Stats NZ has a new website.

For new releases go to

www.stats.govt.nz

As we transition to our new site, you'll still find some Stats NZ information here on this archive site.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Comparing methodologies

This chapter highlights methodological differences that may have had an impact on the seemingly contradictory findings between the census and post-censal surveys. However, this chapter is not a comprehensive record of the methodology of each survey.

How is census data different from survey data?

Census data is different from survey data in a number of ways. The most important difference is that a census sets out to include information from every person in the country. Therefore, it is not subject to sampling errors that occur in a sample survey like Te Kupenga. However, all data sources are subject to non-sampling errors resulting from respondent errors, collection or processing errors, and undercounts. We strive to reduce each of these error types and provide data that is fit for use.

The census includes a broad range of topics providing good contextual information for individuals and families. Te Kupenga narrowed the focus to Māori well-being, and the 2001 HMLS narrowed the focus again to just te reo Māori. As a result, Te Kupenga and the 2001 HMLS collected information in much more detail and depth on te reo Māori than the census.

The population coverage of the census means information is available for much smaller geographic areas – down to the meshblock and area unit levels – and for small population groups; for example, small iwi groups. Surveys often struggle to provide robust statistics at this level.

Respondents fill in the census forms themselves. Like other self-administered questionnaires, self-assessment can lead to more-truthful responses, because the interviewer cannot influence the respondent. However, respondents might not fill in the form correctly, which may lead to issues or errors with the data (Statistics NZ, 2013).

Different questions in the census and the post-censal surveys

The same question on te reo Māori speaking ability was used in the 2001 HMLS and Te Kupenga (2013) in order that the two data sources would be comparable. In the 2001 HMLS, we decided to use self-assessment to determine speaking proficiency. This decision followed a literature review by Te Puni Kōkiri and a series of field tests that indicated we could attach reasonable confidence to self-assessment scores (Statistics NZ, 2002).

In the 2001 HMLS, we asked respondents to place themselves into one of five categories for the following question:

How well are you able to speak Māori in day-to-day conversation?

  1. very well (I can talk about almost anything in Māori)
  2. well (I can talk about many things in Māori)
  3. fairly well (I can talk about some things in Māori)
  4. not very well (I can only talk about simple/basic things in Māori)
  5. no more than a few words or phrases.

While the census similarly uses self-assessment to determine te reo Māori speaking ability, the question is very different. In fact, the census question is not specifically about te reo Māori, but about languages spoken in general. It asks respondents which languages they can have a conversation in about a lot of everyday things, of which Māori is one response option. The question gives respondents no guidance to help them determine their proficiency level.

Methodology used in the post-censal surveys

Te Kupenga

The target population for Te Kupenga (2013) was the usually resident Māori population of New Zealand living in occupied dwellings on 2013 Census night and aged 15+. The Māori population included all individuals who identified with Māori ethnicity or Māori descent in the 2013 Census. We collected the data for Te Kupenga over 4 June to 25 August 2013. The questionnaire was answered by 5,549 individuals, achieving a response rate of 74 percent.

We conducted computer-assisted personal interviews, with the interview lasting an average of 40 minutes. We introduced Te Kupenga to respondents as a general well-being survey. It collected information across 13 modules. The module about te reo Māori was the 12th module, following modules on whānau and tikanga (customs and practices). The speaking proficiency question was the first question in the te reo Māori module.

Respondents could choose whether to complete the survey in te reo Māori or English – the layout of the questionnaire allowed respondents to switch between them. Just 27 out of 5,549 respondents completed their interviews in either te reo Māori or a combination of te reo Māori and English.

We recruited 11 interviewers with Māori language skills (but who were not necessarily fluent) for the collection. These 11 bilingual interviewers were located in geographic areas with a high proportion of Māori-language speakers. If a respondent asked to complete the survey in te reo Māori, a bilingual interviewer in the area closest to the respondent’s location conducted the interview. At times, this required some bilingual interviewers to work outside their designated area.

See Te Kupenga 2013 for more information on data quality.

2001 Survey on the Health of the Māori Language

The target population for the 2001 HMLS was individuals belonging to the New Zealand Māori ethnic group who lived in private households and were aged 15+. We selected 6,072 eligible people for the survey. We received 4,737 full responses – a response rate of 78 percent.

The survey was completed by personal interview over May and June 2001. Respondents could choose whether to complete the survey in te reo Māori or English – the layout of the questionnaire allowed respondents to switch between them. Some 550 respondents completed their interviews in either te reo Māori or a combination of te reo Māori and English, representing 12 percent of the total number of interviews.

Interviews in te reo Māori took an average of two hours to complete, while those in English were closer to 40 minutes. We introduced the 2001 HMLS to respondents as a ‘nationwide survey to find out the number of people who speak and understand Māori’. The question about speaking te reo Māori was near the beginning of the questionnaire, straight after questions on language acquisition.

Sixty-three interviewers worked on the survey, 50 of whom were Māori. Forty-six interviewers were fluent speakers of te reo Māori, and 10 of these had recently worked as Kaitakawaenga (liaison officer) for the 2001 Census. Twelve interviewers were existing Statistics NZ interviewers. Respondents who had identified on their 2001 Census form that they could hold a conversation about everyday things in Māori were assigned a fluent interviewer. In other cases, if a respondent chose to complete the interview in te reo Māori, we referred them to a fluent speaker.

See Final report on the 2001 survey on the health of the Māori language for more information on data quality.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Top
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+