A SURVEY OF STUDENT DRUG USE
AT SECONDARY SCHOOLS
IN AMSTERDAM
F.M.H.M.Driessen
G.van Dam
Report prepared for the Co-operation Group to Combat Drug Abuse
and Illicit Trafficking in Drugs (Pompidou Group) of the
Council of Europe
Utrecht/Strasbourg, 1989
Bureau Driessen
Social Science Research and Advice
This research has been made possible by a grant of the Foundation forPreventive Medicine ("Preventiefonds") in The Hague to the Department ofSocial Medicin of the Free University in Amsterdam
Copyright (c) Bureau Driessen, Utrecht 1989.
CONTENTS
1. INTRODUCTION
2. DESIGN OF THE STUDY AND RESPONSE
2.1. Questionnaire
2.2.Response
3. METHODOLOGICAL RESULTS
3.1.Classical interrogation
3.2.Understanding the questions
3.3. Validity of the questions
3.3.1. Internal consistency
3.3.2. Consistency between different questions
3.3.3. Underestimation of prevalence due to
inconsistent answers
3.3.4. Answering honestly
3.4.Conclusion
4. SUBSTANTIVE RESULTS
4.1. Prevalence of drug use
4.2.Age of first use
4.3.Drug use on prescription
4.4.Percieved availability of drugs
5. CONCLUSION
Tables
Appendix (Dutch Questionnaire)
References
1
CHAPTER 1
INTRODUCTION
In order to forecast developments in drug abuse and treatment demand, to plan policies regarding drug abuse and to evaluate the effects of preventive measures, policymakers and public health officers need information on the prevalence of drug abuse among the population. Since drug use in adolescence is known to be an important risk factor for problematic use later on, developments among the adolescent part of the population are of especially great importance.
In several countries research on drug use among adolescents takes place on a more or less regular basis, but the results of these studies are not comparable because of great differences as to methodology and the design of the questionnaire. International comparability, however, would add to the usefulness of such studies.
For this reason the epidemiology expert group of Cooperation Group to Combat Drug abuse (the so-called Pompidou group) of the Council of Europe has developed a standardized questionnaire.
In the participating countries studies have been carried out to examine the validity of the questions, to optimise the international comparibility of the data and to test the chosen research design (classical interrogation).
In this report the results of such a study in Amsterdam will be presented. This report is a revised version of the report: F.M.H.M. Driessen & G. van Dam. Results of a drug use survey among students of secondary education in Amsterdam. Free University, Amsterdam/ Council of Europe, Strasbourg, 1987. The report has been adapted to the standardized format developed by the Pompidou group. Although in the original report all relevant results are enclosed, the following will facilitate international comparison of findings.
1
CHAPTER 2
DESIGN OF THE STUDY AND RESPONSE
The study was carried out in Amsterdam, because it is generally believed that the prevalence of drug abuse is higher in Amsterdam than anywhere else in the Netherlands. Therefore it can be assumed that estimates of prevalence for Amsterdam give upper limits of prevalence estimates for the Netherlands as a whole.
Five schools (in Amsterdam) were approached for the study. Among other things these five schools differ in: location and the type of education offered. See Table A.
From this table it is obvious that we managed to include schools from different neighbourhoods of Amsterdam, different in respect of the social composition of the neighbourhood and the distance to inner city. In the Netherlands in secundary education many different types of education exist, ranging from vocational training to schools preparing for university. In the sample all types of education are represented, although there are less students from vocational education than in the whole Amsterdam population, since we restricted ourselves -for organisational reasons- to municipal schools. Most vocational schools are not runned by the municipality, but by foundations. Although for this study, designed to test the questionnaire, it is not necessary to have a representative sample, we are confident that our sample approximates to it.
From these schools a number of classes were selected for the questioning, both higher and lower ones. In consultation with the schools, the pupils were handed a questionnaire during school-time. Filling-in the questionnaire took 45 minutes at the most. Introductory information was given to the pupils beforehand, during which incidental mention was made of the anonymity factor (but not too emphatic). Members of the teaching staff were not present in the classroom.
TABLE A. Schools included in the study.
───────────────────────────────────────────────────────────────
Location in Types of secondary Number of
Amsterdam education*Remarks students
in sample
───────────────────────────────────────────────────────────────
1near wealthy Highschool/
suburb, and Prepar.for Univ.171
near area with
appartment-
houses of the
1960s
2mainly post- Highschool/
war flats Prepar.for Univ.132
319th cent. Vocational/Highs.mainly lower
area near Lower Highsch/social class 99
inner city Lower Adm.
4area with Lower Administ./large propor-
flats of Lower Highschool/tion of stud.142
the 1960s Highschoolfrom former
colonies
5inner city Highschool/small school, 86
Prep.for Univ.only higher
grades (5,6)
───────────────────────────────────────────────────────────────
*Voc.=LHNO/LBO. Low.Adm.=LEAO. Low.Highsch.=MAVO. Highsch.= HAVO. Prepar.for Univ.=VWO.
The different classes of each school were, as far as possible, questioned simultaneously, in order to eliminate the possible influence of conversations between those pupils who had already been interviewed and those who had not yet.
Medical students carried out the field-work, supervised by a graduated social scientist. In 36 classes the questionnaire was submitted. The number of students in these classes ranged from 5 to 29. As the students in the higher classes in the schools that prepare for university are taking different courses, in some of these classes only a few students were present when the interview was held.
2.1. Questionnaire
On the one hand the questionnaire consisted of the core questions from the Pompidou-group. On the other hand, questions were included which are of significance in connection with the origins of problematic drug use and the identification of risk groups. These latter questions are, as much as possible, adopted from the optional-questions part of the Pompidou-group questionnaire, giving the possibility of comparison with other countries.
The questionnaire was preceded by an explanation, in which the usefulness of such research is briefly gone into, as well as the method of answering, the anonymity, the fact that there are no 'right' or 'wrong' answers and the importance of answering honestly. It was stated that any question a pupil thought to be unanswerable, should be skipped. This explanation helped to keep the verbal explanation at a minimum, in order to arise as few differences as possible in the interview situation between classes as a result of the verbal presentation. The appendix gives the complete questionnaire.
On the last page of the questionnaire the student could fill in his name and adress, to make possible a follow-up investigation. The students were asked to keep this (removable) last page strictly separate from the questionnaire. These pages were collected seperately from the questionnaires and put in a separate enveloppe. It was stated to the students that the names would be used for a possible follow-up and that they would never be used to make a connection between name and questionnaire. On the last page there was a number and on the questionnaire there was another number to make possible a connection between the data of this interview and those of a follow-up. Only a small minority of the students refused to fill in the last page (33 out of 632).
2.2. Response
The questionnaire was filled in by 632 students. Only one student refused to fill in the questionnaire. Two students obviously did not take their tasks seriously, giving impossible answers (e.g. age: 12; age of first use: 18), and making a lot of dirty remarks on the questionnaire. These two respondents were removed from the data-analysis.
At the moment of the interview 91 students were absent. For one class the number of absentees was not noted (see table B).
TABLE B. Sample and response
───────────────────────────────────────────────────────────────
N of students %
───────────────────────────────────────────────────────────────
Refusal 1 0.1%
Absent 9112.6%
Not filled in seriously 2 0.3%
Response-category63087.0%
───────────────────────────────────────────────────────────────
Total (target sample)724100%
───────────────────────────────────────────────────────────────
CHAPTER 3
METHODOLOGICAL RESULTS
3.1 Classical interrogation
The organisational advantage of classical interrogation is obvious, but some developments in the Dutch school-system make classical interrogation less suitable as a means of inquiry. Teachers stimulate the pupils to cooperate in small groups. In some classes the pupils are habituated to discuss things with each other and as a consequence they sometimes also discussed the questionnaire with each other. The fieldworkers were trained to avoid this sort of discussions, but it was not always possible to keep the class quiet. Especially in the lower classes discussions about the questionnaire were rather frequent, whereas the higher classes were almost all very silent.
Inspection of the questionnaires learned that the same kind of 'funny' remarks was epidemic in some classes, whereas patterns of answers differed markedly between the pupils. This suggest that the pupils filled in the questions seriously. Nevertheless it is possible that these discussions (in the lower classes) made it more difficult to answer honestly the drug prevalence questions. Therefore in future the teacher will be asked to stay in the class during the filling in of the questionnaires. Probably, he will be more capable to enforce silence in the classroom than the medical students, who we used as field workers.
3.2. Understanding the questions
The youngest students are 12 years of age (10% of the sample). Some of them are enrolled in types of education where cognitive skills are less important. Since the questionnaire looks rather complex and requires reading a lot of pages, we were very interested if those younger students in vocational types of education did understand the questions.
In the introduction on the first page of the questionnaire the possibility was stressed that some questions might be difficult to understand. In that case the student was asked to mark these questions by putting a scratch through the question. The in this way marked questions were counted and it turned out that the possibility to mark questions, that were not understood, was used only 4 times. Table C gives the figures for the prevalence questions.
Another indication of the understandibility of the questions may be found by counting the number of times a question is not filled in. It is possible that a pupil just skips a question, because he doesnot understand it, without marking it as such. As can be seen from table C most students filled in all the questions, at least suggesting they understood the questions. Especially questions about alcoholic beverages were more often sipped. Maybe some of the students did not understand that they had to fill in both the question on alcoholic beverages and
TABLE C.Understanding the questions. Number of times a question is marked 'not understood' or not filled in.
──────────────────────────────────────────────────────────────
Question not not
understoodfilled in remarksTotal
N NN %
───────────────────────────────────────────────────────────────
Prevalence of
- alcohol1 443 7.6
- beer0 291 4.8
- wine1 311 5.2
- liquor0 201 3.3
- drunkness0 130 2.1
- tranquilizer0 60 1.0
- amphetamines1 30 0.6
- narcotics0 20 0.3
- methadon0 30 0.5
- barbiturates0 20 0.3
- marijuana0 71 1.3
- cocaine0 10 0.2
- heroine0 30 0.5
- LSD, mescaline1 10 0.2
- inhalants0 40 0.6
───────────────────────────────────────────────────────────────
three more questions about beer, wine and liquor.
Moreover the questions on alcohol were the first prevalence questions, so the pupils probably needed some time to get used to this sort of questions. The seven skipped questions on marijuana maybe incorporate some hidden use.
Besides, most students were able to fill in the questionnaire within 45 minutes. Only a small minority (17 out of 630) did not succeed in reaching the end of the questionnaire and left open (the greater part of) the last two pages. In the higher grades (5, 6) most students performed their task within 20 minutes.
Overall, the students' reactions to the questionnaire were positive. If possible, the fieldworkers talked with the students afterwards and they made short reports on the remarks of the students. According to these reports most students thought the questionnaire interesting.
On the last page of the questionnaire there was a possibility to make remarks. Only three students remarked that they found it difficult to fill in the questionnaire.
All these different findings (number of questions marked as not understandable; number of skipped questions; number of students who did not complete the questionnaire; reactions of students afterwards; remarks made on the questionnaire) indicate that the questionnaire is very understandable.
3.3. Validity of the questions
The validity of the questions regarding the use of different sorts of drugs has been examined in the following ways:
-The internal consistency per question was looked into (use during one's life, during last 12 months, last 30 days).
-Consistency between different questions was examined (lifetime use and first use)
-Whether or not a lot of pupils admit that they would find it difficult to answer truthfully or would answer untruthfully, if they used drugs, was examined.
These analyses give indications of the validity of the questions.
3.3.1. Internal consistency
If the questions are valid, that is to say are filled in according to reality, the internal consistency must be very high. With the prevalence questions this means that lifetime-use will be higher than use during the past year and the latter will be higher than the use in the past 30 days. Table 15 contains the number of times this consistency did not occur. As can be seen, inconsistent answers are not very frequent. Only the questions on alcoholic beverages (the first prevalence questions in the questionnaire) produced inconsistencies.
3.3.2. Consistency between different questions
The prevalence of the use of drugs and alcohol can also, indirectly, be determined with the questions on first use. If a student indicates no use ever on the prevalence questions, but indicates first use on the age of, for instance 12, there is an inconsistency in answering pattern (and vice versa). Tabel 14 contains the counts of these inconsistencies. As can be seen, the questions on alcohol and drunkness produce a rather large number of inconsistent answers. As to alcohol there was a slight difference in the formulation of the questions. In the first-use question the clause 'more than a few sips' was added and this may explain the fact that fewer students indicated first use than prevalence.
Drunkenness was at the bottom of the list of the first-use question, and probably some students filled in 'never', because they did not read the questions thoroughly after marking so many zero's for all those substances.
Rates for illicit drugs inconsistency are all very low.
3.3.3.Underestimation of prevalence due to inconsistent answers
Although we showed that it is unlikely that inconsistency of answers reflects hidden drug use, it is still possible that these inconsistent answers hide some prevalence. For this reason we made an estimation of the prevalence, assuming that all inconsistent answers as well as all questions, which were not filled in, reflect a reluctance to anwer honestly. Of course this estimation procedure gives nothing more than a (not very realistic) upper limit.
As can be seen from table D, for marijuana this procedure results in a prevalence estimation that is only slightly different from the original one. Moreover it turns out that the difference between the prevalence estimation on the basis of the cleaned data is even smaller. Due to the small number to start with, the difference between the original estimation and the upper limit is, in relative terms, greater for amphetamines and heroine, but in percentpoints it is only 1.5 and 0.7.
TABLE D. Different estimations of prevalence in order to establish the possible effect of inconsistent answers and skipped questions on these estimates.
───────────────────────────────────────────────────────────────
estimation of prevalance
───────────────────────────────────────────────────
not cleaned cleanedupper-
Drugdatadata1limit2
% % %
───────────────────────────────────────────────────────────────
marijuana15.416.317.8
amphetamines 1.3 1.4 2.9
heroine 0.2 0.3 1.0
───────────────────────────────────────────────────────────────
1) Cleaned for inconsistent answers within questions. Differences in percentages with other tables are due to the fact that in this table missing values are not excluded, to facilitate comparison between columns.
2) All missing data and inconsistent answers are considered cases of use.
3.3.4. Answering honestly
Following Johnston (1986) three questions were included to establish the willingness to answer honestly. Of course the interpretation of the answers on these questions is precarious and it resembles the classical Greec dilemma regarding the liars from Crete. Why should a student, who hides his use of drugs not also lie about his honesty? According to Johnston these questions nevertheless give an indication of the honesty of categories of students. Table 16 contains the answers.
From this table it becomes clear that there exists a correlation between age and willingness to answer honestly. In the lower age-group there exists more reluctance to be honest. The data show that - for groups - there also exists a correlation between prevalence of use and willingness to be honest. It is easier to be honest about marijuana than about amphetamines or heroine. Obviously as the social controls are less strict, for instance as the influence of the parents becomes less strong, the students both use more often the drug in question and are more often honest about their use. This is in accordance with the findings of Johnston & Harrison (1986) in the U.S. If this correlation, on an aggregated level, between honesty and prevalence holds also true on an individual level, there seems to be no reason to worry about the validity of the questions, as far as honesty is concerned.
Of course other explanations are possible. Logically it is possible that prevalence of use is as high in lower grades as in the higher grades, but that the students in the higher grades are just more frank about it. On the other hand it is logically possible that students in the higher grades do lie about their willingness to be honest, whereas the students in the lower grades are honest about their reluctance to be honest. As we have to do here with an empirical problem instead of a logical one, both possibilities do not seem very plausible, because it is a well-known fact that drug abuse is less frequent in lower age-groups.