Just catching up here - the AEC may claim a 97% turnout but I still believe the reality is different. There is absolutely nothing to stop anyone voting on behalf of someone they know is not going to vote or on behalf of someone no longer alive or out of the country. Although cross checking is supposed to take place there is nothing at all to stop someone voting multiple times and then claiming that they know nothing about it etc. (Very hard to prove unless someone boasts about it or is stupid enough to make a habit of it i.e. at the next election.)
Our electoral and voting systems are antiquated and need an overhaul but we also need to be very careful what we replace them with.
Remember the referendum on the republic? That did not succeed but it was not for want of stacking by certain elements in the community who used the system to insist that people fill in the paper 'under supervision'.

Figures posted to the AEC VTR page for Dobell, updated as at 2:52:32 PM Thu 6 Dec 2007, now show a turnout of 95.21%. See: http://vtr.aec.gov.au/HouseDivisionFirstPrefs-13745-115.htm

At most, 1,713 votes remain to be counted in this Division, if the figures posted on the 'Declaration Vote Scrutiny Progress' page bearing the same update timestamp are correct. If all these uncounted, but acknowledged already received, votes are admitted to the count, turnout would rise to 97.11%. Being realistic, it is to be expected that a significant proportion of the 443 uncounted provisional votes will be rejected from the scrutiny.

Against this anticipated reduction from 97.11% is the possibility that as many as 546 postal votes still unreturned as at the time of update may yet get to be counted. So all in all, somewhere around 97% turnout looks likely to be the end result in Dobell. In 2004 turnout was 95.07% in Dobell, according to this AEC web page:
http://results.aec.gov.au/12246/results/HouseTurnoutByDivision-12246-NAT.htm

The 'Turnout by State' page (see: http://vtr.aec.gov.au/HouseTurnoutByState-13745.htm ) as at 8:23:54 AM on Fri 7 Dec 2007 shows an average 'progressive' turnout for NSW of 93.39%, and an average of 92.95% nationally. Clearly, even on average, turnout is looking like finalising right around the long-term experience of 95%, despite the misleading early and progressing figures quite incorrectly derived from the count instead of from an aquittal for ballots issued and returned in no way dependent upon the count conducted for the purpose of determining who is to be elected.

The really significant point is that at a time when such an aquittal could and should have been available and indicating a turnout in Dobell of around 97%, the official count-derived figures were showing 93%, and earlier even less, turnout. Pointless and unsatisfactory at best. At worst a possible means of withholding information from the public that might otherwise alert it to prospects of electoral tampering.

The AEC Virtual Tally Room page 'Turnout by State' linked to above is, as of an update as at 6:52:35 PM on Sat 8 Dec 2007, showing turnout for NSW of 94.72%, and, nationally, of 94.21%.

The 'Home' page link on the left side of the 'Turnout by State' page takes you to a main page entry to the VTR headed 'The Official 2007 Federal Election Results'. The opening summary states that as of 6:52:35 PM on Sat 8 Dec 2007, 94.21% of the primary vote has been counted for Australia at large.

Note the identicality, 94.21%, of the national figures in each of the above paragraphs. Understand that, although the figures are numerically identical, they refer to altogether different things. Turnout expresses votes claimed as a percentage of total enrolments as at roll close. Count progress expresses votes counted as a percentage of total votes cast. Funny how the two different measures just happened to be coincidentally the same percentage.

At the 2004 Federal Elections the final turnout figure (with respect to the House of Representatives elections) for NSW was 94.70%, whilst nationally it was 94.32%. The turnout with respect to the Senate elections was 95.11% and 94.82% respectively. (This difference arises when electors who have claimed an absent vote incorrectly with respect to the Division within which they are in fact enrolled have their House of Representative ballot paper rejected, but their Senate ballot accepted because they are in fact found to be enrolled correctly elsewhere in the same State.) So it is really the higher of the two percentages which is in fact the true measure of turnout, the attempt to lawfully vote.

Given that the 2007 figures are not yet final, it appears that, if anything, turnout may be slightly greater in 2007 than in 2004. So much for any claim that one in five ENROLLED ELECTORS did not vote!

Adding the 5% who fail to vote to the indicated 15% of the eligible failing to enroll shows Communicat could nevertheless be spot on!

dibs' post earlier in this thread (the 17th post) compared Australian population statistics with official enrolment records. (The latter are published monthly in the Government Notices edition of the Commonwealth Government Gazette, as well as on AEC web pages with reference to roll closures for elections past and present.)

The only point upon which I differ from dibs is with respect to the percentage of the total population that is eligible to be electorally enrolled: dibs suggested 75%, whereas I suggested around 65% (based upon an enrolment level study linked to in the 47th post of the thread) as a more likely figure.

When this comparison is made for enrolments as at the roll close for the 2007 Federal elections when population was around 21.1 million, the 13,645,073 shown as being enrolled compares very closely with the 13.72 million estimate of the electorally eligible population. The enrolment level is virtually 100%.

Courtesy of the link provided by dibs to ABS statistics, Australian population at 30 June 2004 was estimated at 20.1 million persons. Enrolment at roll close for the 2004 Federal elections was 13,021,230 (see this link: http://results.aec.gov.au/12246/results/GeneralEnrolmentByState-12246.htm ) . 65% of 20.1 million is 13.07 million, so electoral enrolment was also 100% of the eligible in 2004!

So what was all the hoo ha about people being disfranchised as a consequence of the legislative changes that (supposedly) closed the rolls at 8:00 PM on the day of issue of the writs? The statistics tell us, in aggregate, that everyone was enrolled! The worst that could have occurred to anyone changing address is that their enrolment became 'trapped' in their old Division. Enrolments were at 100% in both 2004 and 2007, so there was no room for any NETT change in the number of enrolments between events. Electors couldn't have been removed from the roll since leaving, could they?

Not unless other different names had replaced the erstwhile genuine ones that had been in the interim removed, surely?

Puzzle.

TBC

(continued)

First a clarification. After the phrase "change in the number of enrolments between events" in the third last paragraph of my preceeding post, insert the following: ", other than that resulting from the excess of newly eligible electors over deaths of formerly enrolled electors." (Such increase is a consequence of a growing eligible population between elections.)

The claim to the existence of 100% enrolment of the eligible at both the 2004 and 2007 Federal elections does, of course, depend upon the applicability of the extrapolated "E%P" factor from the study "Australia - Aggregate Enrolment Levels 1947 - 1987". This does not appear to have been an official study, but was it to have been seriously inaccurate, surely it would have elicited some form of official rebuttal, or at least the promotion of some alternative official view.

Given that the Australian Statistician has been, since the inception of the AEC in 1984, the non-judicial member of that three-person Commission, it is to be expected that the Australian Bureau of Statistics (of which the Australian Statistitian is in charge) would have been regularly publishing official bulletins with respect to electoral eligibility levels in relation to enrolment levels amongst the Australian population. There are many factors to be considered in attempting to assess the proportion of the population eligible for enrolment at any time, and no organisation would seem better fitted to compile such statistics than the ABS. Does anyone know of any such published studies?

The only other foray into this field I am aware of was a seeming passing observation made under the aegis of the demi-official, Australian Research Council funded, ANU hosted Democratic Audit of Australia by Professor Marian Sawer in relation to the enrolment level in 1903. This has been the subject of comment on OLO. See:
http://forum.onlineopinion.com.au/thread.asp?discussion=208#4002

If the claims that many newly eligible are in fact not enrolled, along with others across the age distribution of the population, are correct, then how are emplaced the enrolments that bring the level up to the 100% statistics indicate?

I commented in the 44th post in this thread that, of the seven Divisions I had spot checked, Robertson had appeared somewhat anomalous, apparently due to incomplete posting of declaration vote issues to the VTR up to that point. This deficiency has since seemingly been rectified.

As at 12:31:17 PM on Tue 4 Dec 2007, the VTR page for Robertson was showing a 'turnout' of 94.50%.

The 'Declaration Vote Scrutiny Progress' page of even timestamp showed that there was a total of 1560 declaration votes still outstanding unreturned, consisting of 159 pre-poll votes, and 1,401 postal votes.

2,602 declaration votes already returned remained then as yet uncounted.

The total votes already counted stood at 89,148. When the 1560 declaration votes outstanding, and the 2,602 votes returned but then still uncounted are added to this total, a total number of 93,314 vote claims had been made in this Division. This equated to an apparent turnout of 98.92%!

Up to this point, only 37 declaration votes had been rejected at preliminary scrutiny, consisting of 11 absent, 10 pre-poll, and 16 postal votes. No provisional vote claims had as yet been scrutinised.

By 3:29:33 PM on Mon 10 Dec 2007 Robertson was showing 95.74% 'turnout', with 1,112 postal votes only unreturned. A total of 1,816 declaration votes had by then been rejected at the preliminary scrutiny, with only 1,171 more votes having been admitted and counted since Tue 4 December. The rejections consisted of 534 absent, 696 provisional, 256 pre-poll, and 330 postal vote claims.

The very high turnout of 98.92% assessed upon the basis of vote claims made had been massaged down to 95.74% with the aid of the rejections of declaration votes at preliminary scrutiny. That's the potential difference between assessing what is misnamed 'turnout' progressively upon the basis of votes counted, as opposed to true turnout at finality upon votes claimed, or even votes cast. For the Australian public it could prove to be an extremely important distinction.