< Previous | Contents | Next >

DISCIPLINING CHILDREN: RESEARCH WITH PARENTS IN SCOTLAND

Appendix A - Research methods

Qualitative component

The purpose of the qualitative component of the project was to provide both breadth and depth to the research: breadth in terms of the context in which the disciplining of children takes place, and depth in terms of the meaning the issues have for parents. It was decided that the qualitative work should precede the survey element in order to ensure that the questionnaire comprehensively covered the relevant issues and that it was couched in appropriate language.

In recent years, focus groups have emerged as perhaps the most common method of exploring public attitudes within applied policy research. This approach certainly has important advantages — for example, some of the tension of the interviewer-respondent dynamic is dissolved and participants often challenge, develop or confirm each other’s contributions. However, it can sometimes have significant limitations in relation to sensitive topics, as the group dynamic may inhibit as well as encourage openness in relation to some issues. Parenting can undoubtedly be a sensitive topic and we therefore decided to use a variety of qualitative interview methods as we anticipated that different aspects of the issues might emerge in different settings. The research therefore combined the use of focus groups with individual, paired depth and peer group interviews. Twenty interviews (five of each type) were conducted between March and April 2002.

Conventional focus groups

These consisted of 6-8 individuals, recruited using conventional face-to-face methods, and not known to each other in advance of the group. The aim was to achieve a high degree of homogeneity within each group (through segmentation based on age of child, sex and social class) but heterogeneity across the group sample as a whole.

Paired depth interviews with couples

We felt that there would be value in using paired interviews with couples, since there are often important gender divisions within households in terms of how issues of discipline are handled. In addition, the couples interviewed together were able to reflect on, build on and challenge each other’s accounts.

Individual interviews with mothers/fathers

Not all children are raised by two parents, of course, and there are particular issues for single parents that can be best explored in the context of an individual interview. Moreover, even within ‘couple’ households, some parents feel more comfortable discussing such issues without their spouse or partner being present.

Peer group interviews

An alternative to conventional focus group work, where participants are typically not known to each other in advance, is to convene ‘peer groups’. For each peer group interview, we recruited a parent and ask them to recruit close friends or relatives with children of the same age. The advantage of this approach was that there was a certain level of existing trust within the group. It was also anticipated that participants’ awareness of each other’s lives would act as a check on the misrepresentation of views or behaviours.

All participants were recruited from the members of the general public by NFO System Three interviewers. The focus groups and peer groups took place in neutral venues such as a room in a pub or hotel. Most of the individual and couple interviews were conducted in the participants’ own homes (one was conducted in a hotel). In keeping with standard research practice for work of this kind, all respondents were paid a small financial incentive to attend. The aim was to encourage participation by those who might not otherwise take part and to offset any childcare or travel costs incurred.

The structure of the sample is detailed below.

*Other than for the interviews with couples, there was no restriction on whether parents were single parents or not. However, to ensure that single parents were included, we selected a single parent as the "lead" member of one of the peer groups and for one of the individual interviews.

The discussion moderators/interviewers were Simon Anderson, Lorraine Murray and Chris Martin of NFO System Three Social Research and Dr Julie Brownlie of Stirling University. The general content and flow of the discussions were dictated by the participants themselves, within the framework of a series of topics introduced by the moderator/interviewer. The topics included:

• General orientation towards young people and issues of discipline
• Own experience of parental discipline as a child
• Assessment of the effectiveness and impact of that experience
• Perceptions of the prevalence of use of different forms of physical chastisement among friends and peers, and in society at large
• Perceptions of the relative effectiveness and impact of different forms of physical and non-physical chastisement
• Situations/circumstances in which different forms of physical chastisement considered acceptable and unacceptable
• Use of different forms of physical and non-physical chastisement in relation to own children
• Attitudes towards the role of the state and the appropriateness of legislation in relation to various forms of chastisement

A variety of techniques were used to try to desensitise the issue and normalise particular attitudes and behaviours. These included asking people to reflect on their own experiences as a child, asking about what they thought of other people’s attitudes and behaviours. We also anticipated using scenarios to ask how people thought they would be likely to react in particular circumstances. In practice, we found that these were rarely necessary as most participants were very willing to talk openly about the forms of discipline they had used in different situations.

With the consent of participants, the discussions were tape-recorded for subsequent transcription. NUD*IST N5 (a qualitative analysis software package) was used to organise, code, search and retrieve the text.

Quantitative component

The quantitative element of the study consisted of a nationally-representative, probability sample of parents in Scotland. In total, 692 interviews were completed during the period May-August 2002. All interviewing took place in respondents’ homes and was carried out by members of the NFO System Three fieldforce using Computer Assisted Personal Interviewing (CAPI).

Data collection methods and instruments

The questionnaire for the survey was developed by the NFO System Three team in collaboration with the Scottish Executive. A small pilot exercise (based on 20 interviews) was carried out in early April 2002 and was used to test issues of wording, question order, acceptability and comprehension. The questionnaire used a mix of interviewer-administered sections and Computer Assisted Self Interviewing (CASI), with the latter used primarily for questions about behaviours and experiences. The final script took on average 30 minutes to administer and covered a range of topics, including:

• Perceptions of how common the use of physical chastisement is among friends or family with children of the same age
• Prevalence and perceptions of the use of different forms of chastisement by respondents’ own parents
• Use of different forms of chastisement with own child, ever, in the last year and in the last week
• Details of characteristics and circumstances of the most recent occasion on which they used some form of physical chastisement
• Sources of knowledge, information and advice about parenting and issue of discipline
• General orientations towards smacking and perceptions of its appropriateness in different situations and for children of different ages
• Understanding of the current legal situation regarding physical chastisement and awareness and understanding of the proposed legislation
• Attitudes towards the various elements of the proposed legislation

The questions on the prevalence of different types of behaviours drew on the Parent-Child Conflict Tactics Scale, developed in the United States by Strauss et al (1998) and on its subsequent modification by the Policy Research Bureau/NSPCC/National Centre for Social Research for use in their ESRC-funded study in Britain (see Finch and Ghate, 2001).

A copy of the full questionnaire is available from the research team upon request from the Scottish Executive.

Because of the relative complexity of the sampling approach and the potential sensitivity of the topic, all interviewers working on the survey attended one of three personal briefings held in late April 2002.

Sample design

The survey was based on a multi-stage probability sampling design, with addresses drawn from the Postcode Address File and ‘screened’ for households containing children/parents. This is a relatively common technique when sampling populations for whom no separate sampling frame exists, but it is dependent on the target population being relatively prevalent within the household population as a whole. Since only around 28% of households in Scotland were expected to contain children, this meant that - across the issued sample as a whole - interviewers would need to visit almost four households for every one at which they could attempt to secure an interview.

Of course, in some areas, the proportion of households containing children was expected to be higher than 28% and, in others, lower. Consequently, different numbers of addresses were issued per sampling point, depending on indications from Census data about the proportions of households within each containing parents/children. The sampling interval was also varied in order to ensure that, in areas of low density of the target population, interviewers were not faced with large distances between eligible addresses. Overall, the objective was to issue interviewers with addresses that would contain broadly equal numbers of eligible households.

Within each eligible household containing more than one parent, one was selected for interview at random using a Kish Grid.

To summarise, the sampling had the following stages:

• Sampling frame definition -For reasons of fieldwork efficiency all postcode sectors with less than 23.33% of households with children were excluded from the sampling frame.
• Selection of postcode sectors - 90 postcode sectors were then stratified by old Scottish Region and the Scottish MOSAIC geo-demographic classification and selected with probability proportionate to the number of households containing children.
• Selection of addresses within selected postcode sectors - Addresses within postcode sectors were sorted by postcode. The first address within each postcode sector was taken as a starting point and the number of addresses then selected and the sampling interval used varied by area according to Census information about likely population.
• Selection of individuals for interview within households -Within any household identified as containing children/parents, a short screening exercise was carried out to establish the number of eligible respondents and to select one, at random, for interview.

Fieldwork and response rates

At the end of fieldwork, 692 interviews had been achieved. Because it was necessary to continue fieldwork until the last possible moment in order to maximise the sample size, many of the field progress forms (known as contact sheets) have not yet been returned by interviewers, making it impossible to calculate final response rates at the time of publication.

Interim calculations, however, suggest a response rate in the range 50-55%. Although this is lower than for many household surveys, it reflects the peculiar difficulties associated with the screening exercise and the problems of persuading fathers to participate. Moreover, the likelihood of significant non-response bias is likely to be less than in surveys of the general population, in which the non-availability of younger people for interview tends to skew final results.

The data have been weighted to correct for unequal probabilities of selection for respondents in households containing different numbers of parents and to correct the gender imbalance in the achieved sample (using information about the composition of Scottish households from the Scottish Household Survey).

Once the relevant fieldwork materials are returned, a more detailed Technical Report on the survey will be prepared and be available upon request from the Scottish Executive.

< Previous | Contents | Next >