Getting it wrong and learning from it

One of the most important distinctions between behavioural psychology (of great importance to communicators) and some economics (of use to communicators) is that the psychological research often demonstrates a much stronger adherence to scientific principles than the economics research.

The blog has written before about Daniel Kahneman’s warning about a ‘train wreck’ in psychology research due to the over-frequent use of samples drawn from US college students and problems of replicability. In response to that call there have been a number of efforts to overcome the problems and an iterative process that has scrutinised the new techniques to ensure they don’t repeat previous methodological problems or create new ones.

One alternative to the college student samples has been the use of the Amazon Mechanical Turk. The Amazon Mechanical Turk (known as MTurk or Turk) is an online labour marketplace which has been increasingly used by researchers to recruit people to respond to questionnaires and to participate in laboratory experiments. In the Weekend Financial Times (11 October 2015) the economist, Tim Harford, pointed out that the Turk also has problems. He quotes one cognitive psychologist, Dan Goldstein, as saying that “The majority of papers presented at conferences I go to now use Turk.” On the other hand, Yale’s Dan Kahan, “wants academic journals to be more sceptical of about MTurk research.”

Harford points out Turk members are not representative of any particular population although they are more likely to be “young, male, poor but highly educated.” He also reports, however, that some familiar psychology results have been replicated with Turk samples and that it is “more diverse than the traditional study pool.” From Goldstein’s perspectives the problems are manageable if the tool is used with skill. From Kahan’s perspective it all harks back to the original 18th century chess-playing ‘robot’ Mechanical Turk which turned out to be a fraud in which a chess playing human was hidden.

Whatever, as the young say, the work is a bit better than some mainstream economic efforts – particularly those espoused by Australia’s Departments of Treasury and Finance and the suppliers of expert views to business leaders. Fairfax Economics Editor, Ross Gittins, in a 2 November 2015 article advocates that our new PM “should think long and hard before accepting advice from the economists in Treasury and the accountants in the department of Finance.” Highlighting their emphasis on monetary incentives and price signals as the foundation of their advice he lists five policy objectives about productivity, saving, investment, innovation and agility (The PM’s current favourite word) and suggests that every time you will get, from traditional economists, the same advice for each of them – “cut the company tax rate and the top rate on individuals” – while failing to mention how little neo-liberal economics tells us about things like productivity and technological advance.

These views often go unchallenged because they suit the rich and the powerful and, for example, just this week Business Council of Australia Chair, Catherine Livingstone, repeated the BCA mantra while extending it to new found heights by suggesting that the most important way to encourage economic growth was to cut company tax. It is this sort of almost religious belief, which coincidentally  suits the interests of business leaders, which colours much debate about the role of government and various policies in economic growth, innovation and productivity.

Vested interest masquerading as evidence-based policy can often be found in the pages of the Wall Street Journal where government (other than when it bombs other countries around the world) is bad and the private sector is good. Before 2008 the WSJ was full of advice about how the de-regulation of markets and their amazing capacity for self-regulation and adjustment would secure a boom for ever more. That antedated the support for the famous Laffer curve of course. Then came the 2008 view that interference in the markets would hamper recovery. Then came the confident prediction that QE would result in hyperinflation which would make the Weimar Republic and Zimbabwe look deflationary. Then came the confident assertion that if government hadn’t have interfered the economy would be booming by now. And in between all this was the constant refrain about welfare sucking initiative – unless it took the form of incentives or what we generally call subsidies or corporate welfare – for the private sector.

Now the WSJ has focussed on the uselessness of government investment in pure science in promoting innovation in a 24 October article by Matt Ridley. To his credit Ridley actually cites evidence. The consensus from Vannevar Bush onwards to the 1995 National Research Council report Allocating Federal Funds for Science and Technology has been that such research investments are “spectacularly successful.” The NRC, and Ridley, question however whether there is a direct, linear, relationship between the investment and innovation. Ridley said: “Innovation emerges unbidden from the way that human beings freely interact if allowed. Deep scientific insights are the fruits that fall from the tree of technological change.” Indeed, one could cite Galileo (as the blog’s friend John Spitzer reminded him when he drew its attention to the Ridley piece and the Physics Today response on 28 October 2015) and the role of technical advances in lens grinding which made many of Galileo’s observations possible. But what is often forgotten is that when Galileo showed sunspots to convinced believers in the Aristotelian view that the Sun was perfect – they didn’t see them. The similarity with the shills for business beliefs is remarkable.

Of course, while Ridley has some evidence for his conclusion – at least as far as linearity is concerned – it is not that popular among some technologists. Bill Gates, for instance, has suggested (although what would he know about technology and innovation?) that the US needs to triple its investment in basic research. Perhaps, however, Gates’ belief stems from the recognition that all we now take for granted in a connected world and a proliferation of IT tools is based on some fundamental principles of physics.

But the ultimate test of a science – social or physical – is its capacity to predict things. Psychology is sort of getting there. Economic forecasting, as J.K.Galbraith famously observed, simply exists to make astrology look good.