Contents

Home

Main pages

Appendices

Downloadable .pdf

About the author
How reliable is science?
If science is about trying to obtain correct answers, an important question is how successful is the process.

In summary, we should say that it has been very successful on the basis that
  • Scientific knowledge has been used as in many human enterprises, for example (a) to predict eclipses, (b) to send rockets into space, (c) to eradicate smallpox, and (d) as a basis for the technology of mobile communications. The success of these enterprises gives us confidence that scientific methods will be successful in these and other spheres in future studies.
  • Replication of new results published in a respected journal generally gives similar results.
  • Replication of systematic reviews in medicine generally gives similar or the same results [1].
  • It is rare for an established body of "knowledge" to be found to be false.
  • The answers in science form an interlinked whole.

Possible sources of unreliability

These include
  • errors and fraud i.e. "research misconduct"
  • concerns about the basis of science, e.g. the "problem of induction".

Research misconduct

Research misconduct has been defined in biomedical sciences as "behaviour by a researcher, intentional or not, that falls short of good ethical and scientific standards" [2].

The "problem of induction"

Some philosophers have raised concerns about the reliability of science. One concern is over whether the world will behave in the future in the same way as it has done in the past, which has been termed "the problem of induction". The argument is that the finding that all observations have conformed to a particular law in the past does not prove that they all will in the future. But this is not a problem for science in particular but for all decisions that we might make about the future. We cannot prove that tomorrow will be more or less like today, but it seems to be a reasonable assumption, and one that most people make without thinking about it (and perhaps without ever thinking about it). People demanding absolute proof that the future will be like the past will not get it. If they say that this means that science cannot be trusted, then they should never step outside of a building in case gravity suddenly fails and they float off into space, and they should never start to write anything about science because there is no certainty that the world will last long enough in its present form for anyone to read what they write.

So we should regard "the problem of induction" as a trivial problem that should not distract people who want to tackle important problems in a scientific way. We should assume in general that what we are studying scientifically will continue in the future according to the same laws of nature as it has in the past because we make the same assumption in every other aspect of our lives.

How ethical it is for people interested in the philosophy of science to propagate concerns about "the problem of induction"? Suppose for example that someone wrote a book (merely for profit and personal gain) with the result of undermining confidence in scientific methods, and this book contributed to a reduction in the uptake of an immunisation programme or life-saving treatment for cancer; would this be an ethical course of action?

Expressions of concern about confidence in scientific statements

In any situation, application of the best scientific methods should give us the most accurate answer to a question together with an estimate of how much confidence we should put in that answer and/or its accuracy. Anyone with concerns that the answer or the associated degree of confidence in it could be improved should participate in the process of answering questions as well as they can be answered; those that only heckle from the sidelines should be suspected of doing so only because they have no good basis for their concerns.

References

[1] Konstantinos Siontis et al (2013) Overlapping meta-analyses on the same topic: survey of published studies BMJ 2013;347:f4501 http://www.bmj.com/content/347/bmj.f4501
[2] Aniket Tavare & Fiona Godlee (2012) Helping institutions tackle research misconduct BMJ 2012;345:e5402 http://www.bmj.com/content/345/bmj.e5402





Last updated: 16 Apr 2018