A Herald data blog post by Andrew Chen (@andrewmeows)

Stereotypes are everywhere, and they apply to political parties too.

Everyone has their own idea of what kind of person votes for each political party, but these are usually informed by a few conversations and personal experience rather than any hard data.

This tool was created to help uncover whether those stereotypes stand up to scrutiny, and provides some interesting insights into the characteristics of voters.


Data was sourced from the 2014 election and 2013 census, and collated into a visualisable form to be compared at the electorate level.

You can see whether the relationship is positive or negative quickly, as well as the strength of the relationship between the demographic characteristic and the political party. Adjust the analysis for differences in population between electorates, and filter the General and Maori electorates.

These visualisations should be interpreted with caution - there are plenty of caveats, and plenty of reasons why there might be a relationship between two variables. Importantly, behaviour and characteristics at the population or electorate level should not be used to draw conclusions about individuals.

Use the buttons at the bottom of the graph to change the demographic variable. To see the fullscreen version, click here.

For example, an often held assumption about the political spectrum is that poorer individuals tend to vote left-wing and richer individuals tend to vote right-wing.

The data from this analysis shows that electorates with more rich people (earning more than $70,000 a year) do vote more for National and ACT, but also vote more for United Future and the Green Party. The Green Party relationship is potentially explained by the support in central urban areas, especially in Wellington, that also happen to be areas with higher income individuals. However, electorates with more low income individuals did not vote significantly more for Labour or other left-wing parties, suggesting that the income-political divide is likely more nuanced than a simple left/right poor/rich split.

Explore the other variables for yourself - we hope that this tool helps spark debate about why these voter demographic relationships exist, and what underlying systemic problems they may represent.