Sections
You are here: Home → Publications → Previous publications → 2001 → Survey on public attitudes towards the physical discipline of children (November 2001) → Appendix C: Research and sampling methodology
Note:

These pages contain material published before October 2003 by the Department of Courts and the previous Ministry of Justice.

 

Appendix C: Research and sampling methodology

Questionnaire Development

  • The questionnaire was developed by the Ministry of Justice and the Ministries of Social Policy and Youth Affairs and the Department of Child Youth and Family were consulted. A market research company, National Research Bureau Ltd (NRB) prepared a version of the questionnaire that was phrased and set out for interviewing. NRB tested the questionnaire, in a cognitive pilot.

Cognitive Pilot

  • This step pre-tested the questionnaire for semantic clarity and comprehension. The procedure for doing this was to administer the questionnaire to a sample of the target people(s), and then to cycle back over the words used, asking what came to mind and what the person thought each question was referring to. This answers the questions:
  • Are they comfortable with the words?
  • Do they correctly interpret them?
  • Do different people interpret the questions the same way?
  • The pilot sub-sample was based on personal interviews with:

6 Pacific Island People

6 Maori People

6 Asian People

6 European People

24 TOTAL PRETEST PARTICIPANTS

  • Twelve were men and twelve were women. A spread of ages was obtained. All have, or have previously had care of children.

Telephone Survey

  • NRB conducted the telephone survey in June 2001. The following outlines the survey methodology and data analysis. NRB supplied the Ministry with the data in table format that was further analysed and written up as the present report.

Sampling

  • The sample frame was all households with a (landline) phone as defined within the most recent white page update of the telephone directory.
  • In order to re-include people who may not have a listing (missed the publication or withheld number) a version of random digit dialling was used. This means that a randomly chosen phone number was drawn on the basis of every xth number of every yth page, and then a "1" is added to that number to randomize it, yet within the range of active numbers assigned to the residential sector. (This is the Waksberg Adaption for RDD).
  • Sampling proceeded by drawing both urban and rural numbers, to the population proportions from Census 1996.
  • The Maori sample of 100 minimum emerged naturally as Maori adults are over 12% of the population. However, any shortfall in participation was made up by continuing to sample numbers and determining ethnicity before interviewing.
  • The Pacific People minimum of 100 is roughly twice their population proportion for adults. In this case sampling and screening proceeds as before, but this activity was focussed in areas where the Census shows Pacific People to be more prevalent.

Respondent Selection

  • The adult aged 18 plus years with the "last birthday" was chosen from all adults normally resident there.

Interviewing and Call Pattern

  • Interviews took place on weekday evenings and on weekends, and were administered over the phone by NRB interviewers.
  • The call pattern consisted of an initial call, then two further calls at different times on different days to attempt contact, and a fourth call where the respondent is willing to give an appointment time when successfully contacted on the third call.

Response Rate Ethnicity Match

  • Interviewing is typically conducted in English. However, New Zealand now contains residents who were either not originally schooled in English as one of their languages or who tend to participate more readily when interacting with an ethnicity matched interviewer.
  • NRB's team was briefed to try to identify the ethnicity/language background of those respondents who are either struggling with the questions and/or who appear to have a different ethnicity/language as a possible source of difficulty. The phone numbers for these were retained with whatever identifying details were obtained, in order that they could be called back by an interviewer with matching ethnicity or language skill.

Analysis and Tabulation

  • The data was entered into a survey analysis package, The Survey System (TSS), being concurrently edited for range and logic. At this point post-stratification or "weighting" is applied to rebalance for any differences between the age/gender/ethnicity mix of the achieved sample, and that of the adult population as determined by the 1996 Census.
  • Cross tabulations were prepared to show the findings in total, and by age, gender, ethnicity, parent responsibility and socio-economic status determined by NZSEI occupational categories.

Quality Control Features

  • NRB observes best practise at each step of the survey. The following quality controls illustrate.

i) The questionnaire was cognitively pretested on the main ethnicity target groups.

ii) Each NRB interviewer was personally briefed and instructed, face to face, by the NRB Supervisor.

iii) Sophisticated RDD sampling was used to provide the best achievable sample frame.

iv) Callbacks were made on 10% of each interviewer's work to ensure quality administration of the questionnaire.

v) Range and logic checks were run on the data as it is entered.

Privacy of Information and Ethics

  • This was achieved by capturing data into electronic form, where each data record contains coded answers but does not contain the name of the respondent, nor the phone number. Thus, the survey data is "anonymised".
  • The original questionnaire forms are held securely for a brief period to allow audit, and are then security shredded.
  • Informed consent is not practised in relation to sample phone polls. This is because people are giving opinion data about a public issue, rather than personal data about themselves. The equivalent process of "informed consent" in phoned opinion surveys however is the "refused option", ie. each respondent may refuse any question they are not comfortable with, or terminate the interview.

Identification of ethnicity

  • People were asked to identify which ethnic groups they belonged too using the Statistics New Zealand criteria. The selection of a respondent's primary ethnicity, where they specified more than one ethnic group, for the purposes of coding followed the following criteria:
  • was coded as Maori, where Maori was one of the ethnic groups given
  • was coded as Pacific person, where Pacific ethnicity was given and another which was not Maori
  • was coded as NZ European/Other where one or more ethnic groups were given that were neither Maori or Pacific peoples.

Identification of socio-economic status

  • To obtain information on respondents' socio-economic status the New Zealand Socio-economic Index of Occupational Status (NZSEI) (Davis et al. 1997) developed by Statistics New Zealand was used. The use of occupation as an indicator of socio-economic status is:

premised on a model of the stratification system in which occupation functions as a latent, intermediate variable converting education into income, measuring "the potential of an occupation to convert a person's main resource (education) into a person's main reward (income) (Kunst and Mackenback, 1994, p.38 cited in Davis, McLeod, Ransom and Ongley 1997, p.39).

  • The NZSEI assumes that a person's occupation provides a reasonable basis on which to assign them a position on the socio-economic hierarchy.
  • Different occupations are allocated scores on a socio-economic index by computing a weighted sum of the average education and income levels observed for their current practice (Davis, et. al. 1997, p.39). The range of scores run from 10-90, for consistency with the International Socio-economic Index. The higher scores represent higher education and income levels and therefore higher socio-economic status. For a full discussion regarding the development of the NZSEI see The New Zealand Socio-economic Index of Occupational Status Research Report #2 (Davis et al. 1997).
  • Respondents were asked to state the occupation of the main income earner in the household. The occupation was classified at the three digit level (minor group) on the NZSEI, which is based on occupational data from the 1991 Census and the NZSCO90. These groupings were further classified to a two digit code (major group), grouping the occupations into ten major groups.
  • Occupations at the major group level are presented in the survey findings, with the exception of the Armed Forces, where the sample size was considered too small.
Document Actions