WA FIVE SEAT SUMMARY

The table below compares the final 5 marginal seat data from Patterson Market Research's final survey conducted over October 2nd to the 6th week, across the five marginal seats of Canning, Hasluck, Kalgoorlie, Stirling and Swan. The “actual” data is from the AEC results published on Monday October 11 2004.

The sample error for 1010 interviews is ±3% at the 95% level of confidence, and about ±1.5% at the 70% level of confidence. Our results are uncannily accurate.

• Canning
• Swan
• Hasluck
• Kalgoorlie

CANNING

The results for Canning shown below are based on almost 80% of the vote (79.4% of the vote) that had been counted. Note that the final survey result for Patterson conducted in the period October 2nd to 6th shows a very close correlation with the actual Canning results. Note that the Patterson survey result for the Liberals was within 1% of the “actual” of 53%, as was our assessment of the Green and Democrat vote. It is only in the area of the Christian Democrats and the ALP that we have been marginally off the mark. We correctly predicted a strong swing to the coalition in Canning.

Top

SWAN

The Patterson result in Swan showed a similar level of accuracy to that for Canning. However, we believed that Kim Wilkie would be returned, based on the strong preferences that they could be anticipated to receive from the Greens in particular, but also the Democrats and other minor parties. The overstatement of the ALP vote of 3 points, and the understatement of the Liberal vote by 1% is effectively the difference between our prediction of Swan being retained for the Labor party to the current prospect that it may be converted to a marginally held Liberal seat.

It should be noted that these comments are made after counting has reached 75.5% of the registered vote. Counting continues as we prepare this document, but it may well be that our prediction of an ALP win will prove to be erroneous. We are nonetheless very happy with the survey result in its very close approximation of the final vote. It may also be of some interest that at this stage, about 5.3% of the vote has turned out to be informal, and it is probable that ALP voters would be disproportionately represented amongst that informal category.

Top

HASLUCK

The Patterson result for Hasluck was incredibly accurate. We predicted that Hasluck would be lost by Sharryn Jackson of the ALP, and that Stuart Henry would become the new member for this seat. We were confident of this prediction, and it turned out to be a reliable assessment, as shown below.

Top

KALGOORLIE

The result in Kalgoorlie is a reflection of the difficult nature of this seat as a survey proposition. We were quite happy with our prediction overall that Kalgoorlie would be safely retained by Barry Haase, but it appears that our assessment of Graeme Campbell’s influence on this seat was overstated by roughly the same proportion by which we understated the success of the ALP candidate. We note that Graeme Campbell’s powerbase is likely to be centred around Kalgoorlie and the Goldfields, and that Tom Stephens did particularly well in the far flung regions of the Northwest. These northern areas are more difficult to survey by telephone, as very often individual households are less able to be contacted by phone (particularly the Aboriginal communities) and this area has a history of strongly supporting the ALP.

Patterson had a similar experience in Kalgoorlie in the last federal election, in which our data showed a quite comfortable win for the Liberal (Barry Haase) but in which he only managed to scrape over the line. It may be that the simplest approach for Kalgoorlie is to conduct the telephone survey as normal, but then to factor in about a 3-3.5% “boost” to the ALP based on their traditional superior showing in the outlying areas.

Top