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+
Appendix 1: Consistency in answers on te reo Māori between the census and the post-censal surveys

2013 Census and Te Kupenga

There appears to be a good deal of agreement between the 2013 Census figure (23.7 percent) and those in Te Kupenga who said they could speak te reo Māori very well, well, or fairly well (22.6 percent). However, the gap is wider between the 2001 Census and the 2001 HMLS (28.2 percent and 19.8 percent, respectively).

The agreement between the 2013 Census and Te Kupenga appears to continue when we look at the percentage of te reo Māori speakers by age group. Figure 2 shows that compared with the 2013 Census, Te Kupenga appears to have slightly undercounted older speakers (aged 45+), but has more similar figures for those aged 15–44 years.

Figure 2

Graph, Proportion of Māori population who are te reo Māori speakers, by age group, comparison between the 2013 Census and Te Kupenga (2013).   

2001 Census and the 2001 Survey of the Health of the Māori Language

The 2001 HMLS undercounted speakers across all age groups when compared with the 2001 Census (figure 3). This undercount was largest in the 35–44-year age group, where there was a 10.8 percentage-point difference when compared with the 2001 Census. Also, the undercount for older speakers (55+) was larger in the 2001 HMLS than it was in Te Kupenga (7.6 percentage points compared with 5.2 percentage points, respectively).

We can analyse this further by looking at individual respondents’ answers to the 2001 Census language question and the speaking proficiency question in the 2001 HMLS.

Figure 3

Graph, Proportion of Māori population who are te reo Māori speakers, by age group, comparison between the 2001 Census and the 2001 HMLS.  

Respondents report different answers in different surveys

Alignment between responses from the 2013 Census and Te Kupenga, and the 2001 Census and the 2001 HMLS is not perfect. For example, 16.1 percent (4,000) of Māori who reported speaking te reo Māori very well in Te Kupenga had stated in the 2013 Census that they could not hold a conversation about a lot of everyday things in te reo Māori. At the other end of the scale, 4.1 percent (9,500) of Māori who reported speaking no more than a few words or phrases of te reo in Te Kupenga had stated in the 2013 Census that they could hold a conversation in Māori.

We would expect these differences given the discussion in this report about different methodologies. But what we do see are differences in this alignment between 2001 and 2013. In 2001, 47 percent of the 2001 HMLS respondents who stated they could hold a conversation in Māori in the census then reported they spoke no more than a few words or phrases or only basic te reo Māori in the 2001 HMLS. This compared with 35 percent in 2013.

Figure 4

Graph, How well Māori report they can speak in te reo Māori, of those who said in the census they could hold a conversation about a lot of everyday things in te reo Māori, by the 2001 HMLS and Te Kupenga (2013).

False positives and false negatives

If we look at false positives and false negatives from the stance that the census has provided the ‘correct’ response, we see an interesting picture. Table 2 shows the figures shaded in green are false positives – those people who said in the census they could not hold a conversation in te reo Māori, but then reported a good level of proficiency in the post-censal survey. Those figures shaded in grey are false negatives – those people who said in the census they could hold a conversation in te reo Māori, but then reported no or low proficiency in the post-censal survey.

Table 2 shows the level of false positives was higher in 2013 than in 2001, but the level of false negatives was higher in 2001. Both these outcomes cause the rate for 2001 to be pushed down – or the 2013 result to be pushed up.

Table 2
Comparison between respondents’ responses to te reo Māori questions in the 2001 Census and 2001 HMLS, and the 2013 Census and Te Kupenga (2013)

How well Māori report they can speak te reo Māori in the 2001 HMLS and Te Kupenga (2013) 

Whether Māori reported in the 2001 and 2013 Census they could hold a conversation about a lot of everyday things in te reo Māori. 

Yes 

No 

Very well, well, fairly well  2001 – 14.5%
2013 – 13.0%  
2001 – 5.0%
2013 – 9.7%  
Not very well, none  2001 – 12.8%
2013 – 7.0%  
2001 – 67.6%
2013 – 70.3%  
  False positives 
  False negatives 

Note: False positives are those people who said in the census they could not hold a conversation in te reo Māori, but then reported a good level of proficiency in the post-censal survey. False negatives are those people who said in the census they could hold a conversation in te reo Māori, but then reported no or low proficiency in the post-censal survey.

Source: Statistics New Zealand  

Many of the factors outlined in chapter 5 may possibly be driving the higher proportion of false negatives in 2001.

  • 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+