Evidence Reduces Trust

A couple of readings and conversations – face-to-face and social-media – recently, that play directly into my agenda of keeping science and humanism honest, and expose where I’m at odds with received wisdom. I’m used to it after 15 years of blogging and, of course, countering with alternatives to received wisdom is the point. I’m not simply being contrary, there are important alternatives being overlooked. Received wisdom is simply a tyranny of the majority.

A number of campaigns I support, many of which fall under Sense About Science, make a lot of sense (obviously) and their intentions are laudable. Laudable enough to actively support as immediate if temporary measures, efforts to get the topics on the public agenda, curb current excesses and abuses of what passes for scientific knowledge. Starting from a ground zero of ignorance and denial, then all progress is positive. But …

But, there is a kind of arrogance that says being right follows from making progress. That’s evolution, innit? And there is a valid line of thinking that says so long as we make progress, who cares about being right. That’s politics, but it’s not science. In politics there will be values, but rarely any concept of ever being fundamentally right. Science on the other hand, whilst knowing it is never right, always contingent, does care about approaching knowledge as truth. If it doesn’t it’s just politics. And this is the Catch-22 again, when you have a political agenda around science we have to be careful to distinguish the politics from the science.

One of SAS campaigns is “show me the evidence” and a corollary of that one is “show me all the evidence” including the null and negative indications, particularly in (say) Ben Goldacre’s #AllTrials demand for publication of all clinical trials, including the failures. Who could argue with that?

Me actually. This is a political extension of the openness and transparency of all considerations and communications. Leaving aside any issues of privacy and security, this may be pragmatically fine from a freedoms and rights perspective, but what is completely impractical is that we all need to consider all available evidence and information. At some point we have to trust the knowledge we’ve got so far and trust the people with & sources of that knowledge. Asking to be shown the evidence is a statement of mistrust or a default to zero trust in the absence of evidence. So it is clearly a judgement where the process stops – when you have enough evidence to trust. That clearly depends on context.

So for the #AllTrials case, where the responsible and expert licensing authorities are in the loop, it will probably be practical to set some rules about disclosure to those bodies (transparently available to anyone, too)(*). The evidence of trust shifts to our relationship with the authority. In the more general “show me the evidence” case, the practical limits will always be a matter of judgement. Evidence that is easily available – and intelligible in all its subtle nuances to whoever is interested – should never be ignored, but we should not expect to see scientifically objective intelligible evidence to support every judgement. (This is Dick Taverne’s argument, and in fact he is a founder of SAS.) The need to trust judgement never goes away, it just gets pushed around. Trust is inevitable and it is where the science and the application of science must part company. Trust, like scientific knowledge, is something we should work to maximise, we cannot entirely replace one with the other.

And there are other competing factors that mean it is counterproductive to pursue the objective evidence line exclusively. One is we will never succeed in achieving watertight definitions of all the objective evidence needed for all situations. And the tighter and more comprehensive such attempted definitions become the more unlikely the nuances will be understand by more people. Simpler communications may give the illusion of wider understanding, but that understanding will be at the expense of actual scientific truth. It may be politically sound to pursue that kind of science communication, be we must be careful not confuse it with the actual scientific knowledge. At some point we always need to trust that the specialist scientists, like the responsible politicians, know better. It’s an illusion to believe we can drive trust out of the system.

Definitions and objective evidence are part of science’s model of the world, and the human world is more than that.

=====

(*)Note: Though even here, where management of the rules and their application has clear authority, it is already possible to predict gaming of the system, whereby potential failures are tested under the radar before bringing into the regulated environment. Rather than selective publication of results we get selective “official” testing. Unintended consequences. ie the devil is in the detail of the execution and management, not in the definitions of the rules and processes. Definitions don’t solve the problem.