« Previous | Contents | Next »
Listen
CHAPTER ONE INTRODUCTION
Background
1.1 This report presents the findings on self-reported drug use from the 2006 Scottish Crime and Victimisation Survey ( SCVS).
1.2 One of the main aims of the SCVS is to measure people's experiences of crime in Scotland by asking a random sample of adults whether they have been the victim of any personal or household crimes in the last 12 months. Interviewing for the 2006 survey was carried out during the second half of 2006 and asked people about their experiences of crime between the 1st April 2005 and the 31st March 2006. The measures of crime which the survey produces are intended to be complementary to police recorded crime statistics. Results from the main survey are published as a separate report.
1.3 At the end of the interview there were 2 separate self-completion modules, where respondents were asked about their drug use and about their experiences of domestic abuse. Questions on these topics are especially sensitive and it is generally accepted that allowing respondents to complete the questions in private through a self-completion questionnaire is likely to lead to more honest answers compared with having interviewers verbally ask such questions.
1.4 The self-completion modules have been a feature of previous Scottish Crime Surveys, with the modules being largely unchanged between 1993 and 2004. However, the 2006 SCVS saw a significant methodological shift, with a move from using a paper self-completion questionnaire to using Computer Assisted Self Interviewing ( CASI). This change has considerably implications for trend analysis.
1.5 The main aim of the drugs self-completion module was to establish whether people aged 16-59 had taken any of 16 specified drugs either at any point in their lives, in the last year or in the last month. Those who had taken any types of drugs were asked a number of follow-up questions, primarily about frequency of usage.
Methodology
1.6 The 2006 SCVS was the first time that Computer Assisted Personal Interviewing ( CAPI) had been used. All previous sweeps of the survey, including the self-completion modules, were carried out using Paper and Pen Interviewing ( PAPI).
1.7 In the 2006 survey, once the interviewer had completed the main part of the interview on the computer, respondents were asked to complete the drugs module by themselves. Interviewers gave respondents a brief explanation of how to use the laptop, including taking them through some practice questions, before handing the laptop over. However, interviewers were always present to help respondents if they needed any technical assistance. Once respondents had completed the modules, they handed the laptop back to the interviewer. While respondents were encouraged to use the computer themselves, if they did not want to use it for some reason, interviewers were allowed to administer the modules provided that no-one else was present in the room.
1.8 In 2006, 96 per cent of respondents aged 16-59 who had already completed the main part of the survey agreed to do the drugs self-completion module, although 21 per cent of interviews involved the interviewer entering the answers for the respondent.
1.9 It was recognised that switching from a paper methodology to a CAPI methodology could have an impact on the ability to compare the 2006 results with previous sweeps of the survey, and in particular, that it could affect the proportion of respondents who admitted to taking drugs. In fact, in the 2006 SCVS, just under 4 in 10 respondents aged 16-59 (37%) admitted that they had taken an illegal drug at some point in their lives. This compares with just less than a quarter of respondents aged 16-59 (24%) who admitted taking an illegal drug in the 2004 survey, which was carried out on paper. Since it seems unlikely that drug misuse in Scotland has increased so rapidly in the space of 2 years, this suggests that the change in methodology had an impact on the reporting of drug prevalence levels. In fact, the impact of a " CAPI effect" has been noted on other surveys where a similar shift from paper to computer has taken place.
1.10 Increases of a similar magnitude in the prevalence of drug taking were found when the British Crime Survey switched from a PAPI to CAPI methodology in 1994 1. Mayhew (1995) suggested three reasons why this increase could have happened, all of which apply equally to the SCVS2.
1.11 First, although the questions used in the 2006 survey were similar to the questions asked on previous sweeps, the way in which they were presented to the respondent was significantly different. On the self-completion paper questionnaire used up until 2004, respondents were presented with a list of all 16 drugs and were asked to tick from the list what drugs they had ever taken, taken in the last 12 months, or taken in the last month. By contrast, the 2006 CAPI questionnaire presented each of the 16 drugs to respondents as a separate question and they were required to give a 'Yes/No' answer to each in turn. Questions were filtered so that respondents only received questions about the last year if they said they had ever taken a particular drug. Similarly, they were only asked questions about the last month if they reported they had taken a drug in the last year. Therefore it seems likely that the CAPI methodology leads to more complete and accurate responses by the way in which the questions are presented.
1.12 Second, a switch from paper to CAPI tends to reduce levels of non-response, both to the overall acceptance of the self-completion, and to levels of item non-response within the modules. In the 2004 sweep of the survey, 88 per cent of respondents aged 16 to 59 completed the drugs self-completion section of the interview compared with 96 per cent of respondents in 2006. It could be hypothesised that those who had taken drugs would be more likely to refuse to complete a questionnaire about drug misuse, as they might not wish to admit to this behaviour. Consequently, an increased response rate may, in itself, lead to an increase in the proportion of people who admit to taking drugs.
1.13 Finally, it is possible that entering responses directly into the computer increases respondents' trust in the confidentiality of their answers and that the technical formality of a computer screen prompts more thoughtful and honest answers. An increase in the proportion of people willing to admit to sensitive behaviours is fairly common when switching from a paper format to a computerised format. For example when the Home Office's Youth Lifestyle Survey made this switch the proportion of youths willing to admit to having committed offences increased significantly 3.
1.14 This difference in methodology means that it is not possible to make any meaningful comparisons between results from the 2006 SCVS and those from previous sweeps of the survey. However, the Scottish Crime and Justice Survey which replaces the SCVS from 2007 onwards will continue to employ a CAPI methodology, so that data collected in future sweeps of the survey will be comparable with the 2006 data.
1.15 A fuller description of the 2006 SCVS methodology is provided in Annex 2.
Limitations of the data
1.16 Although self-reporting drug surveys are valuable in providing information on a sensitive topic where there are few other sources of available statistics about the population as a whole, it is recognised that such surveys do have limitations.
1.17 Firstly, it is likely that there will be an under-representation of the most serious, persistent or 'chaotic' drug users. This will partly be due to the fact that some groups of drug users live in accommodation not covered by a survey of private households such as the SCVS. Examples of types of accommodation that are not included in the survey include hostels, prisons, and student halls of residence. Under representation will also partly be due to the fact that even where such drug users do live in private households, they are likely to be difficult to interview. For example, if they lead busy or chaotic lives they may hardly ever be at home and so cannot be interviewed during the fieldwork period.
1.18 Secondly, despite the switch to a CAPI methodology, it is likely there will still be a certain amount of under-reporting. Drug misuse is an illegal activity and as such some respondents may be uncomfortable with admitting to taking illicit drugs, despite reassurances about confidentiality and anonymity.
1.19 Thirdly, it is possible that some respondents may simply forget occasional uses of a certain drug, particularly if they last took the drug a long time ago.
1.20 While under-reporting of drug use on surveys such as the SCVS are almost certain, it should be noted that the issues discussed above are unlikely to apply equally across all types of drugs. While a household survey such as the SCVS is likely to be a fairly effective measure of the more commonly used drugs, such as cannabis, where the majority of users live in private households, it is likely to be less effective for some of the Class A drugs such as opiates or crack cocaine, where a sizeable number of users may be concentrated in small sub-groups of the population not covered by the survey.
1.21 While under reporting is by far the main concern on a survey such as this, it is also recognised that some people may report taking particular drugs when they have not actually done so for reasons of 'bravado' or other reasons. To try and counter this mis-reporting a bogus drug (semeron) was included in the list of drugs presented to respondents. Including the name of a fictitious drug is a technique that is commonly used in drug surveys. However, in the 2006 survey there was little evidence of over-reporting as only 3 respondents reported that they had 'ever' used the bogus drug and no respondents reported using it in the last year or last month. These 3 respondents were excluded from the analysis.
A note on interpretation of results
1.21 Unless otherwise stated, all data presented in this report has been weighted. The unweighted base sizes have been provided beneath all tables and figures.
1.22 As a general rule, only results that are statistically different at the 95% confidence level are commented upon in the text. In other words, we can be 95% sure that any differences observed are not the result of chance. Statistical tests such as 2-tailed t-tests were carried out where appropriate.
1.23 Within the main body of the report most results are rounded to whole numbers, while in the figures and tables results are presented to one decimal place. Where no responses were recorded the '-' has been used. Where less than 0.5 per cent of respondents gave a particular answer the '*' symbol has been used.
1.24 The drugs module was only completed by respondents aged 16-59. 'All' (in base text etc.) refers to all respondents aged 16-59.
Structure of the report
1.25 The report is structured as follows. Chapter 2 focuses on the prevalence of drug use. Chapter 3 looks at the experience of drug users including the age at which they first took drugs and the drug first taken; regularity of use; and polydrug use.
« Previous | Contents | Next »