Dialogue: It's the science, not the professor

By JOHN ROUGHAN

When you see the whites of their eyes it usually means you have hit something tender and vital to their intellectual architecture.

There are, I agree, more obvious explanations for the way Helen Clark, Trevor Mallard and Michael Cullen quite lost it this week when Wyatt Creech questioned a Health Research Council grant to the Prime Minister's husband, Peter Davis.

There is the injustice of it. Professor Davis is such a conspicuously decent and dedicated social scientist that it takes an effort to imagine him deliberately directing research to Labour's advantage. But that is
not quite what Mr Creech was saying, when he could get a word in.

Then there is the emotional explanation. Who would not be as livid as the woman we saw on Holmes defending her husband? Last week we might have said Helen Clark. She had publicly admonished the poor fellow for e-mailing her office about the appointment of a colleague.

And he was, after all, a member of the Auckland Hospital Board when she, as Minister of Health, sacked the board and put in a commissioner. That speaks for her probity as much as anything.

As for Dr Cullen and the others, loyalty might explain the fact that they could not contain their outrage long enough for interviewers to finish a sentence. But loyalty is not a high value in this Government; it has already banished several ministers on accusations alone, insisting they could no longer credibly do their jobs.

That is getting closer to Mr Creech's charge against Professor Davis, and it is very unfair. But Mr Creech tried to say something more valid about social science generally. It barely came through the cacophony on television.

You might have caught the phrase "experimenter effect." He did not get a chance to explain it. It refers to the phenomenon that people who commission social or economic research nearly always get results that verify their point of view.

Notice that? They say it has to do with the selection of data and the way it is used. It is not dishonest; the same data can give different results depending on the questions asked of them. As anybody who runs an opinion poll knows, it is quite difficult to devise a question which will not influence the answer you get.

Real sciences deal with that problem with a culture of vigorous internal criticism. When you read the results of social science you do not sense the same rigour.

The difference between social science and the real thing is that social science is constantly reinforcing its theories. Real science spends most of its time tearing them down.

When a researcher in the physical sciences believes he has explained something, he publishes a paper setting out exactly what he did and how he came to a tentative conclusion. Thereupon, others in his field set about trying to disprove it.

They repeat his experiments and subject the hypothesis to experiments of their own to check every other conceivable explanation for the phenomenon.

It is not until it has survived years of sceptical examination from different viewpoints that the science may accept the proposition as a reasonable theory.

Social science is not at all like that. When you read its findings you get the impression that nobody has subjected hem to critical tests or considered alternative explanations.

Social science endlessly assembles data to confirm exactly what you would expect: the unemployed are poor, criminals tend to be uneducated, employers discriminate against people with full facial tattoos. Well, not the last; social science would not ask that question. It would be content to call it ethnic discrimination. There are many questions about human behaviour and its consequences that the pseudo science does not want to know.

Medical science is certainly the real thing. It is hard to be as confident of health research when it turns its attention from individuals to the community's health.

Something changes in the tenor of any science when it finds a public mission. Environmentalism has had that effect on the physical sciences. In the name of public health we are beginning to hear extra-ordinary findings on the effects of passive smoking.

It is less a reflection on Professor Davis than on the field in which he works that Mr Creech was right to raise the question about a grant of $750,000 (which seems an awful lot of money) for a survey of the
effects of health changes on patients.

But the answers he received by letter last October from the chief executive of the Health Research Council, Bruce Scroggins, might have put his mind to rest.

The "restructuring" under study seemed to be those of surgical and nursing techniques, which were never politically contentious, and Professor Davis' proposal ran a fairly impressive gauntlet of professional checks before it was approved.

It was one of 303 research applications to the council in November 1999. They were all, says Dr Scroggins, sent to local and international referees to be assessed for health significance, scientific merit, design and methods and the proven ability of the research team.

If they clear that hurdle they are examined again against those criteria by an assessing committee of New Zealand and Australian leaders in the discipline. Less than a third of the applications were finally accepted. Professor Davis' was one.

That sort of collegial approval might not match the rigour of independent science but it probably immunises the system against partisan political considerations. If only all social science faced those tests.

Former practitioners now in power were making some high claims for the objectivity of social research this week. They were a little too shrill.