How to wrongfoot science in three easy lessons

Bill English has been the cabinet’s king advocate of data-mining. Last week he turned to anecdotes. That echoes his squeeze on science funding.

As Finance Minister and chief public service and social services reformer up to December, English gave many speeches on the value to be extracted from data on people, especially children, to find where to get the highest return for taxpayers’ money.

This was a major driver of his development of “social investment”, the setting up of the cross-agency social investment unit and the Ministry for Vulnerable Children and the introduction of the Treasury’s daunting CBAx spreadsheet test of new funding requests.

But at his post-cabinet press conference last week English chose anecdotes, not data, to justify his programme of torrential immigration. One reason he gave — to be fair, he did say it was one, not the, reason — was that employers in some sectors repeatedly tell him they cannot hire enough locals because locals keep failing drug tests.

The actual fail rate on the job, official data say, is around 5%. For pre-work-tested beneficiaries, it is a fraction of that.

English stuck to his anecdotes. And, again to be fair to him, one partial reason for the discrepancy between his anecdotes and the data may be that many don’t take the tests because they know they will fail.

But English can hardly be data-king one day and Winston Peters-style anecdote-baron the next.

So to big-brain Nick Smith’s data on “swimmable” waterways. Smith’s changes to how to measure that risk gave the government a bout of political campylobacter. (Is Smith himself a risk factor?)

It was an standout example of how not to do science communication. A group of disinterested scientific experts spent four days getting a complete understanding of how Smith’s recalibration could work well.

That will be included in a long-gestating, extensive, learned report on freshwater Chief Science Adviser Sir Peter Gluckman will issue in a month or two. Had Smith waited a bit — but he too often doesn’t wait a bit — he could have leaned on this heavyweight work.

Well-grounded evidence is the point of Sir Peter’s appointment of science advisers to major departments. They assess evidence more rigorously than arts-educated policy analysts and managers.

The advisers are now integral to budget bidding, which requires agencies to navigate the CBAx and point towards evidence-based eventual “outcomes”. The advisers had another round of evidence testing last week, with very mixed results from agency to agency.

So science is helping form the budget. But will the 2017 budget return the compliment?

Successive governments’ track records have been suboptimal, judging by the first official annual report of science performance in December.

The 2016 budget trumpeted a $410.5 million rise in government funding of science and innovation from fiscal 2015-16 to $1603.3 million in 2019-20.

Sounds a lot. But as a percentage of GDP the rise is to 0.52% from 2015-16’s 0.51%, which has been the average from 2009-10.

And 0.52% is far below comparable figures for all but cash-strapped Ireland among the other six small advanced economies Sir Peter has gathered into a group that meets on science and now on other matters.

The performance report puts Israel at 0.63%, Switzerland 0.89%, Finland 0.98% and Denmark 1.00% and the OECD average 0.70%.

Also, private sector investment, while growing in recent years to 0.60% of GDP in 2015, is last among small advanced economies and far below the OECD average.

The good news: scientists here seem to do “more with less”, which English pushed while Finance Minister.

While second bottom among small advanced countries (again, to Ireland) in the number of research papers published, New Zealand is top in papers per million dollars and gets a better-than-world-average share into top journals. The papers are also disproportionately cited (though still behind most small advanced countries).

The strongest citation performance per paper is in the physics and astronomy category, then dentistry, then engineering — not cows.

One plus: the funding system is less fragmented than it used to be.

But fewer than 10% of applications are funded. One authoritative estimate is that 20%-25% are of high quality. If so, half the best ideas miss out.

For example, the Endeavour Fund which invests in “smart ideas” and “promising, innovative, research” this fiscal year had only $37 million available for $200 million of applications for the second category. Winners eke out funds over two to three years.

Unsurprisingly, despite a net inflow from offshore in recent years, New Zealand comes near or at the bottom of the small advanced countries in numbers of researchers per 1000 people in employment.

That is not charting a path to a “high-wage” economy.

Well, Steven Joyce was Science and Innovation Minister when the report was issued. As Finance Minister he has the option to reset the funding data. Anecdotes don’t cut it.