New Zealand Election Study Datasets
Please follow the links below to open the relevant Dataverse datasets in a new tab.
The 2020 NZES attracted 3730 valid responses. Participants were sampled from the electoral rolls, containing 94.1 percent of those qualified to vote. 1279 also participated in the 2017 NZES. Persons of Māori descent were oversampled, providing 1246 valid responses. Data in the frequency tables has been weighted by Maori/General electorates, age, gender (from the rolls), highest educational qualification (from the 2018 census), and party vote and turnout (corrected from the marked rolls), in order to be as representative as possible of those on the rolls. The weight variable is lfinwt.
Across those freshly sampled for 2020, on a conservative basis, not removing any of the original sample for non-availability, the response rate (weighted to take account of oversampling) was 32.3 per cent. The response rate of those who could be recontacted from the 2017 sample was 61.6 per cent. All participants received a $20 voucher as compensation for their time.
The 2020 NZES was funded by Te Herenga Waka-Victoria University of Wellington, the University of Auckland, the New Zealand Electoral Commission, and the University of Otago. It was administered by the Public Policy Institute at the University of Auckland.
The questionnaire contains Module 5 of the Comparative Study of Electoral Systems (repeated from 2017). It was sent out by post, and participants could return a hard copy by post or fill out an online version that was also available in Māori and Chinese. 33 per cent of responses were online, 67 per cent by post.
Across those freshly sampled for 2017, on a conservative basis, not removing any of the original sample for non-availability, the response rate (weighted to take account of the oversampling) was 30.6 per cent. The response rate of those who could be recontacted from the 2014 sample was 61.6 per cent.
Questionnaires were in the field 2-3 days after the election. The questionnaire contains instruments to the extent to which individuals’ aspirations for economic advancement and their perceptions of job security or insecurity affect voting choices and turnout. Those who identify a ‘politics of aspiration’ suggest that anticipations of economic advancement by individuals’ own efforts could shape their political behaviour. This may be one reason why the association of income with political choice is often weak: people relate to their anticipations of future rather than present income. Yet aspirational effects may be offset by factors such as low job security. Drawing on a new measure of wealth and assets we will test these conjectures.
Historical Data (1905-1993)
These excel files contain electorate by electorate data for all New Zealand elections between 1905 and 1993. They have been compiled over some years, and should now be free of errors. They include the totals, and both rounds, of the two Second Ballot elections in 1908 and 1911.
Users should note, however, that party labels for early elections were often fluid, and it has proven difficult to make party totals completely consistent with the official data reported in the successive E9 and earlier official results, and with the figures reported in Mackie and Rose’s International Almanac of Electoral History. J. O. Wilson’s New Zealand Parliamentary Record was used as a further source of data.
Thanks to our research assistants Philippa Miskelly, Jean Kite, Sam Martin, and Jason Byrnes for this data.