At KORE CSR we are committed to improving the way we do research. In particular, we are committed to improving political polling.
There’s plenty of laments across the internet about how the ‘pollsters got it wrong’ relating to most recent elections. In most cases, they didn’t really, Australian elections are often a dead heat statistically (within 2-3 points of 50/50) which makes polling predictions difficult.
The Voter Choice Project was a very large study during the 2019 election to understand a whole range of behaviours and factors that go into voter’s decision making. The KORE Poll continues that work, focusing on trying to detect the level of volatility in the electorate to more accurately understand why people are voting the way they are.
We also use our unique model to provide a different research option for organisations and businesses that are more interested in finding out the why, rather than the what or the who.
Can I participate in KORE Poll surveys anonymously?
Yes you can. Samples from the KORE Poll will be drawn from two sources. KORE Panel members will be sent a link unique to them by email. An open survey will also be shared on our website and social media. If you use the link you can participate without ever telling us your name or email address.
But wait, isn’t opt-in surveying bad? How do you get a representative sample if just anyone can do it?
Here’s the myth with representative sampling: all surveys are opt-in. Even with the Census or voting, you actually can’t force someone to give an answer to a question.
We tested sampling a lot in the Voter Choice Project, and found that with quota limiting in and post stratification out, the combination of half panel and half fresh respondents was more reliable than the traditional representative sample where invitations are only sent to a limited number of people fitting demographic criteria. Using standards like people on the electoral roll were both expensive and generated poor quality data, with biased samples.
Social media advertising reached a more ‘average’ voter in the Voter Choice Project than any other sampling method, particularly in terms of gender, age and education spread. Like most methods of online polling, it over samples minor party voters (many of whom do not actually vote for their indicated minor party) but we can correct this in post-survey weighting and data cleaning.