Scottish Government

THE 2000 SCOTTISH CRIME SURVEY: OVERVIEW REPORT

Appendix C - Sampling Error

A sample survey is a small-scale representation of the population from which it is drawn. Survey estimates of characteristics of the population are subject to imprecision due to both sampling and non-sampling errors. An example of a sampling error would be where an estimate of a household attribute might be inaccurate recall or the respondent did not wish to divulge details of an incident. Other sources of non-sampling error which are common to all surveys include systematic data errors introduced by interviewers and mistakes in classification of incidents. Other errors may arise during coding and processing of the data, or because of incomplete sampling frames and bias arising from non-response. These can be minimised by careful training and supervision of survey workers but can never be entirely eliminated.

It is possible to quantify the extent of sampling error from the information collected by the survey whereas to undertake a similar exercise for non-sampling errors would require follow-up work with the respondents. This appendix focuses only on sampling error in the SCS.

The estimates used in this report are based on a sample of the population in Scotland aged 16 or over living in private households and are likely to differ slightly from figures which would have been obtained had the whole population been interviewed. One measure of the likely difference is given by the standard error, which indicates the extent to which an estimate might have varied by chance because only a sample was interviewed. The chances are about 68 in 100 that a sample estimate will differ by less than one standard error from the figure which would have been obtained in a complete census of the population, and about 95 in 100 that the difference would be less than two standard errors.

The following tables show the standard errors for all crime rates produced by the 2000 SCS for central and southern Scotland only and for all Scotland. Simple random sample (SRS) errors are estimates of the sampling error when a straightforward random sample of the population is taken. The SCS did not involve a SRS, however, but used clustering, stratification, and weighting to produce a sample (see Appendix B). Complex standard errors (CSEs) must be calculated, therefore, to take account of these factors. The ratio of the CSE to the SRS standard error produces the design effect*.

Table C1 Victimisation Rates and Sampling Errors for 2000 SCS for Central and Southern Scotland

1. For violence, theft from the person, assault, robbery, other personal theft and all personal offences, rates are quoted per 10,000 adults. For acquisitive crime, vandalism, housebreaking, vehicle offences, bicycle theft, other household theft and all household offences, rates are quoted per 10,000 households.
2. For the distinction between crimes which are 'comparable with police' and 'other survey crimes', see Appendix F.
3. Source: 2000 SCS (central and southern Scotland, weighted data), n=4,271.

Table C2 Victimisation Rates and Sampling Errors for 2000 SCS for all Scotland

1. For violence, theft from the person, assault, robbery, other personal theft and all personal offences, rates are quoted per 10,000 adults. For acquisitive crime, vandalism, housebreaking, vehicle offences, bicycle theft, other household theft and all household offences, rates are quoted per 10,000 households.
2. For the distinction between crimes which are 'comparable with police' and 'other survey crimes', see Appendix F.
3. Source: 2000 SCS (all Scotland, weighted data).