaire to be monitored and revisions were made where necessary.

A3.8 The questionnaire was divided into sections:

• Main Questionnaire (answered by all respondents):
  • Introductory questions on fear of crime and general social issues
  • Screening questions on experience of victimisation
• Victim Forms
  • The victim forms collected detailed information about a maximum of 5 incidents (or series of related incidents) reported in the screener section. Victim forms also covered other factors related to the incident including the impact of the offence on the respondents, whether the police were informed and what kind of help the respondent received.
• Follow Up Modules: respondents were randomly allocated to one of 2 follow up modules:
  • Follow Up A covered a number of topics including:
    • Perceptions of crime levels in local area
    • Perceptions and impact of anti-social behaviour in local area
    • Whether been insulted or pestered since 1st April 2005
    • Whether ever been victim of a violent attack
    • Whether been victim of workplace violence
    • Attitudes to sentencing
    • Personal safety.
  • Follow Up B focused primarily on experience of and attitudes towards the Scottish Criminal Justice System.
• Demographics
  • All respondents completed this module which collected key demographic information about the respondent and other members of the household.
• Self-completion modules
  • There were 2 self-completion modules covering i) drug usage and ii) experience of partner abuse. The drugs module was only asked of respondents aged less than 60 years of age whereas the partner abuse module was asked of all respondents. These modules were presented as computer assisted self-completion ( CASI) modules to ensure respondent confidentiality in answering these questions.
  • Findings from the 2 self-completion modules are presented in two separate reports: 'Drug misuse in Scotland: Findings from the 2006 Scottish Crime and Victimisation Survey' and 'Partner abuse in Scotland: Findings from the 2006 Scottish Crime and Victimisation Survey'.

Fieldwork

A3.9 Fieldwork ran from June to December 2006.

Survey Reference Period

A3.10 The survey collected details of every incident of crime that respondents had experienced since the 1st April 2005. However, the survey reference period on which the incidence and prevalence rates presented in this report are based is the 12 month period between the 1st April 2005 and the 31st March 2006.

Response Rate

A3.11 Table B1 provides a full breakdown of the final response rate. In total 7,854 addresses were issued to interviewers. One in ten addresses (10%) were found to be ineligible or 'deadwood' at which it would not have been possible to gain an interview. Interviews were achieved at 4,988 of the 7,076 eligible addresses giving a response rate of 70.5%.

Table B1 Response Rate

Offence Classification

A3.12 The purpose of the offence classification is to identify a single offence code for each Victim Form from which victimisation rates can be calculated and used for comparison with other statistics (e.g. police statistics and past crime surveys).

A3.13 The framework for offence classification had been developed over the series of crime surveys. Prior to the 1993 SCS, offence coding instructions consistent with previous British Crime Surveys were drawn up. These instructions were used for the SCS in 1996 and 2000 without significant amendments. In 2003 the definition of housebreaking was changed to mirror more accurately the Scottish police recorded crime definition of domestic housebreaking by including housebreakings to non-dwellings such as sheds, garages and outhouses. This same instruction was used for the 2006 SCVS.

A3.14 Offence coding was carried out by a specially trained team of coders using a computer assisted program. Offence coders were issued with a coding manual which contained precise definitions of each crime type. This manual was written jointly by BMRB and the Scottish Executive. Where there was any doubt about the appropriate offence code the victim form was sent to the Scottish Executive for coding. In addition 5% of victim forms were sent to the Scottish Executive for quality checking.

Weighting

A3.15 Separate weights were calculated for household based estimates and for individual based estimates. In addition, grossing factors were calculated to allow population incidence and prevalence national totals to be estimated.

A3.16 Household and individual weights were designed to address 2 types of bias:

• biases arising from variability in sample selection probabilities (the design components of the weights);
• biases arising from non-response (the non-response components of the weights).

Household weight

A3.17 The household weight comprised i) a design weight which was applied to compensate for the fact that where more than one dwelling unit was found at a selected address the interviewer had to select one randomly and ii) a non-response weight which was applied to compensate for any minor differences between the profile of households (in terms of household size and ACORN) at which an interview was achieved and the population profile (by Police Force Area). This weight was applied before any analysis was carried out at the household level (principally experience of household crimes).

Individual weight

A3.18 The individual weight was also comprised of a design weight and a non-response weight. The design weight was applied to compensate for differing numbers of dwelling units at selected addresses and differing numbers of eligible adults within selected dwelling units; both of which affect an individual's chance of being selected for the survey. With the design-weight applied, the achieved sample profile was then compared to the population profile (within each PFA in terms of ACORN and age/sex) and a non-response weight calculated to compensate for any remaining observable bias. This weight was applied where the analysis unit was the individual (such as when estimating experience of personal crimes or when analysing attitudinal data).

Incident weights

A3.19 A victim form can represent either one incident or a series of similar and related incidents. Forms that represent a single incident received a weight of 1. Forms for a series of incidents were weighted by the number of incidents they represented. As in previous sweeps of the survey the weight for the series victim forms was capped at 5 - partly to limit the impact on victimisation rates of outliers and partly to limit the impact of the weights on the survey standard errors.

A3.20 Because the victim forms refer to either a household (in the case of property crimes) or an individual (in the case of personal crimes) the Victim Form weight was multiplied by the household weight (Hhdwt) or the individual weight (Indwt) for the record to which the form related.

Grossing weights

A3.21 In addition to the basic design and corrective weights, the SCS requires weights that:

• express the data from the survey in terms of the number of households and adults in the population as a whole experiencing different types of incident
• express incidents as a rate per 10,000 units (households, adults, households with access to a vehicle, households with access to a bicycle)
• express victimisation as a rate per 10,000 units where the unit varies depending on the type of incident (households, adults, households with access to a vehicle, households with access to a bicycle).

A3.22 These weights are calculated by multiplying the individual or household weights by additional factors.

• Estimates of national incidence / prevalence rates of property crimes are grossed up to a population estimate of 2,286,660 households (2006 projection from 2001 Census data)
• Estimates of national incidence / prevalence rates of personal crimes are grossed up to a population estimate of 4,189,000 adults (aged 16+) (2006 projection from 2001 Census data)
• Estimates of national prevalence incidence / prevalence rates of motor vehicle incidents, are grossed up to a population estimate of 1,637,249 households with use of a motor vehicle (based on survey estimate)
• Estimates of national prevalence incidence / prevalence rates of bicycle theft incidents, are grossed up to a population estimate of 958,111 households with access to a bike.

« Previous | Contents | Next »