The market research agency Ipsos interviews at least 1,000 New Zealand residents who are eligible to vote.
Interviews are conducted over the telephone using CATI (Computer Assisted Telephone Interviewing).
Ipsos conduct the poll over several days including a weekend to allow as many people equal chance of being interviewed.
Ipsos select households randomly and call into urban and rural regions using quotas determined by age, gender and ethnicity. Interviewers ask to speak to the person in the household who is aged 18 years or over and is the next to have their birthday. If the person is not immediately available we call back repeatedly giving them every chance of participating.
At the end of the survey Ipsos examine the data and apply weights to remove any biases that might occur. Weighting the data ensures the final sample looks as close as possible to the actual population.
Ipsos polls provide a maximum sampling error of +/-3.1%-point, at the 95% confidence level. This means Ipsos can be 95% confident that their poll result is within 3.1% of the result had they surveyed the entire population of eligible voters.
Assuming ACT, United Future, Mana-Internet and the Maori Party all retain the seats they currently have, the House would have 124 seats and would look like the graphic above. National would hold 69 seats outright, enough to govern alone but increasing to 71 with the inclusion of United Future and ACT. A Labour-Green Coalition would hold 48 seats, with Mana-Internet and Maori delivering an Opposition total of 53.