ChapterPDF Available

Scientific method

pairwise comparisons. As noted earlier, the
´should not be used for only pairwise
tests. However, both programs allow the user to
specify contrasts, although the user will have to
calculate the critical value himself or herself,
which is hardly a problem. SPSS allows the user
to specify contrasts for the ‘‘One-way ANOVA’’
procedure, and SAS allows the user to define
contrasts under ‘‘Proc GLM.’’ Neither program
will print out confidence intervals, so the user
needs to do that by hand.
The Changing Landscape
When Tukey started working on multiple compari-
son procedures, he largely set the field moving in
the direction of familywise error rates, which is
where Scheffe
´s test fits. But the problem with the
familywise error rate approach is that it is an
extreme case. If a family consists of a large number
of contrasts (pairwise or not), then any error is
considered a failure of the entire set—the confi-
dence intervals no longer cover all parameters.
That is why the critical values are as large as they
are. In the last decade of Tukey’s life, he began to
take an interest in Benjamini’s False Discovery
Rate, which aims to control the number of false
discoveries (Type I errors) that a procedure will
make, rather than focusing on allowing no errors.
As the field moves further in that direction, tests
such as Scheffe
´s, and perhaps even Tukey’s, are
likely to play a smaller role.
David C. Howell
See also A Priori Monte Carlo Simulation; Bonferroni
Procedure; FTest; Multiple Comparison Tests; Null
Hypothesis; Omnibus Tests; Post Hoc Comparisons;
Tukey’s Honestly Significant Difference (HSD)
Further Readings
Howell, D. C. (2010). Statistical methods for psychology
(7th ed.). Belmont, CA: Thomson/Wadsworth.
Maxwell, S. E., & Delaney, H. D. (2004). Designing
experiments and analyzing data (2nd ed.). Mahwah,
NJ: Lawrence Erlbaum.
´, H. (1953). A method for judging all contrasts in
the analysis of variance. Biometrika, 40, 87–104.
The term method derives from the Greek meta and
odos meaning following after, suggesting the idea of
order. Applied to science, method suggests the effi-
cient, systematic ordering of inquiry. Scientific
method, then, describes a sequence of actions that
constitute a strategy to achieve one or more research
goals. Relatedly, scientific methodology denotes the
general study of scientific methods and forms the
basis for a proper understanding of those methods.
Modern science is a multifaceted endeavor. A
full appreciation of its nature needs to consider the
aims it pursues, the theories it produces, the meth-
ods it employs, and the institutions in which it is
embedded. Although all these features are integral
to science, science is most illuminatingly character-
ized as method. Method is central to science
because much of what we have learned from sci-
ence has been acquired through use of its methods.
Our scientific methods have been acquired in the
course of learning about the world; as we learn,
we use methods and theorize about them with
increased understanding and success.
In this entry, scientific method is contrasted
with other types of method. Then, some criticisms
of the idea of scientific method are considered.
Thereafter, four major theories of scientific method
are outlined and evaluated. Finally, the place of
methods in science is addressed.
Four Methods of Knowing
The American philosopher and scientist Charles
Sanders Peirce maintained that there are four gen-
eral ways of establishing beliefs. The poorest of
these, the method of tenacity, involves a person
stubbornly clinging to a familiar idea when chal-
lenged. The belief is sustained by an attitude of
tenacity and unquestioned acceptance. The
method of authority maintains that ideas are held
to be true simply because they are the ideas of
a person who is deemed an expert or perceived to
be in a position of power. Peirce noted that this
method is superior to the method of tenacity,
because some beliefs can be fixed by adopting the
method. The apriorimethod,which is better than
both of the methods just mentioned, involves an
Scientific Method 1325
appeal to the powers of reason independent of sci-
entific observation. It involves accepting beliefs on
the grounds that they are intuitive, self-evident,
and based on reason rather than experience. The
method of science is the method that Peirce himself
advocated. It is superior to the other three meth-
ods because it establishes belief by appeal to an
external reality and not to something merely
human. Unlike the other methods, which are pre-
scientific, the method of science has the character-
istic of self-correction because it has built-in
checks along the way. For Peirce, only this method
has the ability to lead eventually to the truth.
Criticisms of the Idea of Scientific Method
Despite the importance of method in science, the
idea that there is a scientific method characteristic
of scientific inquiry has been the subject of many cri-
ticisms. Perhaps the most frequently voiced criticism
of scientific method is that there is no such thing as
the scientific method; that is, there can be no fixed
universal account of scientific method appropriate
for all disciplines and at all times. This criticism
should be readily accepted because it speaks against
an unrealistic view of scientific method.
Another prominent criticism of scientific
method was proposed by Karl Popper, who often
remarked that scientific method does not exist. By
this he meant that there is no method of discover-
ing a scientific theory, that there is no method of
verification, and that there is no method for estab-
lishing whether a hypothesis is probably true or
not. However, these claims are part of Popper’s
preference for a falsificationist construal of the
hypothetico-deductive method. These claims
might, or might not, be part of other conceptions
of scientific method. For example, advocates of an
inductive conception of scientific method will not
accept the first claim; those who accept the idea of
confirmation, as distinct from falsification, will
argue against the second claim; and Bayesian
methodologists will reject the third claim.
In a book, which was provocatively entitled
Against Method, Paul Feyerabend presented a dif-
ferent criticism of scientific method. He argued
that there are no methodological rules that are
part of scientific method that have not been bro-
ken at some time or other in the interests of genu-
ine scientific progress. Thus, for Feyerabend, the
only rule that does not inhibit progress is the rule
‘anything goes.’’ Feyerabend’s argument has been
endorsed by several commentators who are critical
of appeals to the importance of scientific method.
However, it should be noted that Feyerabend’s crit-
icism strictly speaks against the fixity of methodo-
logical rules only. There is nothing in Feyerabend’s
writing that would counsel against the flexible use
of a variety of different methodological rules that
are revisable in the light of experience, reason, and
other sources of justification.
None of the three criticisms just considered
addresses contemporary issues in scientific
Four Theories of Scientific Method
Modern scientific methodology has given consider-
able attention to the following four general theo-
ries of scientific method: (1) inductive method,
(2) hypothetico-deductive method, (3) Bayesian
hypothesis testing, and (4) inference to the best
Inductive Method
The idea that scientific method involves induc-
tive reasoning goes back at least to Aristotle, and
was given a heavy emphasis by Francis Bacon and
John Stuart Mill. Inductive reasoning takes differ-
ent forms. For example, it is to be found in the
fashioning of statistical generalizations, in the
Bayesian assignment of probabilities to hypotheses,
and in the reasoning involved in moving from data
to hypotheses in the hypothetico-deductive method.
The most popular inductive approach to scien-
tific method is sometimes called naı
¨ve inductivism.
According to this account of method, science
begins by securing observed facts, which are col-
lected in a theory-free manner. These facts provide
a firm base from which the scientist reasons
upward to hypotheses, laws, or theories. The rea-
soning involved takes the form of enumerative
induction and proceeds in accordance with some
governing principle of inductive reasoning. As its
name suggests, enumerative induction is a form of
argument in which the premises enumerate several
observed cases from which a conclusion is drawn,
typically in the form of an empirical generaliza-
tion. However, enumerative induction can also
1326 Scientific Method
take the form of a prediction about something in
the future or a retrodiction about something in the
past. The governing principle for an enumerative
induction to a generalization can be stated infor-
mally as follows: ‘‘If a proportion of As have been
observed under appropriate conditions to possess
property B, then infer the same proportion of all
As have property B.’’ This inductive principle can
be taken to underwrite the establishment of statis-
tical generalizations.
The naı
¨ve inductive method has been criticized
in various ways, although the criticisms are mostly
directed at extreme versions of the method—
versions making the claim that observed facts can
be known infallibly, that observations are made in
an entirely theory-free manner, or that empirical
generalizations can be secured through the use of
a strongly justified principle of induction. How-
ever, the so-called naı
¨ve inductive method can be
defended in a moderate form: Observed facts can
be established reliably, if fallibly; theory has to be
used to guide observations, and theoretical terms
can be used to report observational statements
without threatening the reliability of those state-
ments; and principles of induction can be given an
adequate justification on pragmatic grounds.
In the behavioral sciences, the radical behavior-
ism of Burrhus F. Skinner is a prominent example
of a research tradition that makes use of a nonsta-
tistical inductive conception of scientific method.
The major goals of radical behaviorist research are
first to detect empirical generalizations about
learning, and then to systematize those empirical
generalizations by assembling them into nonex-
planatory theories. Murray Sidman’s Tactic s o f Sci-
entific Research is an instructive radical
behaviorist account of the methodology of phe-
nomena detection.
The Bayesian approach to hypothesis testing
can be regarded as a sophisticated variant of
inductive method. It is considered later as an
account of scientific method in its own right.
Hypothetico-Deductive Method
The most popular account of method in science
is the hypothetico-deductive method, which has
been the method of choice in the natural sciences
for more than 150 years. The method has come to
assume hegemonic status in the behavioral sciences,
which have often placed a heavy emphasis on test-
ing hypotheses in terms of their predictive success.
Relatedly, the use of traditional statistical signifi-
cance test procedures is often embedded in
a hypothetico-deductive structure.
The hypothetico-deductive method is character-
istically described in one of two ways: On one
account, the scientist takes a hypothesis or a theory
and tests it indirectly by deriving from it one or
more observational predictions, which are amena-
ble to direct empirical test. If the predictions are
borne out by the data, then that result is taken as
a confirming instance of the theory in question. If
the predictions fail to square with the data, then
that fact counts as a disconfirming instance of the
theory. The second account is from Karl Popper,
who construes the hypothetico-deductive method
in falsificationist terms. On this rendition, hypoth-
eses are viewed as bold conjectures, which the sci-
entist submits to strong criticism with a view to
overthrowing or refuting them. Hypotheses that
successfully withstand such criticism are said to be
corroborated, which is a noninductive notion of
Even though the hypothetico-deductive method
is used by many scientists and has been endorsed
by prominent philosophers of science, it has
received considerable criticism. Leaving aside Pop-
per’s less influential falsificationist account of the
hypothetico-deductive method, the major criticism
of the hypothetico-deductive method is that it is
confirmationally lax. This laxity arises from the
fact that any positive confirming instance of
a hypothesis obtained through its use can confirm
any hypothesis that is conjoined with the test
hypothesis, irrespective of the plausibility of that
conjunct. Another criticism of the hypothetico-
deductive method is that it submits a single
hypothesis to critical evaluation without regard for
its performance in relation to possible competing
hypotheses. Yet another criticism of the method is
that it mistakenly maintains that hypotheses and
theories arise through free use of the imagination,
not by some rational, methodological, or logical
Criticisms such as these have led some methodol-
ogists to recommend that the hypothetico-deductive
method should be abandoned. Although this might
be a reasonable recommendation about the method
as it is standardly conceived, it is possible to correct
Scientific Method 1327
for these deficiencies and use the method to good
effect in hypothesis testing research. For example,
one might overcome the confirmational defects of
the orthodox hypothetico-deductive method by
employing a Bayesian approach to confirmation
within a hypothetico-deductive framework. With or
without a commitment to the Bayesian approach,
one could use the hypothetico-deductive method to
test two or more competing hypotheses deliber-
ately in relation to the evidence, rather than one
hypothesis in relation to the evidence. Further-
more, in testing two or more hypotheses, one
might supplement the appeal to empirical ade-
quacy by invoking criteria to do with explana-
tory goodness.
Bayesian Method
Although the Bayesian approach to evaluating
scientific hypotheses and theories is looked on
more favorably in philosophy of science than the
hypothetico-deductive alternative is, it remains
a minority practice in the behavioral sciences.
For the Bayesian approach, probabilities are
considered central to scientific hypothesis and
theory choice. It is claimed that they are best
provided by probability theory, which is aug-
mented by the allied philosophy of science
known as Bayesianism. In using probability the-
ory to characterize theory evaluation, Bayesians
recommend the assignment of posterior proba-
bilities to scientific hypotheses and theories in
the light of relevant evidence. Bayesian hypothe-
sis choice involves selecting from competing
hypotheses the one with the highest posterior
probability, given the evidence. The vehicle
through which this process is conducted is Bayes’
theorem. This theorem can be written in a simple
form as: Pr(H/D) =Pr(H) ×Pr(D/H) Pr(D).
The theorem says that the posterior probability of
the hypothesis is obtained by multiplying the prior
probability of the hypothesis by the probability of
the data, given the hypothesis (the likelihood),
and dividing the product by the prior probability
of the data.
Although Bayes’ theorem is not controversial as
a mathematical theorem, it is controversial as
a guide to scientific inference. With respect to the-
ory appraisal, one frequently mentioned problem
for Bayesians is that the probabilistic information
required for their calculations on many scientific
hypotheses and theories cannot be obtained. It is
difficult to know how one would obtain credible
estimates of the prior probabilities of the various
hypotheses and evidence statements that com-
prised, say, Charles Darwin’s evolutionary theory.
Not only are the required probabilistic estimates
for such theories hard to come by, but also they do
not seem to be particularly relevant when apprais-
ing such explanatory theories.
The problem for Bayesianism presented by sci-
entific theory evaluation is that scientists naturally
appeal to qualitative theoretical criteria rather
than probabilities. It will be described in the next
section that scientific theories are often evaluated
qualitatively by employing explanatory reasoning
rather than probabilistic reasoning.
Inference to the Best Explanation
In accordance with its name, inference to the
best explanation (IBE) is founded on the belief that
much of what we know about the world is based
on considerations of explanatory worth. In con-
trast to the hypothetico-deductive method, IBE
takes the relation between theory and evidence to
be one of explanation, not logical entailment, and
by contrast with the Bayesian approach, it takes
theory evaluation to be a qualitative exercise that
focuses explicitly on explanatory criteria, not
a quantitative undertaking in which one assigns
probabilities to theories. Given that a primary
function of many theories in science is to explain,
it stands to reason that the explanatory merits of
explanatory theories should count in their favor,
whereas their explanatory failings should detract
from their worth as theories. The major point of
IBE is that the theory judged to be the best expla-
nation is taken as the theory most likely to be cor-
rect. There is, then, a two-fold justification for
employing IBE when evaluating explanatory theo-
ries: It explicitly assesses such theories in terms of
the important goal of explanatory power, and it
provides some guide to the approximate truth of
The cognitive scientist Paul Thagard has devel-
oped a detailed account of IBE as a scientific
method—one that helps a researcher to appraise
competing theories reliably through the coordi-
nated use of several criteria. This method is known
1328 Scientific Method
as the theory of explanatory coherence. The theory
comprises an account of explanatory coherence in
terms of many constituent principles, a computer
program for implementing the principles, and vari-
ous simulation studies that demonstrate its prom-
ise as a method of IBE.
According to the theory of explanatory coher-
ence, IBE is centrally concerned with establishing
relations of explanatory coherence. To infer that
a theory is the best explanation is to judge it as
more explanatorily coherent than its rivals. The
theory of explanatory coherence is not a general
theory of coherence that subsumes different forms
of coherence such as logical and probabilistic
coherence. Rather, it is a theory of explanatory
coherence, where the propositions hold together
because of their explanatory relations.
Relations of explanatory coherence are estab-
lished through the operation of seven principles.
These principles are symmetry, explanation, anal-
ogy, data priority, contradiction, competition, and
acceptability. The determination of the explana-
tory coherence of a theory is made in terms of the
three criteria. Within the theory of explanatory
coherence, each of these criteria is embedded in
one or more of the seven principles.
Thagard determined that explanatory breadth is
the most important criterion for choosing the best
explanation. This criterion captures the idea that
a theory is more explanatorily powerful than its
rivals if it explains a greater range of facts.
The notion of simplicity that Thagard deems
most appropriate for theory choice is a pragmatic
criterion that is closely related to explanation; it is
captured by the idea that preference should be
given to theories that make fewer special or ad hoc
assumptions. Thagard regards simplicity as the
most important constraint on explanatory breadth;
one should not sacrifice simplicity through ad hoc
adjustments to a theory to enhance its consilience.
Finally, Thagard found that analogy is an
important criterion of IBE because it can improve
the explanation offered by a theory. Explanations
are judged more coherent if they are supported by
analogy to theories that scientists already find
The four theories of scientific method just con-
sidered are commonly regarded as the major theo-
ries of scientific method. Although all the methods
have sometimes been proposed as the principal
claimant for the title of the scientific method, they
are better thought of as restrictive accounts of
method that can be used to meet specific research
goals, not broad accounts of method that pursue
a range of research goals.
The Importance of Method
Even though methodological discussions of sci-
entific method are not fashionable, they are of
vital importance to the well-being of science. For
it is to scientific methods that scientists naturally
turn for the cognitive assistance they need to
investigate their subject matters successfully. The
evolution and understanding of scientific meth-
ods is to be found in the domain of scientific
methodology; this fact makes this interdisciplin-
ary sphere of learning of major practical and
educational importance.
Brian D. Haig
See also Alternative Hypotheses; Bayes’s Theorem;
Falsifiability; Inference: Deductive and Inductive;
Logic of Scientific Discovery, The
Further Readings
Haig, B. D. (2005). An abductive theory of scientific
method. Psychological Methods, 10, 371–388.
Howson, C., & Urbach, P. (2006). Scientific reasoning:
The Bayesian approach (3rd ed.). LaSalle, IL: Open
Laudan, L. (1981). Science and hypothesis: Historical
essays on scientific methodology. Dordrecht, the
Netherlands: Reidel.
Nickles, T. (1987). Twixt method and madness. In N. J.
Nersessian (Ed.), The process of science (pp. 41–67).
Dordrecht, the Netherlands: Martinus Nijhoff.
Nola, R., & Sanky, H. (2007). Theories of scientific
method. Montre
´al, Que
´bec, Canada: McGill-Queens
University Press.
Rozeboom, W. W. (1999). Good science is aductive, not
hypothetico-deductive. In L. L. Harlow, S. A. Mulaik,
& J. H. Steiger (Eds.), What if there were no
significance tests? (pp. 335–391). Hillsdale, NJ:
Lawrence Erlbaum.
Sidman, M. (1960). Tactics of scientific research. New
York: Basic Books.
Thagard, P. (1992). Conceptual revolutions. Princeton,
NJ: Princeton University Press.
Scientific Method 1329
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
A broad theory of scientific method is sketched that has particular relevance for the behavioral sciences. This theory of method assembles a complex of specific strategies and methods that are used in the detection of empirical phenomena and the subsequent construction of explanatory theories. A characterization of the nature of phenomena is given, and the process of their detection is briefly described in terms of a multistage model of data analysis. The construction of explanatory theories is shown to involve their generation through abductive, or explanatory, reasoning, their development through analogical modeling, and their fuller appraisal in terms of judgments of the best of competing explanations. The nature and limits of this theory of method are discussed in the light of relevant developments in scientific methodology.
For years I have urged that there is more to scientific method than meets the philosophical eye. Some highly touted conceptions of method — the hypothetical-deductive (H-D) method, for instance — are remarkably thin when it gets down to details. For when one studies real scientific cases of problem solving and theory construction, the methodological features are almost always more substantive and more interesting than what one reads in general methodological accounts. There are more things in heaven and earth...’
What is it to be scientific? Is there such a thing as scientific method? And if so, how might such methods be justified? Robert Nola and Howard Sankey seek to provide answers to these fundamental questions in their exploration of the major recent theories of scientific method. Although for many scientists their understanding of method is something they just “pick up”; in the course of being trained, Nola and Sankey argue that it is possible to be explicit about what this tacit understanding of method is, rather than leave it as some unfathomable mystery. They robustly defend the idea that there is such a thing as scientific method and show how this might be legitimated. The book begins with the question of what methodology might mean and explores the notions of values, rules and principles, before investigating how methodologists have sought to show that our scientific methods are rational. Part 2 of the book sets out some principles of inductive method and examines its alternatives including abduction, IBE, and hypothetico-deductivism. Part 3 introduces probabilistic modes of reasoning, particularly Bayesianism in its various guises, and shows how it is able to give an account of many of the values and rules of method. Part 4 considers the ideas of philosophers who have proposed distinctive theories of method such as Popper, Lakatos, Kuhn and Feyerabend and Part 5 continues this theme by considering philosophers who have proposed “naturalised”; theories of method such as Quine, Laudan and Rescher. The book offers readers a comprehensive introduction to the idea of scientific method and a wide-ranging discussion of how historians of science, philosophers of science and scientists have grappled with the question over the last fifty years.
This is an updated, revised and enlarged edition of Howson and Urbach's account of scientific method from the Bayesian standpoint. The book offers both an introduction to probability theory and a philosophical commentary on scientific inference. This new edition includes chapter exercises and extended material on topics such as regression analysis, distributions densities, randomisation and conditionalisation.
Science and hypothesis: Historical essays on scientific methodology
  • L Laudan
Laudan, L. (1981). Science and hypothesis: Historical essays on scientific methodology. Dordrecht, the Netherlands: Reidel.
Good science is aductive, not hypothetico-deductive What if there were no significance tests?
  • W W Rozeboom
Rozeboom, W. W. (1999). Good science is aductive, not hypothetico-deductive. In L. L. Harlow, S. A. Mulaik, & J. H. Steiger (Eds.), What if there were no significance tests? (pp. 335–391). Hillsdale, NJ: Lawrence Erlbaum.