Siu L. Chow Keywords: Causation, Control, Correlation, Deduction, Experiment, Induction, Measurement, Non
Siu L. Chow Keywords: Causation, Control, Correlation, Deduction, Experiment, Induction, Measurement, Non
Siu L. Chow Keywords: Causation, Control, Correlation, Deduction, Experiment, Induction, Measurement, Non
IN Methods in
Psychological Research, In Encyclopedia of Life Support Systems (EOLSS), Eolss Publishers,
Oxford, UK. [http://www.eolss.net]
Siu L. Chow
Department of Psychology, University of Regina, Canada
Keywords: causation, control, correlation, deduction, experiment, induction, measurement, non-
experiment, quasi-experiment, objectivity, prediction, psychometrics, random sampling distribution,
reductionism, research design, regression, relativism, reliability, sampling, statistics, substantive
impact, theoretical prescription, validity, variables
Contents
1. Introduction
2. The Confluence of Issues from Different Domains
3. An Overview of Research Methods
4. Reasons for the Multiplicity of Methods
5. Philosophical Issues
6. Some Methodological Issues
7. Current Trends in Methodology
Summary
1
quantitative data as evidential support for qualitative theories? How do psychologists generalize
from their data that are collected in an artificial setting to real-life phenomena? What is the rationale
of experimentation in psychological research? How can psychologists assess their research?
1. Introduction
Conducting research differs from informal gathering of information in that researchers collect data
systematically so as to answer a well-defined question. The research is systematic if data are
collected in strict accordance with a predetermined plan or design. The question is well defined if it
originates from an unambiguous theoretical statement. These requirements seem reasonable for
studying tangible and durable phenomena (e.g. a floating vessel). However, they would seem at first
blush impossible for psychologists to conduct research in view of the following challenges.
First, much of what interests psychologists is not observable in the way physical objects like chairs
and cars are tangible. To skeptics, the incorporeal nature of mental phenomena (e.g. memory,
drives, thinking, etc.) precludes them from being observed or measured. This, in turn, renders it
impossible to study them systematically. Second, individual human beings are unique and
unpredictable in a way physical objects are not. It would seem inappropriate to strive for general
statements, let alone functional laws or explanatory theories that apply to all human beings without
exception. Third, unlike inanimate objects whose nature is relatively durable, psychological
phenomena are colored by fleeting wants, desires, fears, and the like. Is it possible to study
inconstant phenomena systematically and objectively? Fourth, while it may be possible to study a
complex object bit by separate bit if the object is the sum of its components, this analytical
approach is inappropriate for psychology because individual character has a unity that is more than
the sum of its parts. For example, taking away Smith’s motives (if it were possible), psychologists
are no longer studying Smith. Conversely, it is not meaningful to talk about Smith without taking
into account Smith’s motives. Fifth, human affairs are value laden. Can, or should, psychological
research be objective?
This introduction to the research methods used in psychology begins with how the confluence of
psychologists’ conceptual, logical, meta-theoretical, theoretical, and statistical skills serves to
ensure that research data are collected systematically to answer well-defined questions. Next is a
summary of the multiplicity of research methods at the disposal of psychologists. The philosophical
and methodological assumptions underlying the aforementioned reservations about psychological
research are considered before the discussion of how research findings are assessed. The discussion
concludes with an affirmation of the primacy and importance of theoretical considerations.
A description of various skills implicated in conducting research provides the backdrop for
subsequent discussion of various philosophical and methodological issues implicated in
psychological research. It is also essential for understanding why psychologists do things the way
they do.
Psychologists begin a research project when confronted with a phenomenon that cannot be readily
accommodated by the existing conceptual scheme. For example, why does the normally reserved
Joe become an ardent activist who participates in protests? Why do drivers sometimes “see” the red
traffic light after they are half a block past the intersection (see Experimentation in Psychology--
Rationale, Concepts and Issues)? The first step in conducting research is the conceptual task of
coming up with a conjecture that makes sense of the to-be-understood phenomenon. Note that the
phenomenon exists before the research, and that the research is about an explanation of the
phenomenon, not the phenomenon itself (for a discussion of the analogous distinction between the
2
interpretation of test scores and the test scores themselves in The Construction and Use of
Psychological Tests and Measures).
In Psychologist A’s view, Joe’s new found activism is the expression of his previously suppressed
aggressiveness. However, Psychologist B considers the change as a sign that Joe has overcome a
crippling inhibition. Can the two explanations be true at the same time? Which one is correct if they
are incompatible? At the same time, both may be incorrect. For example, Joe’s change may be
brought about by a newly acquired insight into social policies. It is a sophisticated conceptual task
to design appropriate research to sort out these questions.
Psychologist C may suggest that Joe’s previous reserve was, in fact, aloofness, which camouflaged
Joe’s aggressiveness. Such a stance raises the conceptual issue as to how it is possible that
aggressiveness can account for two apparently unrelated, if not opposite, phenomena (aloofness and
activism). A related conceptual issue is whether or not aggressiveness would be evoked to explain
Joe’s reserve had his demeanor been characterized as shy instead of aloof. That is, the direction of
the research depends on how the phenomenon is conceptualized in the first place.
Native speakers of a language are not necessarily good at explaining what makes an utterance
grammatical in the native language. This is because the ability to analyze and talk about natural
language is a meta-linguistic skill different from the ability to use the language. Likewise, a
different conceptual skill is required to explicate or analyze the rationale of psychological research,
namely, a familiarity with some logical rules and relationships, as well as some methodological
issues.
Deductive logic consists of rules that make it possible to derive a valid conclusion from a set of
premises. Of particular interest are the four conditional syllogistic paradigms (or argument forms)
shown in Table 1. The modus tollens argument form may be used to introduce the terms necessary
for subsequent discussion.
Table 1. Four conditional syllogistic arguments whose major premise is a conditional proposition
3
Conclusion: B is definitely true.
4 Denying the antecedent:
Major premise: If A is true, then B is found.
Minor premise: –A (A is not true).
Conclusion: B’s truth-value is indeterminate.
Consider the meanings of “truth-value” and “categorical proposition” with reference to [E1] and
[E2]:
[E1]: An automobile has four wheels.
[E2]: If Theory T is true, then Implication I is true.
If what is described in [E1] is indeed the case, E1’s truth-value is true. If what [E1] describes is not
the case, [E1] has “false” as its truth-value. [E1] is a categorical proposition in the sense that its
truth-value is determined solely by what it says without referring to any other proposition. This is
not the case with [E2], which has three components: the two categorical propositions, “Theory T is
true” and “Implication I is true,” and the logical operator “If . . . then . . .”
The operator “If . . . then . . .” stipulates the logical operations that are permissible on the antecedent
and consequent. [E2] is a conditional proposition in the sense that its truth-value is contingent on
the truth-values of its antecedent and consequent. The proposition that follows “If” is the
antecedent, whereas the proposition that follows “then” is the consequent of the conditional
proposition. As may be seen from Table 2, the conditional proposition is false only when it has a
true antecedent and a false consequent.
Truth-value of the
Truth-value of antecedent Truth-value of consequent
conditional proposition “If A
“A is true” “B is found”
is true, then B is found”
True True True
True False False
False True True
False False True
An argument is a syllogism if it is made up of three propositions such that the first two (called the
“major premise” and “minor premise,” respectively) form the basis for deriving the third one (the
conclusion). A conditional syllogism is one whose major premise is a conditional proposition. The
inferential foundation of empirical research is the conditional syllogism.
As may be seen from Table 1, the categorical proposition used as the minor premise of the
conditional syllogism may affirm or deny the truth of the antecedent or consequent. Such a state of
affairs gives rise to four argument forms. The modus tollens form says that denying the truth of the
consequent warrants concluding that the antecedent is false. The modus ponens form says that
affirming the truth of the antecedent justifies accepting the consequent as true. No definite
conclusion can be drawn when the consequent is affirmed or when the antecedent is denied (see
Quasi-Experimentation).
A graphical representation may be used to overcome the counter-intuitive feel of the “affirming the
consequent” and “denying the antecedent” paradigms. For the purpose of the present exposition, “If
4
T, then I” means that all members of T are members of I. This is represented in Figure 1 by placing
the smaller oval T inside the larger oval I. The area outside oval T is –T. By the same token, the
area outside oval I is –I. The elements x, y, and z represent particular members of T, I, and –I,
respectively. The four syllogistic argument forms shown in Table 1 may now be explicated as
shown in Figure 1.
In terms of Figure 1, to assert –I is to say that there is an element that is not a member of I. This is
the assertion of the minor premise of the modus tollens argument form. Such an element may either
be y or z. As both y and z are outside oval T, neither of them can be a member of T. This shows
why denying I (i.e. asserting –I) warrants the –T conclusion in the modus tollens paradigm.
To say that I is true, amounts to saying that there is an element that is a member of I. Both x and y
fill the bill. However, while x is member of T, y is not. In other words, knowing that there is a
member of I does not warrant the conclusion that there is a member of T. Hence, affirming the
consequent leaves indeterminate the truth-value of the antecedent of the conditional proposition.
The modus ponens rule may be explicated in like manner. Being a member of T, x is also a member
of I by virtue of the inclusion of T in I. Hence, if T is true, I is true. Although neither y nor z is a
member of T (i.e. –T), y is nonetheless a member of I. Consequently, knowing that there is an
element that is not a member of T is not informative as to its membership in I. Hence, denying the
antecedent of “If T, then I” leaves the truth-value of I indeterminate.
The proposition “All X are Y” is a universal (or general) proposition in the sense that it makes an
assertion about all tokens of X, namely, that they are all members of Y. In contrast, “Some X are
Y” is a particular statement that asserts that only some tokens of X are members of Y. The function
of inductive logic is commonly understood as a means that enables us to leap from the particular
proposition to its universal counterpart. For example, having seen white swans without exception in
the past, the budding naturalist may conclude (albeit incorrectly) that “All swans are white.” That is,
induction in such a view is generalization on the basis of enumeration. However, this commonly
held view is unsatisfactory.
For example, what is the minimal number of white swans that warrants the generalization?
Psychologists actually use more sophisticated inductive rules to reduce ambiguities in their
explanation of data (see Experimentation in Psychology--Rationale, Concepts, and Issues).
5
2.2.4. Three Logical Relationships
References are often made in theoretical discussions to one of three logical relationships between
two variables or between a variable and a phenomenon: (a) the necessary condition, (b) the
sufficient condition, and (c) the necessary and sufficient condition. Each of them (particularly the
necessary and sufficient condition) at various times has been identified with the causal agent.
Z is the necessary condition for E if, in the absence of Z, E cannot occur. Suppose that Z refers to
studying conscientiously and E stands for passing the examination. Suppose that students who do
not study conscientiously do not pass the examination. In such an event, studying conscientiously is
the necessary condition for passing the examination.
W is the sufficient condition for Y when Y always occurs in the presence of W. An example may be
the observation that whenever an individual gives birth, the individual is always a female. That is, Y
refers to being a female, and W stands for giving birth. (However, as will be seen in Section 2.2.5.
Formal Rules and Relationships in Theoretical Discussion, giving birth is not the sufficient
condition for being a female.)
Variable A is the necessary and sufficient condition for Event B if (a) Event B cannot occur in the
absence of Variable A, and (b) Event B occurs whenever Variable A is present. Suppose that A
refers to water, and B stands for the flourishing garden. In the presence of plenty of sunshine and
good soil, water is the necessary and sufficient condition for the flourishing garden.
The logical rules and relationships described are applicable to all subject matters. This shows that
logic and substantive issues belong to different domains. Specifically, knowing that water is the
necessary and sufficient condition for the garden to flourish is to become knowledgeable of a
formal relationship between water, sunshine, soil quality, and plant growth. There still remains the
substantive question as to why water bears such a relationship to plant growth, given sunshine and a
good soil. It follows that the necessary and sufficient condition cannot be identified with the cause.
Similarly, neither a sufficient nor a necessary condition should be so identified. A logical relation is
simply not a causal relationship.
The seahorse’s breeding method is unusual because it is the males who give birth. Hence, it is
incorrect to say that being a mother is a sufficient condition for being a female. This counter
example shows that the “induction by enumeration” procedure is an unsatisfactory means for
establishing knowledge. It takes only one negative example to give the lie to the putative sufficient
relationship between motherhood and being a female.
Likewise, the four conditional syllogisms are formal rules. Indicative of their having no substantive
import is the fact that they are used in the same way, regardless of what T and I represent or how T
and I are obtained. This does not pose any problem for logicians because they do not have to
consider what T or I says or why Propositions T and I have the implicative relationship entailed in
“If T, then I.” Nor have they to explain why the minor premise is I or –I. Nonetheless, the logical
relations and four conditional syllogistic argument forms are implicated in theoretical discussion
(see Quasi-Experimentation).
Psychologists use conditional syllogisms to assess the relationship between the theory, the
experimental hypothesis, and research data (see Experimentation in Psychology--Rationale,
Concepts, and Issues). Consequently, psychologists have to specify the substantive meanings of A
and B when “If A, then B” is used as the major premise. Deriving implications of this sort depends
on having an unambiguous theory to begin with, as well as on the psychologists’ theoretical
sophistication.
6
2.3.1. Criteria of Falsification
Reference is made to the well definedness of the research question. Such a question is based on a
testable statement that, in turn, is an implication of the to-be-corroborated theory. The meaning of
“well definedness” may be explicated as follows. Suppose that there are four theories about
atmospheric events: T1, T2, T3, and T4. Further suppose that, in the context of a particular
geographical locale, they imply I1, I2, I3, and I4, respectively:
It will rain [I1]
It will rain tomorrow, given the right conditions [I2]
It will rain on April 1, 2020 [I3]
There will be 2.5 cm of rain on April 1, 2020 [I4]
Theory T3 implies [I3] in the sense that T3 cannot be true if [I3] is not true. Although the common
practice is to say that [I3] is a “prediction” of T3, it is more correct to say that [I3] is a theoretical
prescription of T3 because [I3] is not a forecast of an event in the future. Instead, [I3] instructs
researchers to declare that T3 is not tenable if [I3] is not observed. It is in this sense that an
implication of the theory is a criterion of rejection or falsification of the theory.
Suppose that there is 1 cm of rain on April 1, 2020. This leads to the rejection of T4, but not T3. For
this reason, [I4] is a superior criterion of rejection to [I3] because it is more stringent by virtue of its
being more specific. This is the consequence of T4 being formulated better theoretically than T3.
Nonetheless, both [I3] and [I4] are well-defined statements because they stipulate specific,
unambiguous criteria for rejecting a theory. This is not the case with [I2] or [I1].
In the absence of an independent index for the caveat “given the right conditions,” [I2] is true by
definition (and, hence, circular). That it does not rain may mean that the right conditions are absent.
On the other hand, the right conditions must be present when it rains. At the same time, it is not
possible to say that [I1] is false because of its imprecision. In short, statements like [I1] and [I2]
cannot be used as criteria of rejection in theory corroboration. What this says about theoretical
considerations is that theories like T1 and T2 that give rise to the likes of [I1] and [I2] as
implications, respectively, are not testable theories. Theoretical sophistication is required to avoid
entertaining non-testable theories.
7
corroboration process consists of a concerted effort to refute the theory. Psychologists’ confidence
in the theory increases as more and more deliberate and rigorous attempts to falsify the theory fail.
Such a concerted effort assumes the form of a series of theoretically informed and interrelated
investigations. They are collectively called “converging operations” because they converge on the
theory’s tentative tenability if none is rejected. Important to objectivity is the fact that diverse
investigators (who may not share the same faith in the theory at the beginning) contribute to the
converging operations.
For most psychologists, statistics is an essential research tool whose utility differs in different types
of research. Psychometricians use correlation to establish the validity and reliability of
psychometric tests. Two main issues in extracting the factors from among a set of correlations
coefficients are (a) the number of factors to be extracted, and (b) the nature of the rotation required
in extracting the factors (e.g. oblique, orthogonal, or any other means). The type of validity used in
substantiating the factors thus extracted (construct validity, performance validity, etc.) may
influence the construction of the items of the test. The nature of the test items, in turn, may
determine the type of reliability to be ascertained (see The Construction and Use of Psychological
Tests and Measures).
8
The correlation and regression coefficients are also used in path analysis and structural equations.
These two methods are assumed by many psychologists to be the statistical means of ascertaining
causal relations among a set of variables. However, “cause” is not used in its vernacular sense in
this context. Instead, by “causal,” path analysts have in mind a functional relationship identified in
the path diagram.
The most frequently used statistical procedures are various significance tests (e.g. the χ2 test, t-test,
F-test, etc.). They are procedures used to test whether or not it is warranted to rule out chance
influences on the data collected in different conditions. The essence of these decision rules is that
any test statistic is deemed too unlikely to be explainable in terms of chance influences if its
associated probability is α or smaller than α. The most commonly used value for α is 0.05 (see
Statistics and Its Role in Psychological Research).
Hence, a crucial consideration for psychological researchers is choosing an appropriate statistical
procedure. This decision should be made with reference to the design of the study, the level of
measurement used, and the statistical question being asked. For example, the dependent and
independent t-tests are used in conjunction with the related-sample and independent-sample, one-
factor two-level designs, respectively, when the data are at the interval or ratio scale. The non-
parametric Wilcoxon test and Mann-Whitney U test, respectively, should be used when ordinal data
are collected from two related samples and independent samples.
Despite their critical role in empirical research, a litany of complaints about significance tests has
accumulated since the 1960s. This general dissatisfaction with significance tests was important in
the formation of the Task Force on Statistical Inference by the American Psychological Association.
The more ardent critics of significance tests even advocate prohibiting their use. However, the
criticisms are deconstructionist (see Statistics and Its Role in Psychological Research).
Specifically, no critics have taken into account the fact that the probability basis of significance
tests is the sampling distribution of the test statistic. The criticisms are contentious because they
owe their impetus to a failure to distinguish (a) between statistical and non-statistical concerns, and
(b) between concepts at different levels of abstraction (for a discussion of the appropriateness of
using the standard error of measurement in estimating test reliability see The Construction and Use
of Psychological Tests and Measures).
Not only do psychologists have three types of research methods from which to choose, they also
have multiple options within each type of method. An overview of the research methods,
particularly those subsumed under “non-experiment,” commonly used in psychological research
may be found in Figure 2.
9
3.1. Methodological Evolution
Note that research methods in psychology are related in the sense that they are all means of
soliciting information about human nature from human participants. As a matter of fact, all methods
originate from introspection in which individuals are asked to describe as accurately as possible
what goes on in their minds when sensations or experiences of pains and emotions occur. The
difficulties with introspection became obvious when it was found that research participants had to
be trained extensively before data could be collected. Apart from being time consuming, the picture
obtained from very specialized participants may not be germane to obtaining an account of human
experience in general. Moreover, it is not possible to tell whether the introspective accounts thus
collected are accurate.
The need to collect objective, quantitative data about actual sensations efficiently and without being
beholden to the subjects’ articulateness led to the development of the classical psychophysical
methods (e.g. the method of limits). With some additional assumptions, the new (and presumably
more direct) psychophysical method (sensory scaling and inter-modality scaling) was introduced.
There is another dissatisfaction with introspection. Although interesting, an individual’s
introspective account of sensations or emotions may not be relevant to researchers’ concerns.
Hence, it seems necessary to provide a directive for subjects as to what to report. This is the
beginning of the unstructured interview with open-ended questions. Before long, it is realized that
there is the issue of comparability when more than one interviewee participates in the investigation
with the unstructured format. For example, how can interviewers ensure that the same questions are
asked in the same way of all interviewees? Moreover, the fruitfulness of this method depends too
much on the interviewees’ articulateness. The issue of comparability may be dealt with by
determining the exact set of to-be-used questions before the interview. The difficulty with
differential articulateness among interviewees may be overcome if all interviewees are given the
same set of response options. These solutions give rise to the structured interview. When the
structured interview is conducted in written form, researchers are using a questionnaire.
Crucial to experimentation are well-defined theories, from which research questions and
experimental hypotheses are deduced. This derivation process is not unlike the development of
formal proofs in mathematics. However, mathematicians do simulations when the model on which
they work does not permit any formal derivation. Alternatively, simulations may be used when the
derivation is too complicated. The same situation arises in psychological research. Psychologists
appeal to computer simulations when it is difficult, if not impossible, to derive experimental
hypotheses. This brief account shows that the multiplicity of research methods is partly the result of
10
how research methods evolve when researchers are confronted with difficulties with the current
method.
Any research without a valid comparison baseline is a non-experimental study. As may be recalled
from Figure 2, there are many non-experimental methods: interview, observation, measurement, or
the formalistic simulation method.
Non-verbal information or information about interpersonal interaction may be collected with one of
two methods of observation. Observers adopting the non-intrusive approach would endeavor to be
as inconspicuous as possible, and may record observations with the help of a behavioral checklist.
Psychologists would role-play in the course of collecting data when the participatory observation
method is used. Suppose that the study is about the stresses experienced by medical orderlies. In
11
carrying out a participatory observation investigation, researchers might work as medical orderlies
in order to observe the orderlies (see Interviewing and Observation).
3.2.2. Measurement
Euclidean geometry is a formal system that consists of (a) a minimal number of undefined terms
(i.e. terms whose meanings are taken for granted and understood even by someone who is not
knowledgeable in Euclidean geometry, such as “length” and “point”), (b) all other terms used in
geometry, defined in terms of the undefined terms (e.g. “angle”), (c) a minimal number of logically
independent axioms (postulations whose truth is taken for granted), and (d) all geometry
propositions being deductively derived from the axioms. The simulation approach to psychological
research (e.g. in the study of cognitive activities) shares these characteristics. The simulation
method differs from the deductive inference in Euclidean geometry in that the deductive logic in the
latter is replaced in the former by procedures that can be implemented with the computer.
In conducting experiments, psychologists manipulate the independent variable and measure the
dependent variable while holding the control variable constant. For example, psychologists
manipulate color in the experiment when (a) a subset of its values are used (e.g. pink, brown, and
beige), and (b) the three colors are used to set up three corresponding test conditions. This variable
is called the “independent” variable because the formation of the test conditions is not determined
by what the subject does.
12
3.3.2. Control Variable
Suppose that the experiment is about the effect of color on mood. For theoretical reasons,
psychologists may decide to recruit male subjects within a particular age range only. Consequently,
the variables gender and age are represented at the same level in the test conditions defined by the
three colors in the present example. It is in this sense that experimenters hold constant gender and
age. Consequently, gender and age are the control variables of the experiment.
Consider the situation in which subjects are required to respond as quickly as possible to a stimulus
presented visually. This is the variable of reaction time (i.e. how long it takes subjects to respond).
Its exact value on any trial depends on the subjects’ readiness, perceptual speed, motor dexterity,
and the like. Reaction time is the dependent variable of the experiment.
The extraneous variables of the experiment are identified by exclusion. Specifically, any variable
that is neither the independent nor the dependent nor the control variable is an extraneous variable.
Consequently, there are logically an infinite number of extraneous variables. Fortunately,
psychologists can eliminate most of them on conceptual or theoretical grounds.
Furthermore, psychologists may also exclude some potential confounding variables with special
control procedures such as choosing the appropriate experimental design (the completely
randomized, repeated-measures, randomized block or mixed design), randomizing stimulus
presentation, counterbalancing, and the like. The sole purpose of these control procedures is to
ensure that extraneous variables may justifiably be excluded as potential explanations of the data.
A confounding variable is an extraneous variable that varies systematically with the independent
variable. For example, forgetting has been explained in terms of subsequent learning interfering
with previous learning. Hence, experimenters vary the amount of material to learn in the intervening
trials before testing. However, it takes more time to present more to-be-learned items, thereby
increasing the delay before initial learning and recall. The delay of testing is a variable that is
confounding with list-length under such circumstances.
Confounding variables in experimental studies are invariably discovered after the fact. Otherwise,
they would have been incorporated in the experimental design either as a control variable or as an
additional independent variable. It is quite different when applied psychologists have to collect data
in situ, where they often have to collect data with the knowledge that there is a confounding
variable.
The confounding variable may be used to introduce the quasi-experiment. Suppose that the quasi-
experiment is about the effects of color on employees’ moods in a department store. Psychologists
may not have the luxury of being able to assign subjects randomly to the three color conditions.
Instead, pink, beige, and brown are assigned to the cosmetic department, the kitchenware
department, and the hardware section, respectively. In such an event, color is confounded with the
type of clientele. Important to methodology is the fact that the three color conditions differ in more
than color. Consequently, any difference in the employees’ moods among the three conditions
cannot be attributed unambiguously to the difference in color (see Quasi-Experimentation).
13
It is the presence of confounding variables that distinguishes between experiment and quasi-
experiment. There is also the difference in nomenclature. Instead of being discussed in terms of
independent and dependent variables, quasi-experimental studies are described in terms of treatment
and outcome variables, respectively.
There are many intertwined reasons for the multiplicity of psychological research methods. To
begin with, any psychological phenomenon is open to multiple interpretations that differ in type,
complexity, or level of abstraction. For example, hyperactivity may be explained in terms of an
unhealthy diet, an improper reinforcement regime, or a central nervous system that is not adequately
stimulated. Other reasons for the wide range of research methods are the psychologists’ training,
meta-theoretical perspective, theoretical orientation, research objective, and vested interests.
Researchers come with diverse backgrounds, ranging from formal disciplines like mathematics to
deconstructionist literary criticism. They also bring with them different emphases. For example,
while some psychologists emphasize basic research, other psychologists may consider social
advocacy important. In addition to injecting new insights into psychological phenomena, these
diverse emphases also bring special methodological preferences. For example, psychologists with
extensive training in the physical or biological sciences would find the idea of the “constancy of
conditions” congenial (manipulating some variables in a consistent fashion while holding other
variables constant) (see Experimentation in Psychology--Rationale, Concepts, and Issues). They are
more likely to conduct experiments. Cognitive psychologists well versed in computing techniques
may find it warranted to run simulations to study mental processes. Psychologists who see human
affairs as a dramatic saga are more inclined to collect narrative data by conducting interviews with
open-ended questions.
Psychologists use theories to describe or explain mental or behavioral phenomena. This theorization
exercise is more abstract than the phenomena themselves. In planning their research or assessing
their research data, psychologists have to reflect on methodology or theories. This activity is more
abstract than theorizing about psychological phenomena. This higher-level frame of reference is the
meta-theory that has implications on the method used in research.
14
is the meta-theoretical impetus of many correlational studies in which a large number of participants
are measured on two or more variables, and the research objective is to ascertain the inter-
correlations among the variables. It is important to note that “law” in this “induction” perspective
refers to these correlations (or functional relations between variables). Hence, there is an emphasis
on theory discovery among psychologists who feel that the to-be-studied phenomenon is not well
understood (see The Construction and Use of Psychological Tests and Measures).
Although all experiments share the same formal structure (the experimental design, the
independent, dependent, and control variables, the protocol used to assign subjects to the test
conditions, etc.), they may be categorized as “atheoretical” or “theory-corroboration” experiments.
This distinction also owes its origin to the two meta-theoretical perspectives. Experimenters in the
“induction” tradition (e.g. Skinnerian behaviorists) conduct atheoretical experiments to establish
descriptive, functional relationships between variables, whereas experimenters subscribing to
“deduction” conduct theory-corroboration experiments (see Experimentation in Psychology--
Rationale, Concepts, and Issues).
In terms of their objectives, atheoretical experiments are like correlational studies. That is, the
procedure is to measure the same group of subjects on two or more variables, and the objective is to
ascertain the inter-correlations among the variables. However, appropriate controls are found in
atheoretical experiments, but not in correlational studies.
Adopting the “deduction” perspective, some psychologists conduct experiments to corroborate
explanatory theories. An explication of why it is misleading to represent the “deduction” stance by
the “theory → data” sequence may prove helpful in appreciating the idea of theory corroboration.
Upon encountering Phenomenon P, which cannot be explained in terms of existing knowledge,
psychologists propose a tentative theory (Theory T) that, if shown to be tenable, would account for
the phenomenon. This is the theory-discovery phase of the research. Theory T is then subjected to
rigorous tests in the corroborative phase (by obtaining evidential data, D). However, the discovery
and corroboration activities feed on each other. The collection of evidential data, D, depends on
what Theory T says. At the same time, additional nuances about Phenomenon P may be discovered
in the course of testing the original theory. This necessitates changing the theory. The modified
theory requires corroboration anew, a process that may unveil more nuances.
In short, it is more informative and correct to represent the “deduction” stance as the “phenomenon
→ theory ↔ evidential data” sequence. The double arrow sets in high relief three meta-theoretical
points. First, there is a constant interplay between theory discovery and theory corroboration.
Second, a distinction need be made between the original phenomenon for which the explanatory
theory is proposed and the research data that provide the evidential support for, or the criterion for
rejection of, the theory. Third, research deals with how psychologists theorize about phenomena,
not the psychological phenomena themselves.
Psychologists entertaining the “data → law” meta-theoretical perspective may differ among
themselves as to how the theory-discovery process is carried out. For example, those who subscribe
to the induction by enumeration approach would insist on collecting similar data (e.g. observing the
same thing in similar conditions or similar things in identical conditions). Replicability is crucial,
and the emphasis is on the frequency of replication. The inductive conclusion is assumed to be
stronger the more often replication is obtained.
At the same time, there is an alternative school of thought within the “induction” perspective that
may best be understood by referring to a practice adopted in historical research. Historians abstract
explanatory theories about events in the past. However, they do not treat their task as a matter of
looking for some recurrent characteristics among similar events. Instead, historians use the method
15
of adduction. That is, they look at different aspects of the event, collect information from diverse
sources, and utilize disparate information that may collectively inform them of an explanatory
account of the event.
Under the influence of literary criticisms, psychologists of the social constructionist persuasion
would find the adduction approach more congenial. Moreover, they may treat psychological
phenomena as dramas or “texts.” Consequently, not only would they conduct unstructured
interviews with open-ended questions, they may also collect observation data or utilize archival
materials. Text analysis becomes an integral part of psychological research (e.g. context analysis,
the grounded theory, etc.). In some instances, psychological research is not unlike an exercise in
hermeneutics.
Different psychologists may study the same phenomenon differently because they are committed to
some broad theoretical framework or system such as behaviorism, connectionism, or social
constructionism. While behaviorists conduct experiments or quasi-experiments, connectionists are
more likely to run computer simulations. Psychologists who subscribe to social constructionism
may be less inclined to consider experimentation than those who subscribe to connectionism. The
theoretical orientation has such an effect because it determines the level of analysis adopted by the
psychologists.
Consider the case of the psychology of language. One psychologist may be interested in linguistic
development, whereas another’s concern is the linguistic competence of adult native speakers of the
language. Given the rigid demands of the method, it may not be appropriate, if at all possible, to
study experimentally children at the one-word stage of language development. Moreover, the
crucial information would be what those children produce spontaneously, as well as the frequencies
of certain utterances. Hence, the observation method may be the method to use in research about the
neophyte’s linguistic performance. On the other hand, some theories of adults’ linguistic
competence are sophisticated enough to render experimentation possible.
As another example, some psychologists may find language production interesting while their
colleagues concentrate on language reception. While it is easy to design experiments for language
reception, the same is not true of language production studies. In short, the choice of a research
method may be determined by the psychologists’ theoretical orientation.
Psychologists conduct research for various purposes. Those who emphasize the experiential aspect
of psychological phenomena may be interested in achieving empathy through understanding
participants as unique individuals. Given such an idiographic research emphasis, those
psychologists are more inclined to conduct an in-depth interview with open-ended questions.
However, interview is too labor-intensive a method to use if the objective is to determine what
attitude adults share in general, despite individual differences. Consequently, psychologists may use
a theoretically informed questionnaire for such a nomothetic research.
Suppose that the concern is whether there is a link between the two multifaceted variables life style
and health. Crucial to this sort of epidemiological research is the representative nature of the
sample, the reliability, and the validity of the measures used. To satisfy such a need,
psychometricians’ main research concerns are the validity and reliability of psychometric
instruments. They rely heavily on statistical procedures other than significance tests (e.g. factor
analysis, path analysis, structural equations, etc.) (see The Construction and Use of Psychological
Tests and Measures).
Suppose that a psychologist is commissioned to ascertain whether a new method of training
improves a worker’s output in a particular factory. Note that generality is not a concern in view of
the specificity of the research question. Nor is the issue about why there is a change in job
16
performance. Moreover, the to-be-measured variable is the worker’s job performance itself. Given
these constraints, it is not possible to institute control procedures like random subject assignment,
choosing a theoretically informed experimental task, or dependent variable. Yet, the study would
have superficial resemblance to the experiment. Studies of this kind are quasi-experiments.
5. Philosophical Issues
Reference has been made to (a) “induction,” the “data → law” approach that produces data-driven
generalizations or laws, and (b) “deduction,” the “phenomenon → theory ↔ evidential data” stance
that gives rise to conceptually driven explanatory theories. Underlying the distinction between data-
driven generalizations and conceptually driven theories are several philosophical issues: (i) the
nature of the intervening variable, (ii) the reality of unobservable structures or processes, (iii) the
validity of using animals as theoretical models or analog, (iv) the lure of reductionism, and (v) the
possibility of objectivity if observations are theory dependent.
The onus is on cognitive psychologists to demonstrate the psychological reality of the unobservable
hypothetical mechanisms and processes. It helps to recall the following points. First, it is accepted
that psychological phenomena exist. Second, psychological theories are proposed to make sense of
the psychological phenomena. Third, research is about psychological theories, not the psychological
phenomena themselves. Fourth, a psychological theory is substantiated by demonstrating the
tenability of the hypothetical mechanisms and processes envisioned in the theory.
An implication of this recapitulation is that if the theory is true (i.e. the hypothetical mechanisms
and processes work in exactly the way stipulated in the theory) the consequences of applying the
17
mental apparatus in a specific context are well defined. That is, although unobservable, these
hypothetical mechanisms and processes (hence, “hypothetico”) should have tangible consequences
(“deductive”) if they are to be taken seriously. Hence, the mechanism of conducting research of the
“hypothetico-deductive” is to set up the specific context; the tenability of the theory is assessed by
comparing the actual consequences of the research manipulations with those stipulated by the
theory. The hypothetical mechanisms and processes are falsified if there is no match; the theoretical
intervening variables are deemed tenable, albeit tentatively, otherwise. The hypothetical structures
and processing are intervening variables in the sense that they come between the observable
research manipulations and the tangible consequences.
Understanding has two independent dimensions: experiential and conceptual. The experiential
dimension is characterized by the “Aha” feeling. Individuals who seek empathic solace would find
it acceptable to tune selectively to information that boostered certain emotions. Conceptual
understanding, on the other hand, is assessed in terms of objective evidence. Hence, acceptable
explanations are couched in non-experiential, non-teleological, and impersonal terms. This explains
the general public’s dissatisfaction with much of academic psychology. Likewise, those who look
for motives behind human behavior would find wanting the non-teleological explanations offered
by much of academic psychology.
Conceptual understanding is achieved by extending the current knowledge base through integrating
the new information into individuals’ preexisting frame of reference. The explanation is made up of
the explanandum (the thing to be explained) and the explanans (the element that does the
explaining). The understanding is tested, informally or formally, by testing the explanans with the
hypothetico-deductive procedure. Understanding is achieved only when the attempt is successful, in
other words, when the explanans can be substantiated independently of the explanandum, and the
explanandum “makes sense” in view of the explanans.
During the heyday of positivistic behaviorism, modus ponens was considered the paradigm of
explanation, as may seen from panel 1 of Table 3. The explanans consists of the conditional
syllogism whose major premise asserts the functional relationship between two variables. The
explicit minor premise affirms the antecedent of the major premise. Modus ponens permits
affirming what is said in the consequent of the major premise. The phenomenon is deemed
explained because it is shown to be the special case of the general functional relationship between
the antecedent and consequent of the empirical generalization.
Table 3. Modus ponens as the paradigm of behavioral explanations (panel 1) and modus tollens as
the paradigm of explanation in cognitive psychology (panel 2)
18
strength
Rat R’s bar-pressing response is
Minor
accompanied by a reinforcing stimulus
Suppressed A response that is readily emitted has
minor habit strength
Hence, Rat R’s bar-pressing response is
Explanation Conclusion
readily emitted
Realizing belatedly that one has just
Panel 2 1 Explananum Phenomenon
driven through a red light
If there is a very brief visual sensory
buffer, then sensory information about
2 Explanans
any properly registered visual stimulus
may be processed belatedly
If sensory information may be processed
Justification of
3 Major 1 belatedly, then there is partial-report
the explanans
superiority
If there is response bias in partial report,
4 Major 2
then there is partial-report superiority
5 Minor Partial-report superiority is obtained
The “response bias” explanation can be
Explanation of Experimental rejected by virtue of modus tollens.
6
data conclusions Sensory information may be processed
belatedly
The sensory information retained in a very
Explanation of
7 brief visual sensory store permits the
phenomenon
belated processing of the red light
For the positivistic explanation to work, the major premise of its explanans must be a token of
Aristotle’s efficient cause in view of the behaviorists’ belief that reinforcement shapes behavior.
However, in rejecting hypothetical structures and mechanisms, behaviorists cannot answer the
Humean skepticism about discovering the efficient cause by atheoretical observation. The Humean
argument is that no matter how hard one tries one can never see how one billiard ball causes another
billiard ball to move. One observes only the temporal and spatial contiguity of the two billiard balls’
movement. Moreover, that the contiguity has been found in the past does not guarantee that it will
be observed on the next occasion. The behaviorists’ implicit minor premise is not helpful because
“habit strength” is simply a logical construct in the sense that it is a merely a re-description of the
consequent of the major premise. To be useful, there must be an independent index of “habit
strength” other than the rat’s response readiness.
A more recent account of psychological explanation is shown in panel 2 of Table 3. Although the
appeal is still made to modus ponens, the emphasis is not on the formal relationship between the
two variables. Instead, the force of the explanation is based on the theoretical properties of the
hypothetical structure described in the antecedent of the major premise (brief visual sensory buffer).
Specifically, the explanation works by showing first how the explanandum (i.e. realizing belatedly
that one has just driven through a red light) can be accounted for if the explanans is tenable. The
boldface in row 2 of panel 2 highlights the fact that the explanans is the antecedent of the
conditional proposition (a very brief visual sensory buffer), not the entire conditional proposition.
The consequent of the conditional proposition (only sensory information about any properly
registered visual stimulus may be processed belatedly) is an implication of a theoretical property of
the explanans, which is used to make sense of the explanandum (see Appendix 1 of
Experimentation in Psychology--Rationale, Concepts, and Issues).
19
The italicized “can be” in the previous paragraph alludes to the second element of the legitimate
explanation. Can be is used instead of is because the explanandum can also be explained by other
explanans. Hence, for an explanation to be accepted, it must have evidential support at the same
time that its contenders are rejected by the evidence. That is, the explanans is justified only when
the theoretical property of the hypothetical structure is substantiated. This theory-corroboration
aspect of the explanation may be seen from the two conditional syllogisms depicted in rows 3
through 6 of panel 2.
Rows 3, 5, and 6 make up the conditional syllogism used to substantiate the sensory nature of the
very brief visual buffer. Rows 4, 5, and 6 constitute the conditional syllogism used to challenge the
sensory nature of the visual buffer in question. That is, the entries in rows 3 and 4 are the major
premises of two separate conditional syllogisms. The entry in row 5, supplied by the data of a
properly designed experiment, is the minor premise of both syllogisms. The data rejects, by virtue
of modus tollens, the syllogism headed by the entry in row 4. Hence, the tentative experimental
conclusion is the antecedent of the conditional proposition in row 3, which is an implication of the
hypothetical structure, the very brief visual buffer.
Behavioral psychology owes it attractiveness to the claim that they have understood behavior
because they can forecast (hence, can shape) behavior. However, there is neither self-evident nor
empirical reason for the behaviorists’ practice of defining “understanding” in terms of forecasting
or shaping behavior. The behaviorists’ explanation is wanting because there is no justification for
its explanans. The more serious problem is that behaviorists have misled friends and foes alike
when they claim to have identified the efficient cause in their causal statement. Behaviorists’ laws
are simply functional relationships, not causal statements that have been substantiated.
Aristotle also recognized the material and formal causes. Pouring molten chocolate into the mould
of a rabbit results in a piece of chocolate in the form of a rabbit. The shape of the mold is the cause
of the shape of the solid chocolate by virtue of its form. This is an example of a formal cause.
Various musical notes are produced when a pianist hits different keys. The explanation in terms of
an efficient cause is inappropriate because the same striking action is used to produce the different
musical notes. The differential musical notes are determined by the properties of the different
strings of the piano. This is an example of Aristotle’s material cause.
Likewise, the sensory nature of the very brief visual buffer is an example of the material causes.
That is, in corroborating the properties of the very brief visual buffer, cognitive psychologists are
establishing a formal or material cause of behavior. In other words, understanding may be achieved
when an empirically substantiated explanation of a material or formal cause of behavior is
appreciated by the individual who seeks understanding. This sort of understanding occurs with or
without the “Aha” feeling.
Behaviorists’ psychology owes its empirical basis to studies with animals as experimental subjects.
The validity of their assumption depends on the applicability to all species of the functional
relationships between observable behavior and environmental variables thus established. The
behaviorists’ functional laws are true to the extent that “behavior” and “environmental variables”
are defined so broadly that the functional relationships become true by definition. For example, a
“response” is the behavior that is followed by reinforcement. This definition applies to the rat
pressing the bar in the Skinner box and to a student diligently studying when “environmental
variable” refers to the presence of a food pellet and passing the examination with high distinction,
respectively. However, such success is more apparent than real because, as has been shown, the
behaviorists’ account does not extend the knowledge base.
20
Nonetheless, it would be incorrect to dismiss animal studies for the simple reason that behaviorists
had oversold their promise. Given the similarity of the cat’s and the human visual systems, much
about the basic operation of human vision was acquired by using cats as subjects (e.g. the ideas of
visual field, the hierarchical organization of the visual system) (see Animals as Models or Analogs).
The presumed incompatibility between idiographic uniqueness and nomothetic constructs is more
apparent than real. In actual fact, the former cannot be established without the latter. For example,
Jane is deemed unique because she differs from her peers in one or more ways (e.g. dress code,
hobbies, demeanor, etc.) Any of these differences is predicated on a uniformity or a norm. For
instance, Jane may be unique because she wears a red blouse without a belt, whereas all her
colleagues wear white blouses with blue belts. An example of uniqueness with reference to a norm
is that Jane is taller than 99% of her peers. A subtler example is the case where Jane is unique
because she is so erratic that she cannot be expected to behave in the same way twice in the same
situation. This case owes its uniqueness to the assumption that unique individuals are nonetheless
consistent in their uniqueness. Jane simply departs from this common feature. In short, it is not
possible to talk about uniqueness in the absence of commonality. Nomothetic constructs deal with
what is true of human nature, despite individual differences.
21
language plays on its users). However, the physiological account becomes less convincing once it is
shown that mental terms are more than being just the product of the language game.
The force of the language-game argument is based on the putative invalidity of referential meaning.
Specifically, it is argued that terms such as “memory,” “thinking,” “wish,” and the like do not refer
to anything real. Instead, they are merely social acts of a particular sort. However, this meaning-as-
use view is debatable because social action presupposes meanings. Consider the following example.
Speech-act psychologists are correct in saying that “I like it that C is the case” is an act of avowal of
satisfaction. However, they are incorrect in saying that “like” does not refer to a particular
psychological phenomenon. For example, the same avowal may be expressed with “I am happy that
C is the case” or “I have no objection that C is the case.” That a particular mode of making the
avowal is chosen over other alternatives suggests that the speech act serves more than a mere
performative function. It also serves a referential function, namely, to say something about a
particular psychological phenomenon.
Stripped of its linguistic sidekick, physiological reductionism is founded on a host of correlations
between physiological and behavioral indices. For example, the blood flow or electrical activity in
region X in the frontal lobe is expected to change when subjects are given a particular cognitive
task to do. Short of any theoretical foundation, data of this sort are no more than correlation data. At
best, they demonstrate that the brain or the central nervous system is necessary for psychological
phenomena. They cannot give an adequate account of psychological phenomena. For example, how
are the differences between envy and jealousy reflected by physiological indices? Moreover, is it
possible to identify the psychological phenomenon that gives rise to a particular pattern of
physiological indices? How would a pattern of physiological indices represent sympathy?
Constructionism in literary criticism is the outcome of pushing to the extreme the view that beauty
is in the eye of the beholder. To such critics, a literary text does not refer to the reality envisaged by
its author. Instead, what the text says depends on what readers can read into it. Adopting such a
meta-theoretical position, some psychologists doubt the possibility of having an objective
psychology. In their favor, constructionist psychologists may avail themselves of the following
evidence.
Consider the situation in which John’s demeanor may suggest shyness or aloofness. How he is
being treated depends on the initial assessment he receives. Of interest is the fact that someone who
is used to social interactions is more inclined to see John as shy, whereas a timid person may see
John as aloof. That is, what one sees is the outcome of one’s own experience or perspective.
Constructionist psychologists may also point out the phenomena of perceptual constancies (e.g.
shape constancy, size constancy, etc.) visual illusions (e.g. the Muller-Lyer illusion), as well as the
effects of linear perspective on perception. That is, seeing per se is no guarantee of objectivity.
The “theory-dependent observations” view would indeed be fatal to objectivity if there were only
one omnibus theory. However, the theory to which experimental data owe their identity is not the
to-be-corroborated theory (see Experimentation in Psychology--Rationale, Concepts, and Issues).
It is assumed that readers understand that “validity” means correctness. That concerns about validity
are raised often in many contexts indicates that research can be wrong in more than one way.
Hence, a distinction is made between the internal validity and external validity of the research (see
Quasi-Experimentation).
22
6.1. Internal Validity
When research results are assessed in terms of the structure and implementation of the research
itself, the concern is its internal validity. There are two components to internal validity. First,
questions about the appropriateness and correctness of the experimental design or the non-
experimental plan are basically about whether recognized alternative interpretations of the data may
be excluded. As these questions are answered with reference to the inductive principle underlying
the design or plan, this component of internal validity may be called the inductive conclusion
validity.
Data from nomothetic studies are tabulated and analyzed with the appropriate statistical procedures.
Researchers have to choose the appropriate procedure with reference to (a) the level (or scale) of
measurement of the data, (b) the statistical question being asked, and (c) the design of the research.
These are questions about the statistical conclusion validity of the research.
There are also two components to the external validity of empirical research: generality and
construct or hypothesis validity.
Empirical research is conducted to answer questions about the world. Hence, research data may be
assessed as to their range of applicability. The generality of the conclusion of empirical research is
the extent to which it can be applied to other settings. It is for this reason that the issue of
representativeness becomes important. Representativeness is determined by both the
representativeness of the research participants and that of the situation in which data are collected.
However, it is important to note that generality is independent of the other component of external
validity.
It helps to talk about the second component of external validity separately for psychometric and for
quasi-experimental or experimental studies. Consider the psychometric research conducted to
establish an instrument for measuring aggression. It is imperative to ascertain that the test measures
aggression, not defensiveness or some other phenomenon. To the extent that the test measures
aggression, the test has construct validity.
It is a bit different with experimental research. Recall experiments are conducted to corroborate
theories (hypothesis), not the phenomena about which the theories are proposed. Hence, the data are
different from the to-be-studied phenomenon itself. Nonetheless, the data must be relevant to the
hypothesis. For this reason, it is more helpful to talk about the hypothesis validity of the data, that
is, the extent to which the data inform about the tenability of the to-be-tested hypothesis (see
Experimentation in Psychology--Rationale, Concepts, and Issues).
Psychologists are very conscious of their methodology, so much so that they constantly question
their research approach and develop new methods, particularly statistical methods. One reason is
that psychologists have expanded their research interests as a result of (i) involving themselves in
conducting studies that are strictly non-psychological (e.g. the epidemiological approach to life-
style or “wellness”), and (ii) collaborating with researchers from other disciplines (e.g. education
and sociology). The trend in some areas of research is to introduce non-methodological issues into
methodological considerations. Two notable examples are ecological and practical validity. These
two non-research issues are conflated with methodological issues, much to the detriment of
conceptual rigor, hypothesis validity, or conceptual clarity. Consequently, issues about internal
validity and external validity seem to have been neglected.
It is easy to show that many psychological investigations, particularly psychological experiments,
are artificial. The rhetorical question then asked by the skeptics relates to the usefulness of the data
thus obtained. This is the charge that research data collected by experimental psychologists may
23
lack ecological validity (i.e. people do not behave in the way depicted by the data because the data
are not collected in natural or real-life settings). The artificial setting in which experimental data are
collected can be defended (see Experimentation in Psychology--Rationale, Concepts, and Issues).
ffice it to say here that advocates of ecological validity have failed to appreciate the fact that
research is about theories (i.e. tentative explanations of phenomena), and not the phenomena
themselves.
Much of the contemporary dissatisfaction with using significance tests stems from the desire to use
research to inform practical decision making (e.g. whether to use a new method of teaching, a new
drug, etc.). While this is a legitimate concern, it should not be conflated with the proper role of
statistics. For example, some psychologists advocate using confidence-interval estimate and effect
size in addition to, if not in place of, tests of significance. They insist that experimenters should use
the sample-size tables envisaged in power analysis to determine the sample size. It is said that effect
size is more useful than the decision about statistical significance because the magnitude of the
effect has substantive impact, but not statistical significance. Note the two italicized phrases are not
synonymous (see Experimentation in Psychology--Rationale, Concepts, and Issues and Statistics
and Its Role in Psychological Research). It can be shown that (i) the tests of significance cannot
(and ought not) be replaced, (ii) there are fundamental differences between statistical decision and
causal inference, and (iii) statistics cannot be used to reduce conceptual ambiguities in data
interpretation.
Glossary
Adductive reasoning: The process of putting forward a theory by looking at different aspects of
the event, collecting information from diverse sources, and utilizing disparate information.
Atheoretical research: Research that is not guided by any theoretical perspective.
Categorical proposition: A proposition whose truth-value is assessed on its own.
Conditional proposition: A proposition that has an antecedent and a consequent. The two are
linked up by the logical operator “If . . . then.” Its truth-value is contingent on the truth-
values of its antecedent and consequent.
Consequent: The proposition that comes after “then” of the logical operator “If . . . then” of the
conditional proposition.
Control: A control is a means in research to exclude an explanation of the data.
Construct validity: The construct validity of a test is the extent to which it correlates with the
behavioral index of the to-be-measured psychological structure or mechanism.
Correlation: The association between two random variables.
Deduction: Some psychologists use “deduction” to represent the “theory → data” sequence of
research.
Deductive logic: A set of rules that render valid the derivation of a proposition from one or more
propositions.
Ecological validity: The extent to which various aspects of the research mimic their counterparts in
the real-life phenomenon.
Experiment: Empirical research that has two or more data-collection conditions that are identical
in all aspects but one.
Explanandum: The thing to be explained.
Explanans: The explaining element in an explanation.
External validity: The correctness of empirical research assessed in terms of criteria exogenous to
the research.
Factor: The statistical term for “variable.” The ways in which a factor may be identified are the
“levels” of the factor.
Functional relationship: The relationship between two variables expressed as a mathematical
function.
24
Hypothesis validity: The extent to which the theoretical process envisaged in the hypothesis is
supported by empirical evidence.
Idiographic research: Research about what is unique about an individual.
Induction: Some psychologists use “induction” to refer to the atheoretical “data → law” view of
research.
Inductive logic: Inductive logic consists of rules that permit the exclusion of unwarranted
interpretations of the data.
Internal validity: The correctness of empirical research assessed in terms of criteria endogenous to
the research itself.
Nomothetic research: Research about the common characteristics of a group.
Proposition: A statement that is capable of being true or false.
Quasi-experiment: Experiment-like research in which at least one recognized control is missing
due to logistic or practical constraints.
Reductionism: Reductionism is the meta-theoretical assumption that incorporeal psychological
phenomena can be explained with terms used in the physical, biological, or computational
sciences.
Reliability: The extent to which the same test will produce comparable results.
Research: Research is the systematic collection of data to answer a well-defined question.
Scientism: An exaggerated faith in the propriety or efficacy of applying the methods used in
physical sciences to the social sciences.
Syllogism: An argument made up of two premises and a conclusion.
Theory corroboration: The attempt to assess whether or not a theory is tenable with empirical data
collected for that specific purpose.
Truth-value: Whether or not a proposition is true.
Validity: “Validity” is the technical term for the correctness of empirical research. As the research
may be correct in more than one way, psychologists are concerned with various types of
validity, namely, internal validity, external validity, ecological validity, and practical
validity.
Variable: A variable is something that may be identified in more than one way. For example, color
as a perceptual category may be represented by white, blue, green, and the like. Hence,
color is a variable, three of whose representations (called values or levels) are white, blue,
and green.
Bibliography
Boring E.G. (1954). The nature and history of experimental control. American Journal of Psychology 67, 573–589.
[This is one of the few attempts to show explicitly the technical meanings of “control” and their inductive foundation.
Much of the contemporary misunderstanding of psychological research could have been avoided had textbooks not
neglected Boring’s contribution.]
Campbell D.T. and Stanley J.C. (1966). Experimental and Quasi-Experimental Designs for Research, 84 pp. Chicago:
Rand McNally. [This is a non-technical introduction to the inductive basis of empirical research.]
Chow S.L. (1992). Research Methods in Psychology: A Primer, 320 pp. Calgary: Detselig. [This textbook gives an
overview of commonly used research methods to a greater depth than most introductory textbooks. It also presents a
more critical look at some commonly held belief about research methods such as the social psychological psychology of
the psychological experiment.]
Manicas P.T. and Secord P.F. (1983). Implications for psychology of the new philosophy of science. American
Psychologist 38, 399–413. [Apart providing a good alternative to positivistic behaviorism, they present a convincing
case as to why the logic of explanation is not modus ponens, but the juxtaposition of modus tollens (for the rejection of
an untenable explanation) and the indeterminate affirming of the consequent (for the tentative retention of the tenable
explanation, albeit tentative).]
Meehl P.E. (1967). Theory testing in psychology and in physics: A methodological paradox. Philosophy of Science 34,
103–115. [This is one of the first articles that remind psychologists not to conflate the statistical hypothesis with the
substantive hypothesis. By implication, readers will learn that making the statistical hypothesis is not corroborating the
substantive hypothesis.]
25
Popper K.R. (1968). Conjectures and Refutations: The Growth of Scientific Knowledge, 417 pp. New York: Harper &
Row. [Chapter 2 is a less technical introduction to Popper’s hypothetico-deductive view of science. Much of the
dissatisfaction with Popper’s argument is based on his critics’ neglect of Popper’s emphasis on “growth.”]
Turner M.B. (1968). Psychology and the Philosophy of Science, 240 pp. New York: Appleton-Century-Crofts. [Much of
the discussion in Section 5 ("Philosophical Issues") is based on Turner's book, particularly the distinction between
logical construct and hypothetical mechanism.]
Wilkinson, L. and Task Force on Statistical Inference (1999). Statistical methods in psychology journals: Guidelines
and explanations. American Psychologist 54, 594–604. [This report is an advocacy document, in which a set of research
agenda is set forth. Despite the Task Force’s disavowal, much of what the report says is about research methods in
general, and psychologists’ putative departure from good research practice in particular. The report sets in high relief
the importance of the philosophical and methodological issues raised in the present article.]
26