You are visiting the website of


Return to Home Page

Return to List of Articles


This article (slightly amended here)first appeared in 'The Skeptical Intelligencer', Vol. 18, Winter 2015, pp1-2.

Recently there was much interest shown by the media in an investigation reported in Nature which concluded that 'intrinsic risk factors contribute only modestly (less than around 10-30% of lifetime risk) to cancer development'. That is, most cases of cancer result from avoidable factors such as toxic chemicals and radiation (note 1).

One reason for the attention given to this research was that earlier this year a cancer study was widely reported that concluded, 'Only a third of the variation in cancer risk among tissues is attributable to environmental factors or inherited predispositions. The majority is due to "bad luck," that is, random mutations arising during DNA replication in normal, noncancerous stem cells' (note 2).

The results of the two studies appear to be contradictory. Perhaps in the fullness of time this contradiction will be resolved but for the moment the lay public will once again be asking, 'What are we supposed to believe?' and bemoaning the fact that hardly a week passes when the media announce that scientists have made a discovery that contradicts previous research findings (which may also have contradicted research findings prior to these). Inevitably people ask, 'Can we really trust science?'

Any research that involves human beings is particularly liable to a confusion of evidence. This is not just the case for the social sciences but also biological sciences (human physiology, biochemistry, genetics, medicine, etc.). Perhaps the main reason is that human beings are so complex. But in fact an attitude of skepticism to the announcement of any new scientific finding or discovery is in itself good scientific practice. This is certainly the case with any observations or conclusions that are not immediately predictable from current scientific knowledge.

The rules of science are simple in theory but deceptively difficult in practice because so much else intrudes into the thought process - notably cognitive biases of various kinds. The major rule is that the preferred explanation for any observation is the one that is most consistent with all available knowledge. So if someone claims that once, during the night, they were visited by strange beings, lifted out of their bed and taken aboard some kind of aircraft, were subjected to a surgical operation, etc. and then returned to their bed, then we stick with explanations that fit best with what we already know about the world. One such explanation in this case is that the person was dreaming or experiencing sleep paralysis rather than being temporarily abducted by extra-terrestrials.

Our adopting the most likely explanation is not the same as saying that the unusual explanation is wrong, and it may be that by applying this rule rigidly scientists may at times be missing something very important - e.g. alien beings are visiting planet Earth. But there is a very good reason for applying this rule. If we adopt the unusual explanation we are left with many more things to explain than if we go for the most likely one. In the present case we have to account for how the extra-terrestrials managed to make a journey of trillions of miles; why their craft was not spotted by radar defences; why the neighbours didn't hear what was happening; how the surgical operation was accomplished with no evidence of scarring; etc., etc.

So the rule is if, based on our existing knowledge of the world, there is at least one explanation more likely than the one being offered, don't accept the latter. So suppose you tell me that you have just done an experiment on remote viewing and obtained highly significant results. I ask you, 'Was there ever a time when you left the participants alone in the room with the test materials?' 'Well, er … yes', you reply, 'but none of them would cheat - they're not like that!'. I then say, 'Sorry. I can't accept that your results demonstrate ESP'.

There is a subtle point to be made here that may clarify what is good scientific thinking. In the above case I do not say, 'The positive results were due to cheating by the participants'. This is a serious allegation and I should provide evidence to support this (other than the positive experimental results). I simply say that so long as there is at least one more likely explanation, one that is more consistent with existing knowledge of the world, I will not accept the paranormal explanation. The onus is then on the experimenter to eliminate this 'more likely' explanation - and others - by further research.

This rule holds not just for the case where a paranormal explanation is being considered; it applies throughout science (and in other scholarly disciplines such as history and indeed in everyday life) when new findings or discoveries require an unusual or far-reaching explanation. For example, in October 2015 I was thrilled to hear that a researcher at the California Institute of Technology, Ranga-Ram Chary, reported that while mapping the cosmic background radiation he detected an unusual glow in one area (note 3). Chary believes that this could be evidence of the existence of other universes (the multiverse hypothesis). He suggests that this radiation could be 'due to the collision of our Universe with an alternate Universe whose baryon-to-photon ratio is a factor of around 65 larger than ours'. However, theoretical astrophysicist David Spergel, from Princeton University, thinks it is worth looking into explanations that do not involve other universes, such as dust. 'The dust properties are more complicated than we have been assuming, and I think that this is a more plausible explanation' he says' (note 4). So for the time being it is good science not to accept the multiverse explanation.

In this way science reflects on itself skeptically. So too in another way: when there are observations or discoveries that, if correct, have significant consequences for our understanding of the world, the reaction should be to wait until the results have been replicated a sufficient number of times, maybe with adjustments to eliminate competing explanations (note 5).

So, although scientists themselves are certainly fallible, the public need not perceive a lengthy timeline of contradictory scientific findings as evidence that there is something at fault with the scientific method. It is the nature of scientific enquiry that this must be so. We do our best to construct the material world accurately through our senses, but to arrive at real, reliable truths about this world is more difficult than we imagine. We see through a glass darkly, as St Paul said. And science, skeptically applied, is the best means we have of penetrating this darkness.


1. Wu S., Powers S., Zhu W. & Hannun Y.A. (2015) 'Substantial contribution of extrinsic risk factors to cancer development', Nature doi:10.1038/nature16166. here

2. Tomasetti C. & Vogelstein, B (2015) Variation in cancer risk among tissues can be explained by the number of stem cell divisions. Science, 347, 78-81. here

3. R. Chary (2015) 'Spectral variations of the sky: Constraints on alternate universes'. here


5. For this reason, amongst others, accounts of one-off observations such as UFO sightings, ghosts and strange coincidences are not usually good material for scientific investigation.