Colmar Brunton is delighted to announce a major change to the 1 NEWS Colmar Brunton Poll. To ensure we continue to provide New Zealanders with reliable polling information we will use a mix of mobile and landline telephone numbers.
For each Poll, we interview a sample of 1,000 New Zealanders who are eligible to vote. Previously we have randomly generated landline numbers to collect these samples. In households with two or more adults we ask to speak to the person with the next birthday. Historically this approach has delivered representative samples, and the Poll has proved a reliable guide for the election result – including 2017.
New Zealanders’ relationships with their phones have changed, and we need to respond to that shift. It is becoming increasingly challenging to use this approach to build representative samples. The 2013 census found 20% of those aged 18-34 had access to a mobile phone but did not have access to a landline. [Source: Statistics NZ, customised report and licenced by Stats NZ for re-use under the Creative Commons Attribution 4.0 International licence.] This proportion is likely to have increased in the past four years.
The recent 2017 election demonstrated the importance of having a representative sample of younger New Zealanders. The youth vote was key to explaining Labour’s surge in support in the polls in the lead up to the election.
The Jacinda-effect appeared to have an influence, propelling Labour ahead of National in the 1 NEWS Colmar Brunton Poll published on 31 August. Support for Labour had increased across the generations, but noticeably with those aged 18-34.
In the subsequent two Polls support for Labour softened amongst the older generations (35-54 and 55+) after Labour’s tax policies came under scrutiny. However, Labour retained its lead in the 1 NEWS Colmar Brunton Poll thanks to increasing support from younger voters.
However, on 20 September in the final poll before the election, support for Labour amongst 18-34s fell dramatically.
Our analysis shows Labour support amongst those aged 35+ remained steady, but at the same time it fell noticeably amongst 18-34s year olds. This enabled National to regain the lead in our final Poll -46% National vs 37% Labour; in line with the final election result.
However, getting the right representation of younger people has become more challenging and we are committed to ensuring the 1 NEWS Colmar Brunton Poll continues to deliver robust and representative samples of young New Zealanders. Therefore we have decided to shift to a mix of randomly calling mobile and landline phones.
We can be confident of the accuracy of this approach having piloted this in the final poll ahead of the Election. The 1 NEWS Colmar Brunton poll that was published on 20 September used the same methodology as all our polls leading up to the election – based only on calling landline numbers, but at the same time we conducted a trial that included the results of both randomly calling mobile numbers and landline numbers. By blending the landline and mobile samples we achieved a more representative sample by key demographics such as age and ethnicity including Pacific Peoples and Asian New Zealanders. This further improved the accuracy of the survey findings relative to the Election result.
We are now conducting 50% of the Poll using randomly generated landline numbers and 50% using randomly generated mobile numbers. The new approach is currently being used for the end of year poll for Q+A announced on 10 December, and will used for the 1 NEWS Colmar Brunton Poll from 2018. We look forward to continuing to provide reliable and pertinent data that informs political discourse, in a rapidly changing world.
For more information please contact:
Alida Shanks
Marketing & Communications Director
E: alida.shanks@kantar.com
M: +64 21 782 502