This chapter outlines the general methodology of the study, the research design, research instruments, and the interviewing procedures used. As this is a longitudinal study involving the follow-up of individuals interviewed and assessed ten years previously, the procedures used to locate the sample are also described.
The initial phase of this study was carried out in 1990 (Fitzgerald & Jeffers, 1994). In the original study, 2029 primary-school children in the Dublin area were assessed on a number of social, psychological and educational dimensions. Behaviour was measured using the Rutter B2 Questionnaire (Rutter 1967), completed by the teacher for each child, and IQ calculations were based on a non-verbal test, the Standard Progressive Matrices (Raven, Court & Raven, 1983). A sample of 185 children was extracted from this population and these children and their families were analysed in greater detail. In each family standardised instruments were used to assess the psychological health status of the child and his/her mother. The health status of the children was assessed using the Isle of Wight Parental interview on Child's Psychiatric State (Rutter & Graham, 1966). The mothers were administered the Clinical Psychiatric Interview (Goldberg, Cooper, Eastwood, Kenward & Shephard, 1970) and the Malaise Inventory (Rutter, Tizard & Whitmore, 1970). Information was also gathered on a range of social and economic items.
Sixteen percent of the children studied had a definite psychiatric diagnosis and this was related to maternal mental health and role satisfaction, socio-economic factors and the parents' marital status. A child was twice as likely to receive a diagnosis if his/her father was unemployed and children living with married parents were less likely to be disordered than those whose parents were separated. A higher IQ score also appeared to be protective for the child in terms of developing disorder. One-third of mothers were categorised as suffering from psychiatric disorder and this was associated with economic disadvantage, dissatisfaction with domestic roles and lack of social and recreational opportunities.
2.2 Ten Year Follow-up of Children and their Families: Aims of the Study
The primary aim of the study was to investigate psycho-social outcomes in a cohort of children (n=185) first studied ten years previously. Another objective was to delineate those factors (both protective and potentially threatening) in a person's environment which best predicted negative or positive psychosocial outcomes in adulthood. The specific research aims were:
1. To examine social and psychological outcomes in a sample of 185 young people. The main outcomes of interest were:
- Mental and physical health
- Educational attainment and employment status
- Drug and alcohol usage
- Involvement in crime
2. To investigate if certain factors, identified in Phase One of the study, were associated with these outcomes. These included:
- Presence or absence of psychiatric disorder during childhood
- Presence or absence of psychiatric disorder in the mother or father
- The child's IQ score
- The social class and economic circumstances of the family
- The marital status of the parents
2.3 Research Design
This study employed a longitudinal follow-up design, which involved the re-examination of a cohort of children and their mothers first assessed in 1990 across a variety of social and psychological dimensions. In 2000, these children and their mothers were located and, if they consented, re-interviewed. In the baseline study, the mother was the key informant while in the follow-up enquiry, the young person was the key informant although the mother was also interviewed.
2.4 Method of Data Collection
Data was collected from each key informant and his/her mother during a structured interview using mainly standardised instruments. In addition to basic demographic characteristics, information was collected from the young person on:
- Psychological and physical health (including alcohol consumption and drug use)
- Education and employment status
- Social networks and relationships
- Self-esteem and Locus of Control
- Contact with the Law
The mothers' schedules included :
- Psychological health (self and family)
- Socio-economic and housing circumstances
- Marital relationship and parenting
- Social networks and relationships
- Alcohol, substance abuse and contact with the Law
2.5 Research Instruments
Five main research instruments were used and these are described in detail in Appendix One. The SCID (Structured Clinical Interview for DSM-IV Axis Diagnosis) (First, Spitzer, Gibbon & Williams, 1996) was used to assess psychological functioning in both the key respondent and the mother
1 . The Beck Scale for Suicide Ideation (Beck & Steer, 1991) was employed to ascertain suicidal ideation and behaviour. Self-esteem was assessed using an instrument devised by Rosenberg (1965), the Rosenberg Self-Esteem Scale. The Arizona Social Support Interview schedule, devised by Barrera (1980), was administered to measure various indices of social support. Locus of Control was measured using a scale devised by Pearlin, Menaghan, Lieberman & Mullan (1981). The remaining areas of interest (employment, educational attainment, physical health, alcohol and substance misuse, criminal activity etc) were examined using a questionnaire devised for the study by the authors. A similar instrument was developed for the mothers' interview, which included questions relating to their marital and parenting role, and to the health and social and economic situation of their husbands (where relevant) and other children.
1 Training in the use of this instrument was provided by a psychiatrist (MF) who also monitored its use. The results were scrutinised by him and, independently, by another psychiatrist.
2.6 Pilot Interviews
Before carrying out the main interviews, a number of pilot sessions were carried out. The purpose of the pilot procedure was to examine and refine questions and question order, as well as to test the acceptability of the interview content. Pilot interviews were carried out among a sample of young people and older women, matched for age with the participants in the study. Some minor modifications were made to the schedules following these interviews.
2.7 Locating the Sample
The aim was to locate and interview as many individuals as possible who had participated in the first part of the study in 1990. As the young person was now the key respondent efforts were concentrated on trying to locate him/her. The mother was only interviewed if the key informant agreed to be interviewed. Preliminary checks were made, using the Telephone and Thom's Directories, to determine whether individual families were still living at the same address. Once residence was satisfactorily established according to these criteria, key informants were contacted by letter in which the purpose and scope of the study were detailed and they were requested to contact the research team by phone or letter. A relatively small number of participants (n=19) responded to this initial contact. When no reply was forthcoming after a specified time period, participants were contacted by telephone (n = 30) or by calling to the house (n = 78). Various tracing methods were used for those families which had moved house. Neighbours proved to be a very useful source of information and twenty-six families were located in this way. The remaining untraced families (n=32) were then cross-checked with tenants on the Dublin Corporation housing register. If a match was found the Corporation forwarded a letter about the study. The researchers had negotiated this arrangement with Dublin Corporation's Housing department in order to protect the tenants' privacy. This method yielded no respondents.
All interviews took place between April and December 2000. In almost all cases, the interview took place in the person's home and was carried out with only the participant and interviewer present. At the start of the interview a short summary of the study was provided and the confidentiality of the process emphasised
2. The entire interview was conducted using a laptop computer. The researcher read the questions aloud to the participant and his/her answers were entered immediately. This method of administering the schedules presented no problems and respondents appeared to be quite at ease with the procedure.
2 Both the key informant and the mother were assured that information would not be divulged to either side.
2.8.1 Response Rate
Table 2.1: Details of participation and non-participation rates
A breakdown of the response rate is shown in Table 2.1 above. Participation and non-participation rates are reported both as percentages of the total sample and of the sample of families that were located. Of the original sample of 185 families, 83% (153 families) were contacted. Thirty-two families were not traceable. Of the 153 families who were contacted, 97 (63%) of the young people agreed to take part. This figure represents 52% of the total sample. Over a fifth (21%) of the total sample of young people refused to participate. Three people had died. Two of these deaths were drug related and the third a possible suicide. Eight had emigrated and six were no longer in contact with their families. Overall, the response rate compares well to other, similar, studies (see, for example, Dalgard, Bjork & Tambs (1995) who obtained a 50% response rate, and Serbin et al. (1998) who obtained a 47% response rate.
In addition to the 97 young people, 80 mothers were also interviewed. The discrepancy between the number of key respondents and mothers who were interviewed was due to a number of factors. Twelve mothers refused to take part, one was living abroad and could not be interviewed, and two young people requested the researchers not to contact their mother. In the remaining two cases, the mothers' whereabouts were unknown to the young people.
2.8.2 Respondents and Non-Respondents
Based on data collected in from Phase One of the study respondents and non-respondents were compared in terms of gender, IQ score, behavioural deviance, psychiatric symptomatology, maternal mental state, family history of mental illness, economic background and employment status of the father. This analysis revealed a statistically significant difference between the two groups in relation to behavioural deviance (Rutter B2 Scale). Fifty-eight percent of non-participants had been categorised as behaviourally deviant compared with one-third of Phase Two respondents ( = 10.662, df = 1, p < 0.01). There was no significant difference between the two groups in relation to psychiatric diagnosis or the other variables analysed.
A further analysis was carried out to discriminate between those who refused to participate and the untraced group. There were no differences between those who participated and those who refused, on any of the variables investigated. There was, however, a significant difference between participants and the untraced group in terms of psychiatric diagnosis at Phase One ( = 8.949, df = 1, p < 0.01). Specifically, 59% of non-contacts received a psychiatric diagnosis at Phase One, compared with 30% of those interviewed. Additionally, 63% of non-contacts were categorised as behaviourally deviant during Phase One, compared with one third of Phase Two participants and this difference was also significant ( = 8.063, df = 1, p < 0.01).
These findings would ordinarily predict a higher level of negative outcomes amongst the non-contact . To explore this possibility we took the investigation a step further. The names of all Phase One participants were checked against Garda files for a history of juvenile offending and this revealed important information about the various groupings. Only six (6%) of those interviewed had a recorded juvenile offence in contrast to 21% of the non-participants (i.e. refusals and non contacts together) and this difference was significant ( = 8.486, df = 1, p < 0.01). More crucially perhaps, juvenile offences were much more common among the untraced group. Almost 29% had a juvenile record, compared with just 13% for those who refused to participate. Differences between study participants and refusals were non significant but a significant difference was evident between participants and untraced individuals ( = 13.403, df = 1, p < 0.01). These findings are illustrated in Figure 2.1 and have important implications for the overall investigation.
Figure 2.1: Record of juvenile offence among participants, refusals and non-contacts.
2.9 Statistical Analysis
Examination of SCID results was carried out according to its customised computer programme and the remainder of the data was downloaded onto the SPSS package for analysis. Modelling techniques were used to test the interrelations of key variables.
This section outlined the methodology of the study, a longitudinal analysis involving the follow-up of individuals initially interviewed ten years previously. The main aims of the study were to examine outcomes for the sample of young people who participated (n=97) and to determine if factors identified in Phase One were associated with these outcomes. Outcome measures which included psychological health and well-being, contact with the law, and educational and economic status, were assessed using standardised instruments. Of the original sample (n=185), 153 young people and their mothers were located. Of these 63% agreed to participate and 26% refused. A comparison of participants and non-participants revealed a significant difference between the groups in relation to childhood psychological and behavioural disorder. An examination of Garda records revealed that considerably more of the untraced individuals (but not the refusals) had a juvenile record for offending.
Structured Clinical Interview for DSM-IV Axis I Diagnoses (SCID)
Beck Scale for Suicide Ideation (BSSI)
Rosenberg's Self-esteem Scale
Arizona Social Support Interview Schedule (ASSIS)
Locus of Control
Frequency of SCID Diagnostic Categories
Search for your local office