False findings show a scientific truth

First came the global headlines, then the doubts - and now the recriminations. Two huge science stories, both hinting at potential revolutions in our understanding of reality, now look fatally flawed.

Sub-atomic particles don't travel faster than light after all, and nor can people foretell the future, as two widely-reported studies suggested last year.

This should hardly come as a surprise: scientific revolutions are rare, and two in one year always seemed a bit of a stretch.

The real shock has been the response of the scientific community to the claims, even before their demise. There has been a chorus of criticism of the researchers for ever taking seriously such "obvious" nonsense, while those responsible for publishing the ESP study have been castigated for sullying their journal's reputation.

Clearly, those levelling such charges believe they possess a paranormal ability - one that allows them to discern truth by thought alone. We should be grateful such individuals were not on hand to dismiss all previous cases of "obvious nonsense", like crystals emitting apparently limitless energy (or as it's now termed, radioactivity), clashing continents (plate tectonics) and rocks that fall from the sky (meteorites).

Science is - or at least, is supposed to be - a process that allows anyone to challenge established ideas, and to get closer to understanding how the universe really works. One begins with a belief, subjects it to experimental test, and then draws appropriate conclusions.

Yet as anyone who ever tried to perform a simple "practical" during school science lessons knows, this is easier said than done. Experiments are fraught with difficulty - as the two claims currently generating such fuss amply demonstrate.

The challenge to Einstein's famous injunction against superluminal travel came from a simple-sounding experiment in which particles called neutrinos were emitted from a source in Switzerland and detected 700km away in Italy.

Data collected over several years pointed to the neutrinos arriving so quickly that they appeared to have been travelling faster than the speed of light. The effect was tiny - just 0.0025 per cent faster - but the team could find no obvious explanation for it.

This led them to announce their findings, hoping others would conduct the sine qua non of any genuine scientific claim: a replication. This has now been performed by another team of neutrino researchers in Italy, and their preliminary results support what most physicists always believed: that Einstein was right, and the particles obey his speed limit.

At the same time, the original team has found two potential sources of error in their own experiment: a faulty fibre-optic connection in a timing mechanism, and a time-stamping error. Whether this is the real explanation of the superluminal effect should become clear over the coming months, as other teams report their own replication efforts.

It is a similar story with the apparent evidence for the paranormal. In a series of experiments involving more than 1,000 people, Dr Daryl Bem, emeritus professor of psychology at Cornell University in New York state, uncovered what appeared to be perhaps the strongest evidence yet for the reality of precognition.

Participants had to predict which of two screens was about to show an image randomly chosen by a computer. Guessing would produce a hit rate of exactly 50 per cent, but those taking part in Prof Bem's study managed 53 per cent.

While not huge, the effect was consistent enough to be hard to ascribe to mere fluke. But Prof Bem went further, looking at whether precognition behaves like conventional psychological phenomena - where for example people respond more strongly to stronger stimuli, and less strongly after repeated exposure. In virtually all the experiments, Prof Bem found results consistent with the reality of precognition.

Even before the results were officially published in the well-respected Journal of Personality and Social Psychology, sceptics were denouncing them as "pure craziness" and an "assault on reason".

Fortunately, others decided to put their faith in the scientific process, and attempted to replicate the claims.

Early efforts produced mixed results, but last week the most comprehensive attempt yet was published in the online journal Public Library of Science ONE by a team led by Dr Stuart Ritchie of the University of Edinburgh.

As with the neutrino claim, the replication focused on matching the original experiment as closely as possible - in this case, right down to using the same software as Prof Bem. And, as with the neutrino experiment, the attempt at replication failed completely.

It's still too early to say that Prof Bem's claim has been entirely disproved - not least because absolute proof is beyond the reach of science. What it can and does do is to build up weight of evidence for or against a claim, and by that standard, the evidence that Prof Bem has demonstrated precognition is now looking pretty weak.

As for what went wrong, various suggestions are being put forward - notably questionable statistical analysis that can make fluke results appear "significant".

Faulty wires, unreliable timers, dodgy statistics - it hardly sounds like scientists have covered themselves in glory over these two claims. Some believe the resulting media hoopla has brought science into disrepute.

Yet rarely does the public get to see the scientific process working so clearly, with a puzzling result being subjected to careful scrutiny rather than brushed aside precisely because it doesn't fit existing paradigms.

Should studies into controversial issues like precognition even be conducted at all? Absolutely: while the probability of its reality is low, if it were found to exist the implications are sufficiently huge to justify research.

While not on a par with the search for the cure for cancer, the study of such bizarre claims does serve a key public service. It provides a showcase for the power of science as a cure for self-deception.


Robert Matthews is visiting reader in science at Aston University, Birmingham England

Leave a Reply