Topic Papers

QLD Election 2012 results show landline telephone is still gold standard

The research industry has been closely monitoring the impacts of mobile only households on survey accuracy in recent years. As there is no currently available method for determining the geographic location of a mobile number, use of mobile numbers in RDD sample has proven prohibitively expensive in many cases. Real world comparisons such as election outcomes allow us to benchmark the SamplePages landline telephone sample currently used by the majority of the industry. The 2012 Queensland election result both vindicates Newspoll methods and supports the ongoing use of probability samples based on landline numbers.

NSW Election 2011 - Judgement as well as accuracy

By now it will be no surprise to know that Newspoll surveys reflect real world events better than any other company. The NSW state election was to all analysis, a ‘no brainer’ from some time out. Nonetheless, a number of other pollsters estimated different results and the fairly wild calls of some about Labor retaining only 14 seats were not proven out on the day.

Federal Election 2010 - Newspoll accuracy again superior

Newspoll uses both online and telephone methods according to the nature of the project at hand. A number of sample suppliers offering only online have suggested that growth in mobile only households and changing response rates might have a negative effect on the accuracy of telephone surveys that use probability sampling. Newspoll’s outstanding result in measuring real world events such as the 2010 federal election supports a growing body of evidence that this suggestion is simply untrue.

How good is survey data? - 2010

It appears to many users of survey research that that there has been a large and continuous increase in total survey error (TSE) over the last 20 years or so. The main evidence used to support this view is the steady decline in response rates together with increasing numbers of our target population having zero chance of being selected in our research. Has survey data eroded in quality to the point where it is no longer a value proposition for informed decision making? Managing this risk (that is, minimising the chances of making a bad decision because of poor quality data) is the focus of this paper.

Queensland election 2009—Newspoll on the money again
As the count for the 2009 Queensland state election progresses, the accuracy of
Newspoll surveys has been once again vindicated.

QLD Election 2009 - Newspoll on the money again

As the count for the 2009 Queensland state election progresses, the accuracy of Newspoll surveys has been once again vindicated.

How to generate high quality regional estimates from phone surveys - 2008

Many research projects require information at a regional or sub-state level. Most require samples to be controlled for uneven population distributions. Designing efficient surveys and providing high quality regional estimates requires considerable expertise as there are many traps for the unwary.

Federal Election 2007 - Newspoll accuracy wins out

In the lead up to the Federal election many commentators were questioning the accuracy and predictive validity of the polls. Here is a comparison between Newspoll’s published results and the official Australian Electoral Commission result.

Matters to consider when choosing Online - 2006

Online research is gaining momentum and Newspoll is expecting its use to again grow rapidly in the year ahead. At Newspoll we encourage and assist our clients to use online when it is appropriate for their needs.