|
Methods
Full details of methods are given in the
Technical Report which has been made available to ESRC along with the
report of findings.
The core of the study is a cohort of
4,300 young people aged 11½ to 12½ at the start of fieldwork in the autumn
of 1998, when they entered secondary school. Essentially the cohort
consists of all the young people in the City of Edinburgh in the relevant
age group. The strategy is to collect information from multiple sources
about all members of the cohort once a year. At each sweep the period
covered is the previous 12 months, so that the study provides a continuous
account of events in the lives of the cohort, and not just an account of
selected time segments. The advantages of this design, which focuses on
the largest possible number of young people within a single city, are
discussed in the paper by Smith and McVie that is submitted with this
report. Closely integrated with the cohort study is a parallel study of
social geography and crime patterns in Edinburgh, which primarily makes
use of data from the 1991 census and police-recorded crime data. This
makes it possible to analyse the findings for cohort members in the light
of the characteristics of the neighbourhoods where they live.
At the first two sweeps, information
about individual cohort members was collected from questionnaires normally
completed in the classroom; from the files of the City of Edinburgh Social
Work Department; from the files of the Scottish Children’s Hearing
Reporters Administration; and from school records. At the second sweep,
pastoral teachers completed a short questionnaire about each cohort
member.
The first, major, exercise in mapping
the social geography of Edinburgh made use of 1991 census data and
police-recorded crime data for 1997.
We carried out case studies of two
Edinburgh neighbourhoods with similar social composition but contrasting
crime patterns.
Semi-structured interviews were
conducted with 40 young people selected from the cohort.
An initial, detailed analysis of
findings from the first two sweeps was carried out. Multivariate analysis
is currently in progress.
Ethics and data
protection
It is important to ensure that the study
is carried out with the informed consent of parents and children, and that
no child or parent can be harmed by taking part. The Data Protection Acts
prevent a court from requiring the researchers to hand over information
obtained solely for the purposes of research, for example to aid a
criminal investigation. This means that we are able to offer children and
their parents an absolute assurance of confidentiality, with one
qualification. This arises because the child protection guidelines of the
education department of the City of Edinburgh state that any disclosure of
physical or sexual abuse must be reported to the appropriate school
authorities. For that reason, no direct questions about abuse by adults
will be asked until respondents reach the age of 16. On the rare occasions
(two in the first three years) when respondents provide unsolicited
information, researchers report the matter to a pastoral teacher, and also
encourage the child to report it.
At the beginning of the study, a
two-page letter from the researchers with a covering letter from the head
teacher was posted by schools to parents. The letter set out the purposes
of the study, and clearly described each of its elements. Parents who
wished to withdraw their children from the study were invited to return a
tear-off slip. On the first occasion that children were asked to complete
a questionnaire, they were given a single sheet summarising the purposes
of the study, describing the main elements, and explaining how the
confidentiality of the information would be maintained. At this point, and
on all subsequent occasions, children could decline to participate.
Parents are sent a newsletter about the study about once a year, which in
future will contain a summary of key findings. Parents are always given
the opportunity to withdraw if any new element is added to the
study.
Advisory Group
An Advisory Group was established under
the Chairmanship of Professor Sir Michael Rutter of the Institute of
Psychiatry, London, and met for the first time on 26 May 1998. The members
comprise senior representatives from all the agencies involved in the
study, including education, police, social work, the children’s reporter,
the Scottish Executive, the Home Office, independent schools, state
schools, and parents, together with several academics and practitioners
with an interest or involvement in research into crime and young people.
The Advisory Group meets formally once a year, but is regularly kept
informed of progress and approached for advice at key stages of the
study.
Access and relationships with other
organisations
The research can only be carried out
with the close co-operation of a range of organisations, including the
City of Edinburgh Department of Education and individual state schools,
independent schools in Edinburgh, the city’s Department of Social work,
the Scottish Children’s Hearing Administration and its Edinburgh office,
and the Lothian and Borders Police. Our initial, formal approach to the
education department was followed by a series of presentations to the
Director of Education, the Convener of the Education Committee, head
teachers, and the Parents’ Consultative Committee. In February 1998 the
Education Committee, and shortly after the full Council, agreed in
principle that the Edinburgh schools could participate, although final
agreement had to be sought from individual schools. Separate approaches
were then made to the other organisations, and to the individual
independent schools. The study has received a remarkable degree of support
and active participation from these organisations. The only small
qualification is that a few independent schools refused to participate
(see below).
Rates of participation and
response
A total of 49 schools in Edinburgh were
approached to take part in the study, including mainstream state schools,
special needs schools, and independent schools. All of the mainstream
schools agreed to take part, 9 of the 12 special needs schools, and 8 of
the 14 independent schools. Of all independent school pupils, 63 per cent
were in participating schools (at sweep 1), and 75 per cent of special
needs pupils were in participating schools. Overall, 92 per cent of
Edinburgh pupils at sweep 1 were in participating schools. Although new
pupils were allowed into the cohort at sweep 2, these percentages remained
very similar.
Parents were given the opportunity to
withdraw their children from the study, and the young people themselves
could decline to participate. Of all children in the participating
schools, 3.5 per cent were withdrawn or opted out at sweep 1, and this
proportion declined slightly at sweep 2. The proportion opted out was the
same at mainstream state and independent schools (but as expected, higher
at special needs schools), which suggests there was no systematic bias
according to social class. Questionnaires were completed at both sweeps
for nearly all (over 99 per cent) of those in scope (who had not been
withdrawn by their parents or opted out).
Questionnaire development and
piloting
The sweep 1 questionnaire was developed
after a systematic trawl of the literature and collation of questionnaires
used in current and recent longitudinal studies focusing on crime. Very
close attention was given to making the questions understandable and
interesting for the age group. There were several stages of pilot work at
both sweeps, carried out at schools outside Edinburgh (so that members of
the cohort would not be involved). This moved from qualitative and open
methods, through trials of short sections, to a full-scale pilot of the
whole procedure.
Fieldwork in schools
Most children filled in the
questionnaires in the classroom, under the close supervision of one or two
researchers. All children were carefully briefed about the purposes and
methods of the study before they began, and had the opportunity to
withdraw. Children with reading or writing difficulties were given help
adapted to the level of the difficulties. This usually involved reading
the questions out to them. In the most difficult cases, they were
interviewed on a one-to-one basis. Recalls were made to find children not
present on the first occasion. Where children were still not present after
several recalls, researchers made efforts to contact them at
home.
Teachers’
questionnaire
The teachers’ questionnaire consisted of
a shortened version of the Goodman Strengths and Difficulties Scale.
Teachers’ questionnaires were completed during the autumn term of 1999 for
94.2 per cent of members of the cohort included in sweep 2.
Social work and children’s hearing
records
Monitoring forms were designed so that
relevant information could be transferred for every member of the cohort
who had had contact with the social work department, and, separately, with
the children’s hearing system. A total of 468 cohort members were
identified as potentially having a social work record by the department’s
central computer system. Files were found for 363 of these young people
(further details in the Technical Report). It is likely that most of the
remainder were either incorrectly shown as having a file, or had had
little contact overall, and no recent contact. Children’s hearing records
were identified for 374 members of the cohort, and 356 of these files were
successfully located. At sweep 1, information was collected about contacts
with the agencies at any time up to August 1998. At sweep 2, information
was collected about contacts during the 12 months ending August
1999.
Personal interviews
A total of 40 semi-structured interviews
were carried out in the summer of 2000 with a sample of boys and girls
selected from the cohort to include a larger group of frequent offenders
and a smaller group with low offending scores. Interviews were
tape-recorded and transcribed in full.
Mapping social geography and crime
patterns
Mapping the social geography and crime
patterns of Edinburgh was a large undertaking which involved bringing
together data from a number of sources and processing it in a variety of
ways (see Chapter 14 of the report of findings). As one outcome of this
process, Edinburgh was divided into 91 neighbourhoods, choosing the
boundaries so as to maximise the homogeneity of each neighbourhood, and
hence the contrasts between them. The main criterion used was an index of
social and economic stress derived from six census variables. Post-codes
were obtained for the home addresses of most cohort members, so that each
one could be assigned to a neighbourhood. From this point on, there were
two datasets: one containing 91 cases (neighbourhoods); the other
containing 4,300 cases (cohort members). Information could be exchanged
between the datasets. Information about individuals could be analysed by
reference to characteristics of their neighbourhood, and neighbourhoods
could be characterised according to information derived from cohort
members living in them.
Analysis
The first stage of the analysis is
described in the report of findings. This involved deriving a large number
of scores to summarise results across groups of questions. Next, basic
results were described in detail, with an emphasis on examining
relationships between a wide range of variables and self-reported
offending. The pattern of relationships was found to be very different,
depending on whether the analysis was at the individual or at the
neighbourhood level. An extensive programme of multivariate modelling is
now starting. This will focus on attempting to predict, or explain,
self-reported offending, victimisation (including bullying), and
adversarial police contact. Explaining change from sweep 1 to sweep 2 will
also be a major objective. There will be an intensive examination of the
relative importance of neighbourhood and individual characteristics in
explaining both offending and
victimisation. | 