Under Obama the Republicans in Congress did everything they could to stop the stimulus package, even threatening default.

Tax cuts for the rich and high military spending are a deadly deficit-creating combination unlike economy-boosting public spending on infrastructure and welfare. And Trump is at it again.

He should have stayed with his earlier view that going in to Iraq and Afghanistan was wrong. But like presidents before him he could not admit that the blood and treasure spent was futile or that western interference in Afghanistan, Iraq and the rest of the Middle East in the cause of the terrorism not the solution.

The folly of his action is illustrated by Pakistan. Parkistan harbours anti-US terrorists, so surely it, too, should be invaded and bombed. But no-one is suggesting that

DUAL CITIZENSHIP

Our seven dual-citizen Members of Parliament may not be aware of another peril they face. The 1975 Common Informers (Parliamentary Disqualifications) Act provides that if found ineligible any citizen can sue them for $200 a day for every day of their ineligibility for the past year. That is $73,000 each.

It is a case of first in, first served. Once one citizen sues, further suits will not be entertained.

There is a strong possibility that at least one will be found ineligible, making the whole exercise quite a good little earner for some bright young do-it-yourself lawyer or law student.

THE ABS

Last week I mistakenly wrote that the Australian Bureau of Statistics will be adjusting the figures it gathers on the marriage survey to correct for sampling bias. This is not correct. The ABS will publish the Yes/No result for each electorate and the age and gender break up of those who vote. It will not be breaking up the Yes and No vote according to age and gender.

My apologies.

The error occurred because I wrongly assumed that the ABS would be following standard rigorous statistical practice of correcting for sampling error wherever it is possible.

However, on this occasion the ABS has been given a ministerial directive as to what statistical information it shall collect. That’s politics.

Usually, standard statistical practice – say on gathering economic data – is to adjust the results if the sample proves not to correlate with the general population as measured by census data.

So, for example, if you are collecting unemployment figures and find that the people who reply “unemployed” have an age profile different from the population at large (say twice as many over 65 than in the general population and half as many under 25 than in the general population) you would adjust the weighting of their replies by halving those of the over-65s and doubling those of the under-25s).

This could have been done in the marriage survey. It would have been a simple matter to mark the outside of the barcoded envelope (which has age, gender and electorate information) with the additional information of whether the vote was Yes or No, before sending the “Yes/No” part of the survey anonymously to one pile and the barcoded envelope to another.

But the ministerial direction said not do it. We can only assume because the Government did not want the result properly adjusted for age, gender and geography as good statistical practice demands.

The ABS’s hands are tied by the ministerial direction, more’s the pity.

However, there would be nothing to stop someone armed with a spreadsheet doing an adjustment based upon the electorate break-up combined with the age break up within electorates and the age break up in opinion polls.

The ABS has statistics based on electorates’ age profile. Some electorates are noticeably “younger” than others. It would be a messier exercise than if the ABS had been allowed to do it while collecting the marriage survey results, but it could be done, and probably will be done if there is a narrow No “win” on the raw count.

Make no mistake, though, this is a deeply flawed political exercise stacked against the Yes vote. Any statistical exercise is suspect where the sample is self-selecting (as this one is) and not selected randomly with equal probability of any member of the population being selected in the sample, unless the result is corrected so that the sample as far as possible matches the population as a whole. That is what the ABS does for economic data. It must be holding its nose while it conducts this political exercise, whatever its official public view.

This is not an exercise to gather accurate statistical information on the public’s opinion on marriage equality. If you wanted to do that you could do a survey of 600 in each electorate and get a much more accurate picture. This is a stalling chance to delay marriage equality. The only way to foil it, is to hold your nose and take part.

www.crispinhull.com.au

Thoughts? Let us know in the comments below.

* Readers are encouraged to use their full details below to ensure comment legitimacy. Comments are the opinions of readers and do not represent the views of Newsport or its staff. Comments containing unlawful, obscene, defamatory or abusive material will not be published.