Contents

Home

Main pages

Appendices

Downloadable .pdf

About the author
Methods used in science
One implication of the answer that science is "Trying to give correct answers to questions that we feel have correct answers" is that science is a human activity i.e. the process of trying to find correct answers. This process involves a variety of methods which are used in different combinations to try to find correct answers in different situations. So, there are scientific methods (pleural) rather than a scientific method (singular), which some authors have written about.

A key part of the process is making observations and conducting experiments, and preferring conclusions based on these to mere opinions. The earliest recorded example of an experiment is in the Bible (The Book of Daniel, Chapter 1:1-16) where Daniel preferred not to eat the food and wine assigned to him, persuaded his guard to allow a ten-day trial of vegetables and water instead, and (it was said) looked healthier for it [1].

Scientific methods can be divided into
  • methods used for clarity and precision in discussing topics in science
  • methods used in planning investigations
  • methods used in carrying out investigations
  • methods used in analysing and interpreting results
  • methods used in publishing results

There is a wide range of methods, depending on the field of science. The following list gives some of the main examples.

For clarity and precision in discussing science

Use of an agreed terminology
  • Agreed standards such as using metre for length or kilogram for weight
  • Agreed definitions e.g. using Body Mass Index (BMI) [2] when discussing normal or abnormal body weight
  • Agreed rating scales such as in the definition of obesity [2]

This agreed language means that researchers can compare findings directly, without having to assess (or guess!) whether they are talking about the same things. Sometimes this agreed language comes from someone making a sensible suggestion and then everyone else taking it up, and sometimes it is from an international conference where a consensus is reached after discussing and debating alternatives, such as the internationally agreed staging for cancer [3].

Definition of any new or uncommon terminology or abbreviation
Good practice is for terms to be defined if there is any doubt about their meaning.

Distinguishing between facts, uncertainties and opinion
Writen and spoken text distinguishes between
  • facts that do not need justification (such as that the earth is round rather than flat) - evidence is readily available elsewhere
  • facts where justification might be helpful - the statement is followed by reference to the research or summary text that supports it, either in brackets or referenced via a number (generally bracketed or as a superscript) that refers to a reference list at the foot of the page or at the end of the document
  • uncertainties that are denoted by appropriate text such as "this suggests that", or "a possible explanation is"
  • opinions

In planning investigations

  • Replication of previous experiments: in many fields, research findings are not generally accepted until successfully replicated by an independent group of researchers.
  • Sufficient repetitions of an experiment to ensure that results are representative of what is being studied
  • If there is a large variation in results, then subjects should be selected to be representative of the group to whom the results of the study will be applied.
  • Specification of all procedures in advance as a protocol.

In carrying out investigations

  • Adhering methodically to the planned protocol wherever possible
  • Recording any departures from the planned protocol
  • Dealing with departures from the planned procedure in a fair way
  • Making use of chance findings e.g. as happened in the discovery of penicillin

In analysing and interpreting results

  • Not jumping to conclusions on the basis of a small number of observations.
  • Not treating an association as due to cause and effect unless there is other evidence for this.
  • Construction of laws as summaries of results that seem to follow a pattern.
  • Construction of models and theories i.e. provisional explanations for results, which might lead to futher studies that either support or invalidate those models and theories. There is not a clear distinction between models and theories, but generally models are taken to be smaller in scope and perhaps more mathematical than theories.

In publishing results

Publication of scientific papers in scientific journals so that
  • The articles are widely disseminated around the world through subsciptions by scientific libraries and by individuals.
  • The articles are available for study for years, decades or even centuries - the article never goes "out of print".
  • There is "competitive publication": journals are keen to attract the best articles, and authors are keen to have articles accepted by the most prestigious journals. This maintains and improves standards.

Peer review of draft scientific manuscripts
Scientific articles when submitted for publication in a so-called peer-reviewed scientific journal are reviewed by two or three experts in the field of research who give their opinion (usually anonymously as far as the authors are concerned) on whether the article should be accepted for publication without alteration (which is rare), or should be accepted after some changes (more common), or should be rejected. This
  • prevents poor articles from being published
  • may raise standards since authors will know that their work will be assessed before publication
  • improves articles via the incorporation of helpful comments from the reviewers.

Reporting of all results fairly, not just those that fit some preconceptions.

Publishing the results of all completed trials, not just those that happen to fit in with a desired result.

Inclusion of a clear summary, generally at the start of the document.

Is this page fact or opinion?

What I have written on this page is my opinion of how science is generally carried out, based on personal experience and non-systematic observations, and how I feel that it should be carried out. So, each sentence could be preceded by "In my view, ...". There is a circular argument here in that if an activity does not meet my criteria, I do not consider it as part of science.

But, to varying degrees, the methods can be justified on the grounds of firstly plausibility, and secondly that we have compelling evidence that they have worked in the past (in generating predictions which have been found to be accurate).

References

[1] James Lind Library: Principles of testing (viewed 15.10.2013) http://www.jameslindlibrary.org/illustrating/records/the-book-of-daniel-chapter-11-16/key_passages
[2] NHS Wales: Glossary (viewed 20.11.13) http://www.wales.nhs.uk/sitesplus/922/page/49823: "BMI is calculated as weight (in kilograms) divided by the height squared (in metres). Adults with a BMI of 25 or more are categorised as overweight and with a BMI of 30 or more as obese."
[3] UICC TNM classification http://www.uicc.org/resources/tnm





Last updated: 10 Apr 2018