A canna’ change the laws of physics

Scotty, The Naked Time, stardate 1704.3, Episode 7

Living with uncertainty

Posted by apgaylard on June 11, 2008

Dice in blackI’ve been following a debate between various bloggers and Dr John Briffa about MMR and autism.  The good doctor seems to take the position that we are not in a position to know for sure that the MMR vaccine is safe, i.e. doesn’t cause autism.  This raises interesting questions about uncertainty, proof and evidence.

Of course, in science, there is generally room for doubt.  Particularly when we are dealing with complex interventions in complex systems (like vaccinating people), we are never in a position where we can be absolutely sure.  However, life needs to go on: we have to make the best decisions we can.

This raises important questions: how can we live with this kind of uncertainty?  How can we make the decisions that we need to when there are no guarantees?

Sometimes this sort of debate tends to degenerate, with protagonists taking up unedifying, entrenched positions.  Either it’s someone requiring absolute proof and certainty before they will accept a position, or contending that science has failed because it cannot get to this position.

To a degree, this is understandable: faced with serious decisions about the health of loved ones we would all prefer to have certainty.  So how can we cope when this is not, or is never going to be, forthcoming?

Central to this kind of debate are ideas about ‘proof’ and ‘evidence’.  Proof can mean different things to different people.  Some want absolute proof, something like a formal mathematical or logical proof.  Unfortunately, that’s the domain of logicians and mathematicians, not the experimental sciences. 

Therefore many scientists are, quite understandably, loath to talk about proof.  Others do, but they are talking about something more like a commonly understood sort of proof – legal proof.  This is where something is considered ‘proven’ by the evidence beyond ‘a reasonable doubt’ by a jury of peers.

Casting this debate in terms of the former is pointless.  It’s not what science, or life, is about.  Using the latter standard may get us somewhere.

Next, evidence; debates like this usually focus on anomalies – unexpected, controversial results (Wakefield and his ideas about MMR and autism, for example).  Advocates for a particular position tend to latch on to them – and go no further.  Some even keep hold of them long after they have been discredited.

From a scientific perspective anomalies are the beginning of a process, not the end.  They are rigorously scrutinised, subjected to careful measurement – and mostly evaporate.   A real scientific anomaly will survive and help drive change; it emerges from a background of rigorous science and survives intense scrutiny.

Now, for example, it is clear that the MMR/Autism hypothesis of Wakefield did not emerge from a background of rigorous science.  Neither has it survived scrutiny

Most importantly, rather than focussing in on individual hypotheses and pieces of evidence a more holistic view is required.  We need to bundle together all the hypotheses and evidence associated with each of these and see where they stand relative to each other.

This is, in essence, a plea to consider the philosophy of Imre Lakatos.  This distinguished, but often overlooked, philosopher of science asserted that rather than assessing the merits of individual scientific hypotheses, groups of hypotheses should be considered together.  He styled these research programmes.

A research programme generally consists of a hard core of hypotheses and a protective belt of anciliary hypotheses.  The former is the core of the programme, the latter protects it from premature rejection (such as discarding an idea before it has been sufficiently refined or accepting refutations before they have been properly evaluated) and provides the assumptions and mathematical tools required to develop the programme.

Programmes are then compared on the basis of which ones are making progress and which are not.  When comparing poitions like “MMR is safe” and “MMR causes autism” we can then ask: which is more progressive? That is, which is best at answering perplexing problems; has the most consistency; elegance; simplicity; explanatory power; unifying power; and, dare we say, truth about it? (See here for a contemporary application of these ideas in physics – string theory.)

Taking the MMR and autism debate as an exemplar, I’d say that the answer is pretty clear.  All things considered the ‘MMR is a valuable intervention’ programme is progressive.   What swings it for me is the lack of rigour in the oft-cited anomalies and the direction the best evidence points in.  Science, like life, does not come with absolute guarantees.  But it can point us in the most useful, helpful and productive direction.

The stuff of science is not sterile (and synthetic) debate about proving absolutely that an intervention can never harm; but rather settling reasonable doubt by looking at the balance of carefully tested evidence.

[This is an edited version of a comment I made on jdc325’s Weblog.]

4 Responses to “Living with uncertainty”

  1. dvnutrix said

    Excellent. This makes a fine companion pair with Orac’s The Fallacy of the Perfect Solution.

    It is particularly irritating that some of the Wakefield anomalies have been accounted for by Nick Chadwick’s testimony that the findings of measles in the tissue samples were false positives and that he informed Wakefield that this was the case. Brian Deer gives an account of Chadwick’s objections and his actions and even reproduces the abstract and a relevant table from Chadwick’s PhD thesis. Chadwick’s thesis was published in 1998 and documented results from 22 autistic children – including children reported in the notorious Lancet paper. Chadwick concluded:

    The results do not support previous data implicating persistent measles virus infection with the aetiology of IBD or autistic enteropathy.

  2. apgaylard said

    Thanks. Chadwick’s testimony and thesis really kill the core of Wakefield’s ‘programme’. It’s a scandal that anyone still gives it any credence. I don’t think that there are many better examples of a degenerate research programme. (direct link to Deer’s summary of this now included in the piece)

  3. hustleandfloe said

    Along with Lakatos model, as you refer to, I am reminded broadly of Richard Feynman’s position. He characterized doubt as the fuel for discovery. If I want an absolute truth, I am essentially setting up an artificial roadblock to future discovery. I’d rather just keep faithfully approaching the limit (proof, truth, whatever). I thought out loud on these things myself here:


    In this light we can understand where Feynman notes no dichotomy between science and religion as he would probably agree: “MMR is a valuable intervention’ programme is progressive.”


  4. hustleandfloe said

    …I mean: http://tinyurl.com/usain-bolt-feynman-richard

Sorry, the comment form is closed at this time.

%d bloggers like this: