Epistemological and Methodological Bases of

Naturalistic Inquiry
Egon G, Guba
Yvonna S, Lincoln

It is important, at the outset, to recognize

Egon G. Guba is professor of education, Indiana Uni-
versity, Bloomington, IN 47405. Yvonna S. Lincoln is what naturalistic inquiry is and what it is
associate professor of education, University of Kansas, not. Naturalistic inquiry is a paradigm of
Lawrence, KS 66045. inquiry, that is, a pattern or model for how
inquiry may be conducted. It is frequently
asserted that its distinguishing features are
that it is carried out in a natural setting
(hence the term naturalistic), that it uses a
case study format, and that it relies heavily
on qualitative rather than quantitative
methods; however, none of these features is
naturalistic inquiry. All of these assertions
are essentially correct, but no one of them,
nor indeed all of them together, captures
How suitable is the rationalistic paradigm the full significance of the term paradigm.
for research focusing on human Paradigms differ from one another on mat-
behavior? Proposing that naturalistic ters much more fundamental than the locale
inquiry befler serves the social/ in which the inquiry is conducted, the for-
behavioral sciences, the authors define mat of the inquiry report, or the nature of
the differences between the two the methods used--namely, they differ on
paradigms and suggest criteria for the basic axioms on which t h e y rest.
ensuring the trustworthiness of naturalistic Paradigms are axiomatic systems charac-
inquiry. terized essentially by their differing sets of
assumptions about the p h e n o m e n a into
which they are designed to inquire.
There are many different paradigms of
inquiry. We are all intimately familiar with
most of them, and we use them on virtually
a daily basis. The system of jurisprudence
common in the Western World, for exam-
ple, is based on an adversarial paradigm, in
which opposing parties press their cases
with as much vigor as can be managed; it is
believed that a jury can sift out truth from
ECTJ, VOL. 30, NO, 4, PAGES233-252
falsehood when the opposing views have
ISSN 0148-5806
been heard. Religious faiths are based on
234 ECTJ WINTER 1982

theological paradigms, most of which hold aggregatability of data is one of the chief
that necessary truth is directly revealed; virtues of operating in the scientific mode;
study of revelation can yield the answer to each investigator "stands on the shoulders"
any question. Outcomes of many athletic of his or her predecessors to make the next
contests, critiques of works of art, and peer logical contribution. But where is this ag-
reviews of research proposals are based on a gregatability in sodal and behavioral sd-
judgmental paradigm, which assumes that ence? Where is the essential body of knowl-
competent persons (experts, critics) can, by edge, systematically and patiently built up
virtue of training and experience, recognize over decades of work? It is true that exten-
truth when they see it. sive literatures exist, but they are charac-
Those persons concerned with disciplined terized at least as much by conflicting find-
inquiry (in the sense that that term is de- ings as by reinfordng ones. Even the ap-
fined by Cronbach and Suppes, 1969) have proach known as "meta-analysis" has been
used almost exclusively what is commonly found inadequate by its critics, on such
called the scientific paradigm but which we grounds as failing to differentiate adequate
will here, for a variety of reasons, term the from inadequate studies or, more impor-
rationalistic paradigm. 1 A second paradigm tantly, focusing on "main effects" while ig-
that is also aimed at disciplined inquiry-- noring contingendes or important side ef-
one we term the naturalistic--is currently fects.
receiving a great deal of attention, and it is We may also cite the inability of inves-
that paradigm which this paper explicates. tigators to apply the scientific model in ac-
Whereas related paradigms, for example, cordance with the design principles on
the anthropological or ethnographic, have which it is based. Random sampling, for
been in use for some time, the naturalistic example, is a virtual impossibility in any
paradigm has only recently emerged as a real world situation, as is random assign-
serious contender in social/behavioral in- ment to treatments. Scientists often argue
quiry. that such failures are more apparent than
One may well ask why anyone would real, suggesting that if there were more so-
contemplate using a competing paradigm cial commitment to discovering "truth" and
when the rationalistic one has gained such more understanding of the need to conform
widespread legitimation and achieved such to design conditions to ascertain it, then
conspicuous successes. After all, one need application of the paradigm would indeed
only look about to see the tangible proof of be possible. Scientists are quick to excoriate
the utility of scientific inquiry; the current political figures, administrators, parents,
that powers the typewriter on which this and even the "subjects" themselves (to use
paper was drafted, for example, not to men- their pejorative term) for their unwilling-
tion more spectacular feats such as moon ness to cooperate, obviously making the
shots and bypass surgery, depends on the judgment that scientific values stand unas-
fruits of such research. How could one sailably above political, human, or ethical
doubt the efficacy of the sdentific mode for values. But clearly, political figures and de-
all inquiry? John Stuart Mill urged inves- cision makers must take account of the vari-
tigators in what might now be called the ety of factors in their milieu; individuals
social science areas to shift to the scientific have a right to privacy and to opt in or out of
mode as long ago as 1843; can there by any a study depending on how they view the
question, a century and a half later, that his risks to which they are to be exposed; ethics
advice was well founded? contravene the withholding of a potentially
It seems to us that a variety of evidence effective treatment simply for the reason of
can be cited in counterargument. First, we testing it experimentally. Finding a
believe that the j u d g m e n t that the paradigm that can tolerate real world condi-
rationalistic paradigm has enabled con- tions surely makes more sense than manip-
spicuous successes in social and behavioral ulating those conditions to meet the arbi-
inquiry is mistaken, however well founded trary design requirements of a paradigm.
it may be in the so-called hard or life sd- We may also note, in contrast to the claim
ences. For example, it is often argued that that the scientific model has achieved such
 NATURALISTICINQUIRY 235

conspicuous successes, the general lack of sciences, often continue to act as if the
impact on practice of research conducted in rationalist paradigm had validity, continu-
this mode. The failure to use evaluation in- ing to accept a position that is essentially
formation, for example, verges on a na- analytic, reductionist, empiricist, associ-
tional scandal. And to what practices in our afionist, reactivist, nomological, and monis-
schools can we point with confidence as tic. As we shall see, this posture is inconsis-
having originated in research data? Indeed, tent with the characteristics of much social/
virtually all school practice has emerged on behavioral inquiry.
the basis of experience. One possible expla- Finally, we suggest that the rationalistic
nation of this lack of impact is that users and paradigm, like all paradigms, rests on cer-
practitioners lack the insight and creativity tain fundamental axioms or assumptions,
to see how research results can be applied. and that the particular axioms of rationalism
An equally compelling argument is that the are but poorly fulfilled in social/behavioral
research results have so little meaning that inquiry. This fact can be appreciated from
the effort to apply them is wasted. the preceding few paragraphs; indeed, the
The realization that the rationalistic shortfalls of rationalistic inquiry (in the so-
model is difficult to apply and results in cial/behavioral arena) are so pronounced
information that is used infrequently is not that if naturalistic inquiry did not exist as an
new to the research community. Educa- alternative, it, or something like it, would
tional research is often dated from the 1894 have to be invented. It is our intention to
publication of Joseph Mayer Rice, who at- devote a major segment of this paper to a
tempted to show, by relating scores on discussion of these rationalistic axioms and
achievement tests (which he developed) to of their naturalistic counterparts, and to
amount of time spent studying (an early deal with the question of which set of
version of time on task?), that schools were axioms is better fulfilled in the phenom-
wasting time in teaching the three R's. enological field customarily designated as
Rice's first tests were found to be invalid "social/behavioral."
because of the lack of field controls (the As we shall demonstrate, the motivation
teacher taught to the tests, for example, or for considering naturalistic inquiry as an al-
provided cues to "help" students); his later ternative paradigm to rationalism is not
test (confined to spelling) provided infor- founded simply on the desire to avoid the
mation that a p p a r e n t l y s u p p o r t e d his shortfalls of that latter model. Naturalistic
hypothesis, but publication of results influ- inquiry has many characteristics to recom-
enced virtually no one to change methods of mend it on other grounds. For example, it
teaching spelling. It is astonishing that the offers a contextual relevance and richness
research community, faced with similar ex- unmatched by any other paradigm. It dis-
periences for nearly a century, persists in plays a sensitivity to process virtually ex-
finding fault with almost every aspect of the cluded in paradigms stressing control and
situation except the paradigm that guides experimentation. 2 It is driven by theory
their inquiry in its effort to explain these grounded in the data; the naturalist does
observations. not search for data that fit his or her theory
As the second point in our counterargu- but develops a theory to explain the data.
ment, we question the utility of the Finally, naturalistic approaches take full
rationalistic paradigm, as we shall describe advantage of the not inconsiderable power
it below, on the grounds that it reflects ear- of the human-as-instrument, providing a
lier rather than emergent epistemologies of more than adequate trade-off for the pre-
science. It is apparent that sophisticated sumably more "objective" approach that
m o d e r n - d a y scientists cannot accept characterizes rationalistic inquiry.
positivism; even a casual acquaintance with Even without these a priori claims for the
the field of particle physics provides ample advantages of naturalistic inquiry, it seems
evidence of its inadequacies, as, for exam- clear that the examination of alternative
ple, in the Heisenberg Uncertainty Principle paradigms has utility, since such examina-
(Tranel, 1981). Yet practitioners of scientific tions force out otherwise hidden assump-
inquiry, in the hard but especially in the soft tions and meanings. If it is true that the
236 ECTJ WINTER 1982

examined life is "better" than the unexam- With these four axioms, Euclid was able to
ined, it is surely the case that the examined derive the first 28 of the eventually much
paradigm is better than the unexamined. larger set of theorems, but the 29th proof he
This paper, then, has two major pur- attempted was intractable; Euclid had in-
poses: (a)to distinguish the rationalistic stead to assume it as a fifth axiom:
and naturalistic paradigms on five basic 5. If two lines are drawn which intersect
axiom differences, and to describe six post- a third in such a way that the sum of the
ures on which practitioners of these inner angles on one side is less than two
paradigms have traditionally differed; and right angles, then the two lines inevitably
(b) to suggest some methods for respond- must intersect each other on that side if
ing to four basic criteria for trustworthiness extended far enough.
(analogs to the traditional rationalistic Another way to state this axiom is: Given a
criteria of internal and external validity, re- line and a point not on that line, it is possible
liability, and objectivity) that might be used to construct only one line through the point
by naturalists to counter charges of lack of and parallel to the given line.
discipline (sloppiness). Now, c o m p a r e d with the first four
axioms, the fifth seems strained and inele-
BASIC AXIOMS THAT DISTINGUISH THE gant, and Euclid was sure that eventually he
NATURALISTICFROM THE RATIONALISTIC would be able to find a way of proving it in
INQUIRY PARADIGM
terms of the first four. But his hope was not
Axioms can be defined as the set of undem- to be realized within his lifetime, or indeed,
onstrated (and undemonstrable) propo- ever; two millennia of effort by mathemati-
sitions accepted by convention (even if only cians have failed to provide a proof.
intuitively) or established by practice as the Early efforts to prove this axiomJtheorem
basic building blocks of some conceptual or were of what mathematicians would call the
theoretical structure or system. Before direct variety, but no direct proofs could be
examining the axioms that underlie the two found. Later mathematicians fell back on
paradigms of interest here, it may be useful indirect proofs, one variant of which is to
to clarify the nature of axiomatic systems. assume the direct opposite of what one
Probably the best k n o w n and most wishes to prove and then to show that the
widely experienced system of axioms is that opposite assumption leads to absurd con-
undergirding Euclidean geometry, to which clusions (theorems). It was exactly this ap-
virtually e v e r y o n e is exposed in high proach, however, that led to so-called
school. Euclid had set himself the task of non-Euclidean geometries; not only were
formalizing everything k n o w n about the consequences of such non-Euclidean
geometry at his time; essentially that meant assumptions not absurd, they were in fact
systematizing the many rules of thumb of great utility for many purposes. One such
used by land surveyors, who, although geometry is called Lobachevskian, after its
using them, could not provide any proof of developer, Nicholas Lobachevsky. This
their validity other than experience. It was form of non-Euclidean geometry begins
Euclid's powerful and ingenious insight with the axiom, "Given a line and a point
that perhaps all of these rules could be not on that line, it is possible to draw a
"proved" by showing them to be logical bundle of lines through the point, all of
derivatives from some simple and basic set which are parallel to the given line." Now
of "self-evident truths." Euclid began with this axiom seems absurd and certainly flies
four such axioms (Hofstadter, 1979): in the face of all human experience, yet it
1. A straight line segment can be drawn yields results of great interest, for example,
joining any two points. to astronomers. One of the theorems
2. Any straight line segment can be ex- "provable" from the Euclidean fifth axiom
tended indefinitely in a straight line. is that the sum of angles in a triangle is 180~
3. Given any straight line segment, a cir- but the sum of angles in Lobachevskian
cle can be drawn having the segment as triangles is not always 180~ but approaches
radius and one end point as center. 180~ as triangles become "small." It seems
4. All right angles are congruent. clear that earth-sized triangles are all small,
 NATURALISTICINQUIRY 237

TABLE 1
Axiomatic Differences Between the Rationalistic and Naturalistic Paradigms

Paradigm
Subject of Axiom Rationalistic Naturalistic
Reality Single, tangible, convergent, Multiple, intangible, divergent,
fragmentable holistic

Inquirer/respondent relationship Independent Interrelated

Nature of truth statements Context-free generalizations-- Context-bound working hypo-

nomothetic statements--focus theses--idiographic state-
on similarities ments--focus on differences

Attribution/explanation of action "Real" causes; temporally prec- Attributional shapers; interactive

edent or simultaneous; manipul- (feedforward and feedback);
able; probabilistic nonmanipulable, plausible

Relation of values to inquiry Value-free Value-bound

Note. In certain of our previous writing (Guba, 1978,1981; Guba & Lincoln, 1981) we focused on only the first three of
these five axioms. However, the latter two now seem to us as equally, if not more, important.

since no such triangle has ever yielded a Hence, the axioms to be described in this
sum of angles less than 180~. But astronomi- section s h o u l d not be j u d g e d on the
cally sized triangles are very much larger, grounds of their self-evident truth, their
and astronomers find that Lobachevskian common-sense qualities, or their familiarity
geometry provides a better "fit" to the to the inquirer, but in terms of their fit to the
phenomena they investigate than does Eu- phenomena into which one proposes to in-
clidean. quire. When the rationalistic axioms fit, the
From this digression we may deduce sev- rationalistic p a r a d i g m should be used;
eral crucial points: when the naturalistic axioms fit, the natu-
9 Axioms are arbitrary and may be as- ralistic paradigm should be used.
sumed for any reason, even if only for "the Five axioms differentiate the rationalistic
sake of the game." and naturalistic paradigms; these five
9 Axioms are not self-evidently true, and axioms are summarized in Table 1. Im-
they need not even appear so; indeed, some mediately following is a more formal state-
axioms may appear very bizarre at first. m e n t of the five axioms in both their
9 Different axiom systems will have dif- rationalistic and naturalistic versions. We
ferent utilities, depending on the phenom- attend to the question of which set provides
enon to which they are applied. These a better fit to social/behavioral phenomena
utilities are not determined by the nature of in a later section.
the axiom system itself but by the nature of
the interaction between these axioms and
the characteristics of the area in which they Axiom 1. The nature of reality
are proposed to be applied. Thus, Euclidean Rationalistic version: There is a single,
geometry is fine for terrestrial spaces, but tangible reality fragmentable into indepen-
Lobachevskian geometry is better for inter- dent variables and processes, any of which
stellar spaces. can be studied independently of the others;
9 A decision about which of several al- inquiry can converge on this reality until,
ternative axiom systems to use in a given finally, it can be predicted and controlled.
case is made by testing the "fit" between Naturalistic version: There are multiple,
each system and the case, a process analo- intangible realities which can be studied
gous to (although not nearly as well under- only holistically (to dissociate the wholes is
stood as) testing data for fit to assumptions to alter them radically); inquiry into these
before deciding on which statistic to use in multiple realities will inevitably diverge
analyzing them. (each inquiry raises more questions than it
238 ECTJ WINTER 1982

answers) so that prediction and control are method for establishing cause-effect rela-
unlikely outcomes, although some level of tionships is the experiment that demon-
understanding (verstehen) can be achieved. strates via manipulation that introducing
the cause(s) produces the effect.
Axiom 2. The inquirer-object relationship
Naturalistic version: An action may be ex-
Rationalistic version: The inquirer is able
plainable in terms of multiple interacting fac-
to maintain a discrete and inviolable dis-
tors, events, and processes that shape it and
tance between himself/herself and the ob-
are part of it; inquirers can, at best, establish
ject of inquiry; when the object is a human
plausible inferences about the patterns and
being, special methodological safeguards
webs of such shaping in any given case. The
may need to be taken to guard against reac-
best method for assessing these patterns
tivity, that is, a reaction of the object to the
and webs is the field study that deals with
conditions of the inquiry that will influence
them holistically and in their natural con-
the outcome in undesirable ways.
texts.
Naturalistic version: The inquirer and the
object interact to influence one another; es-
Axiom 5. The role of values in inquiry
pecially is this mutual interaction present
Rationalistic version: Inquiry is value-free
when the object of inquiry is another
and can be guaranteed to be so by virtue of
human being (respondent). Special safe-
the objective methodology employed.
guards must be taken against both kinds of
These methods guarantee inquirer neu-
reactivity. 3
trality and inquiry rigor and produce data
Axiom 3. The nature of truth statements that "speak for themselves."
Rationalistic version: The aim of inquiry is Naturalistic version: Inquiry is always val-
to develop a nomothetic body of knowl- ue-bound. It is value-bound in at least four
edge; this knowledge is best encapsulated ways, captured in the corollaries that fol-
in generalizations (truth statements of en- low.
during value that are context-free). The Corollary 1: Inquiries are influenced by
stuff of which generalizations are made is inquirer values, especially as those values
similarities among units; differences are set are expressed in the choice of a problem and
aside as intrinsically uninteresting. the framing, bounding, and focusing of that
Naturalistic version: T h e aim of inquiry is problem.
to develop an idiographic body of knowl- Corollary 2: Inquiry is value-influenced
edge. This knowledge is best encapsulated by the paradigm selected which guides the
in a series of "working hypotheses" that investigation into the problem.
describe the individual case. Generaliza- Corollary 3: Inquiry is value-influenced
tions are impossible since phenomena are by the choice of substantive theory and
neither time- nor context-free (although methods used to guide the collection and
some transferability of these hypotheses analysis of data relevant to the problem
may be possible from situation to situation, selected and in the interpretation of find-
depending on the degree of temporal and ings.
contextual similarity--an empirical matter); Corollary 4: Inquiry is influenced by the
differences are as inherently interesting as values inherent in the context. Those values
(and at times more interesting than) are not those that characterize individuals,
similarities. but those that specifically characterize so-
ciobehavioral, human, organizational
Axiom 4. Attributionlexplanation of action phenomena.
Rationalistic version: Every action can be Corollary 5: With respect to corollaries 1,
explained as the result (effect) of a real cause 2, and 3, inquiry may be said to be either
or causes that precede the effect temporally resonant (reinforcing or congruent) or dis-
(or are at least simultaneous with it); inquir- sonant (conflicting) with the nature of the
ers can, using suitable methods, establish problem to be studied. For instance, an in-
cause-effect relationships unequivocally in quirer could determine a problem to be
functional form, that is, y = (axl + bx2 9 9 9 studied, choose a paradigm, and/or choose
rxn), or at least probabilistically. The best methods that have the possibility of being
 NATURALISTIC INQUIRY 239

either value-resonant or value-dissonant ality but exist only in the minds of people.
with the context into which he or she will As Filstead (1979) suggests, "There are mul-
take the inquiry. Problem, p a r a d i g m , tiple realities . . . . Individuals are concep-
method(s), and context must be congruent tualized as active agents in constructing and
(value-resonant) with each other to produce making sense of the realities they en-
meaningful findings. counter" (p. 36). There are as many con-
The decision about which paradigm to structions as there are people to make them.
use in a given case depends, we again as- These constructions are also not equiva-
sert, on an assessment of the area to be lent to perceptions which, although cer-
studied to determine the degree of "fit" be- tainly influenced by factors inhering in the
t w e e n the axioms of each c o n s i d e r e d perceiver, are mainly determined, as the
paradigm and the area. If we limit ourselves Dictionary of Psychological and Psychoanalyti-
to a consideration solely of the area com- cal Terms phrases it, by the "excitation of
monly designated as social/behavioral in- sensory receptors." We are not belaboring
quiry, 4 we make the following observations here the well-known fable of the blind men
about fit. and the elephant. If that fable were to
provide a useful metaphor, it would only do
Axiom 1. The Nature of Reality. In the hard or so were there no elephant. We mean to
life sciences, there can be little doubt that suggest precisely that there is no tangible
there exists a tangible reality which is the reality that can be touched as the blind men
focus of inquiry: actual events, objects, and touched the elephant. There we dealt with
processes found in nature that can be ob- their perceptions of the elephant; here we
served and often measured. The utility of deal with constructions that are developed
breaking this physical world into variables from whole cloth in the mind of the con-
is well demonstrated by such terms from structor.
physics as time, mass, velocity, accelera- Since these constructions reside wholly
tion, distance, charge, and the like. Such in the minds of people, where they are sub-
variables can be studied independently and stantially inaccessible, they must be dealt
related to one another in expressions of the with in a holistic fashion. They cannot be
form y = f(axl + bx2 + . . . rxn). Of course, divided into parts or variables, however
even in the hard sciences, recent work runs appealing that prospect may be; nor can
afoul of these assumptions; for example, they be understood as the mere sum of
many of the particles studied by physi- parts. An understanding of the quality of
cists (e.g., mesons) cannot exist (to do so shirtness, by way of analogy, requires more
would violate the law of conservation of than an analysis of cut-up sleeves, cuffs,
mass/energy), yet they do. Physicists get collars, and so forth.
around this difficulty by referring to them as Finally, since the realities are multiple (as
"virtual" particles (Zukav, 1980). Neverthe- many constructions as there are people), it
less, for all practical purposes, and certainly is futile to expect inquiry to converge. One
at macro levels, the rationalistic axioms cannot converge on a " c o m m o n " or "typi-
serve the hard and life sciences well. cal" reality since each is idiosyncratic. The
In the social/behavioral sciences, how- more individuals one explores, the more
ever, the class of phenomena typically ad- realities one encounters; inquiry diverges as
dressed in inquiry has no reality in the phys- a result. Knowledge cannot be represented
ical sense. The phenomena we deal with as a " m a p " of territory to be explored and
cannot be touched, seen, tasted, smelled, or finally to be fully understood; knowledge is
heard. That is not to say that tangible ob- rather like a sphere in space, which repre-
jects, events, and processes do not enter sents the unknown. As the sphere is en-
into human behavior, for example, to shape larged (i.e., more knowledge is gained), one
it. However, it is not these tangibles that we simultaneously comes in contact with new
care about, but the meaning and interpreta- unknowns that were not appreciated be-
tion people ascribe to or make of them, for it fore. Every inquiry finally raises more ques-
is these constructions that mediate their be- tions than it answers.
havior. These constructions do not have re- Axiom 2. The Inquirer-Object Relationship. In
240 ECTJ WINTER 1982

the hard and life sciences, it is not unrea- spondent's return. And so on.
sonable to posit the ability of the inquirer to Finally, far from deploring inquirer-re-
maintain a discrete distance from the s p o n d e n t interactivity, the naturalist
phenomenon under study. Balls rolling exploits it. If interactivity could be elimi-
down inclined planes, chemicals interacting nated by some magical process, the natu-
in a test tube, or cells subdividing under a ralist would not think the tradeoff worth-
microscope are unlikely to be influenced by while, because it is precisely the presence of
the fact that someone is watching; nor is the interactivity that makes it possible for the
watcher likely to be influenced (in any way inquirer to be a "smart" instrument, honing
adverse to the investigatory outcome) by in on relevant facts and ideas by virtue of his
what he or she observes. Of course, when or her sensitivity, responsiveness, and
the phenomena being studied are at the adaptability. This will be discussed more
level of subatomic particle, the so-called later.
Heisenberg Uncertainty Effect (Tranel, 1981)
comes into play; the act of determining one Axiom 3. The Nature of Truth Statements. The
characteristic of such a particle--for exam- development of generalizations is said by
ple, the mass--makes it impossible to de- many to be the ultimate purpose and aim of
termine other characteristics such as the inquiry. If generalizations cannot be de-
velocity. So even in a realm indisputably rived, one can deal only with the unique
well served by the rationalistic paradigm, event. Why would anyone want to invest
grave doubts emerge about the indepen- time and effort in a study that can yield
dence of observer and observed. nothing more than the single occurrence
In the social/behavioral sciences, a good has to offer? Context-flee statements of en-
deal of attention has been paid to reactivity during truth value clearly are highly prized.
of subjects (Campbell & Stanley, 1963; The question that confronts us is whether
Webb et al., 1966). It is commonly recog- they are achievable.
nized that objects of inquiry, when they are In the hard and life sciences that question
human, may react to inquirers or to their must be a n s w e r e d with a r e s o u n d i n g
inquiry methods, s Less appreciated is tile "Yes." Statements like F = ma and e = mc 2
fact that the inquirer is also subject to in- are derivable in physics, for example, and
teraction. Just as the inquirer may shape the they hold "true" whether tested in the 18th
respondent's behavior, so may the respon- or 20th centuries, on Earth, Mars, the
dent shape the inquirer's behavior. Anyone moon, or anywhere else in the universe.
who has done research involving humans is Such statements form the cornerstones of
well aware that one cannot abandon one's most disciplines; indeed, the phrase
own humanness in the interest of "objec- "nomothetic science" implies exactly the
tive" inquiry; it is both impossible and ethi- development of law-like generalizations
cally undesirable to do so. which can be depended on to provide bases
It should also not be supposed that the for prediction and control.
interpolation of a layer of "objective" in- Even in the hard sciences, however, there
strumentation between the inquirer and the is a real question of whether generalizations
respondent(s) is sufficient to overcome or can be made that will be true "forever."
offset this interaction. Images of what the Cronbach (1975) uses an interesting
respondent may be like or how he or she metaphor, that of the decay of radioactive
might respond (very often these images are materials, to make the counterpoint.
formalized in the theory that guides the in- Generalizations, he asserts, like radioactive
quiry-itself a construction) guide the in- substances, decay and have half-lives. He
quirer in devising instruments. Images of gives numerous examples from both the
what the inquirer wants, or what he or she hard and the social/behavioral sciences to
will do with the responses, guide the re- make his point--for example, the failure of
spondent in dealing with the instruments. DDT to control pests as genetic transforma-
Images of what the respondent meant or tions make them resistant to the insecticide,
intended guide the inquirer in coding, in- the shifting of stars in their courses so as to
terpreting, and even in accepting the re- render star maps obsolete, Ghiselli's sug-
 NATURALISTIC INQUIRY 241

gestion that the superiority of distributed ing "to probe causal hypotheses" (p. ix), see
practice over massed practice may not re- causal connections as "real but imperfectly
main valid from one generation to another, perceived" (p. ix), and address their book to
and Bronfenbrenner's conclusion that class those " w h o have already decided that they
differences in parenting observed in the want a causal question answered" (p. 2).
1950s were just the reverse of those ob- For them the question is not whether to
served in 1930. One is reminded of the long- entertain a concept of causality but which
ing of the king in "Anna and the King of concept to accept.
Siam" for a return to his childhood when The meaning to be imputed to the term
"what was so was so and what was not was causality has been discussed for centuries,
not." Things absolutely known to be true at despite which, as Cook and Campbell (1979)
some point in time turn out not to be true at note, "the epistemology of c a u s a t i o n . . , is
some other time, or in some other cultural at present in a productive state of near
or social context. chaos" (p. 10). Causality originally was con-
What seems to be the case even in the ceived in a common-sense way in "if-then"
hard sciences is undoubtedly much more terms, probably because of the tendency of
the case in the social/behavioral sciences. It early scientists to view the world as one
is doubtful whether generalizations can be huge machine whose linkages produced
made about h u m a n behavior with impun- and drove all observable phenomena. In the
ity. Time is an enormously important factor, early 18th century David Hume noted that
and who can offer an example of human causality was never directly observed but
behavior that is context free? merely imputed by the observer when two
N o w none of this line of argument should events were contiguous and temporally fol-
be interpreted to mean that there can never lowed one another. H u m e e s p o u s e d a
be any transfer (to resurrect an old familiar "regularity" or "constant conjunction"
term) from one situation to another. What theory of causality that denied the need for
we mean to say is that statements cannot be the concept of causality at all. Later, an "es-
made about human phenomena that are sentialist" view emerged based on the idea
likely to be true for even substantial num- of necessary and sufficient conditions; es-
bers of years (not to mention forever) or for sentialists sought functional laws express-
any substantial number of contexts (not to ing inevitable cause-effect relationships
mention any and all contexts). Condition- (Weir, 1980). An activity theory of causation
als, contingencies, and disjunctions must placing heavy emphasis on manipulation as
all be taken into account (Wiles, 1981). the test for inferring cause-effect relation-
Moreover, differences in times or contexts ships n o w has wide currency, lending
are as important to know about in making legitimation to the notion that the best test
the j u d g m e n t of transferability as are for cause-effect relationships is the experi-
similarities; it is as important to know the ment (Cook & Campbell, 1979). Cook and
ways in which fit does not occur as to know Campbell opt for the "critical-realist" posi-
the ways in which fit does occur. The natu- tion:
ralist, then, is concerned first with develop- The perspective is realist because it assumes that
ing an adequate idiographic statement causal relationships exist outside of the human
about the situation he or she is studying, mind, and it is critical-realist because it assumes
accompanied by sufficient "thick descrip- that these valid causal relationships cannot be
tion" to make judgments about transferabil- perceived with total accuracy by our imperfect
ity possible, should anyone care to ask that sensory and intellective capacities. (pp. 28-29)
question. [emphasis added]
Formulations such as any of those above
Axiom 4. Attribution/Explanation of Action. have meaning (to some degree) within the
The search for causality is the mainspring rationalist paradigm insofar as it is applied
that drives conventional research. Even to the hard and life sciences. There seems to
such authors as Cook and Campbell (1979) be little question about the propriety of seek-
who recognize that causality is a slippery ing cause-effect relationships when one is
concept nevertheless define designs as serv- talking about gas laws, electric circuits, or
242 ECTJ WINTER 1982

an impact that mashes the fender of an au- pattern of neural connections in the brain.
tomobile. But these ideas are highly suspect All neurons are interconnected and can
when applied to the arena of social/behav- stimulate one another; an arm moves, a
ioral inquiry. The realities that we are deal- memory emerges, a creative insight is expe-
ing with are constructed and exist only in rienced, but how could one say what the
the minds of people; if the realities are con- cause-effect connections were that pro-
structed, why not the attributions or expla- duced those actions? They seem not to be
nations of causality? And if that is reason- isolatable to particular neurons or chains of
able, emergent attributional and/or seman- neurons, and virtually any set of neurons
tic theories of causation (if that is now the can perform the functions "normally" pro-
proper term) are more likely to be meaning- duced by other sets. Which neuron stimu-
ful than any of the formulations that have lated which in producing the action? Ac-
developed in relation to the other inquiry tions emerge, and there is no doubt that
areas. In these views, causality is not merely they emerge in conjunction with activity in
empirical or contingent, but depends heav- the brain, but the concept of causality seems
ily on meaning. Such questions as, Is the to provide little insight in understanding
treatment applied via a particular instruc- what is happening. Simultaneous shaping
tional program effective in increasing stu- rather than temporally contingent causa-
dent learning?, imply a cause-effect relation tion seems a more useful concept under
between "treatment" and "student learn- such circumstances.
ing," but the nature of that relationship
surely depends on what is meant by treat-
ment and student learning and what criteria Axiom 5. The Role of Values in Inquiry. The
of effectiveness are taken to be. In other customary presupposition of rationalists is
words, causality is a construction less trace- that inquiry is value-free, that is, that the
able by empirical linkages than by plausible outcomes of the inquiry are guaranteed by
semanticdattributional linkages. The con- the methodology employed by rationalists
cepts of constructed reality and constructed to be purely empirical. The data, it is often
causality are congenial to and supportive of said, "speak for themselves," that is, they
one another. transcend the values of both inquirers and
Thus, the naturalist argues, there can be respondents. Naturalists, on the other
no certain way of determining a cause-effect hand, presuppose that inquiry is inevitably
relationship; indeed, the very concept of grounded in the value systems that charac-
causality seems to have outlived its useful- terize the inquirer, the respondent, the
ness. Positivists such as Hume believed the paradigm chosen, the methods selected,
concept of causation to be unnecessary; and the social and conceptual contexts.
naturalists believe it to be archaic. Instead, Values cannot be set aside, methodologi-
the naturalist prefers to think of multiple cally controlled, or eliminated. It is more
factors and conditions, all of which interact, reasonable, they assert, to acknowledge
with feedback and feedforward, to shape and take account of values, insofar as one
one another (note that causes and effects are can, than to delude oneself about their im-
inextricably intertwined). Action can be portance or to hope that methodological
understood not as having been caused but hedges (guards against threats to internal
as having emerged from the constant inter- validity, for example) will compensate for
play of its shapers, all of which are them- their intrusion.
selves part of the action, indistinguishable Values, naturalists insist, may enter into
from it, and shaping and being shaped si- and influence the course of inquiry in five
multaneously. Whereas rationalists seem to ways, all of which are by definition ex-
have given up certainty in specifying func- cluded in the rationalists' construction.
tional/causal relationship and have fallen 1. Values influence decisions about what
back on probabilistic statements, the natu- to study, how to study it, and what in-
ralist is satisfied to tease out plausible con- terpretations to make of the resulting data.
nections between phenomena. The evidence for such influences is over-
A useful metaphor for the naturalist is the whelming (Bahm, 1971; H o m a n s , 1978;
 NATURALISTICINQUIRY 243

Kelman, 1969; Krathwohl, 1980; Scriven, (increasing the odds for falsification) (Per-
1971), and most rationalists are willing to row, 1981).
concede at least this point. Both rationalists 4. Inquiry is also influenced by the mul-
and naturalists could agree, too, that values tiple value and belief systems that inhere in
also enter into decisions on the part of re- the context into which the inquiry is taken.
spondents about such matters as what in- Those values are comprised of the sets of
terpretations to make of requests for infor- values that belong to individuals and those
mation and what responses to give (for that specifically belong to social/behavioral,
example, whether open or guarded). human, and organizational phenomena.
2. Inquiry is also influenced by the Thus, the values held by an inquirer may be
paradigm selected to guide the investiga- reinforcing to the sets of values he or she
tion into the problem. The rationalist, for encounters in a context or they may be con-
instance, who believes that reality is singu- flicting. Of course such incongruences (or
lar and convergent, will impose that con- conflicts) may occur only when the objects
struction on the findings, even when he or of inquiry are human beings. Inquirers in
she hears respondents in a context saying the hard and life sciences virtually never
again and again that their constructions of experience such resonances or dissonances
the problem, or their lives, are at variance because the objects of their inquiries (e.g.,
with those of other respondents. Thus the atoms or genes) patently do not hold
rationalist proceeds much as does a court of values, but social/behavioral inquirers con-
law, "constructing" and reconstructing into front this phenomenon constantly. It is not
singular reality that which represents the surprising, for example, that many subjects
"truth" to him or her. of research characterize inquiry as de-
3. Inquiry is value-bound, too, by the humanizing or that ethical principles are
choice of substantive theory and the frequently found to contravene what scien-
methods chosen to collect and analyze data tifically oriented researchers would like to
relevant to the problem selected, and in the do.
interpretation of findings. The values un- 5. Finally, inquiry can be characterized as
dergirding the substantive theory selected being either value-resonant (reinforcing or
to guide the inquiry may be resonant or congruent) or value-dissonant (conflicting)
dissonant with the values undergirding the with the nature of the problem to be
methodological (paradigmatic) theory. Both studied. So, for instance, an inquirer could
the substantive theory and the methodolog- bound a problem to be studied, choose the
ical paradigm are constructions (in the paradigm within which to operate, choose
sense that we have used that term) and methods and substantive theory to guide
therefore have roots in assumptions and the inquiry, and still have to determine
value position. If those assumptions and whether the inquiry is value-resonant or
values happen to be resonant, that is, con- value-dissonant with the context of the in-
genial to and reinforcing of one another, quiry. When making this decision, prob-
well and good; if they happen to be disson- lem, paradigm, method(s), and context
ant, the research findings may be quite in- must exhibit internal coherence, value fit,
valid. and congruence (value resonance) for the
As an example, consider the study of or- inquiry to be deemed appropriate and fit-
ganizations. If organizations are conceived ting and to produce meaningful findings.
as bureaucracies (in the Weberian sense), The naturalist admits the role that values
conventional methodological approaches play in shaping an inquiry and appreciates
serve quite well, since they, like bureau- the possibility of difficulties arising if there
cratic theory, are very systems-oriented. is value dissonance either between inquirer
But if organizations are defined in emergent and subject or context or between theories
terms such as "loosely c o u p l e d , " of substance and method. Although natu-
methodologies based on systemic formula- ralists cannot eliminate value effects (any
tions are likely to be dissonant, resulting in more than can rationalists), they endeavor
curious findings and probably loading the to set up safeguards, to expose and expli-
dice against the "loosely coupled" theory cate the values whenever possible, and to
244 ECTJ WINTER 1982

test insofar as possible for value resonance. (over and above the fact that they will han-
In this latter regard we may note that the dle different types of data): the quantitative
naturalist's propensity for grounding the methods have greater precision and are
inquiry (see next section) provides a virtual mathematically manipulable, whereas qual-
guarantee of value resonance, since the sub- itative methods are richer and can deal with
jects' constructions and the substantive phenomena not easily translatable into
theory are both extracted from the data numbers (or cases in which the isomorph-
rather than laid on them.6 ism implied by the translation is suspect).
For the naturalist, the propensity toward
SOME CHARACTERISTICPOSTURES the use of qualitative methods is less rein-
Whereas the axioms represent basic distinc- forced by these advantages, however, than
tions in premises between the rationalistic by the fact that qualitative methods are the
and naturalistic paradigms, certain pos- methods normally preferred by humans
tures typically assumed by practitioners fol- using themselves as prime data collection
lowing these two orientations also provide instruments. Techniques such as interview,
important insights into the differences be- observation, use of nonverbal cues and un-
tween them. These postures are not com- obtrusive measures, and documentary and
pelled by the axioms, in the sense that they records analysis seem more appropriate
are necessary, logical derivatives (like the and (if the reader will forgive what is not
theorems of a geometry), yet they are con- intended as a pun) natural in that case.
genial or reinforcing to the practice of the
paradigms and probably would be insisted Source of Theory. Rationalists prefer a priori
on by each paradigm's followers. theory; indeed, they are likely to insist that
Six of the most common postures are de- inquiry without a priori theory is impossi-
scribed below. It should be noted that, un- ble. Theories always exist, they say, even if
like the case of axioms, where "either-or" only at the implicit level. It is better to make
decisions must be made, postures can often them explicit than to be uncertain about
be compromised. Yet compromises are in- what is guiding one's inquiry. Naturalists
frequently found. The reason for this ap- suggest that it is not theory but the inquiry
parent intransigence cannot be laid to the problem that guides and bounds an in-
obduracy of the proponents, however; quiry. A priori theory, they are likely to
rather it stems from the fact that the collec- assert, constrains the inquiry and intro-
tivity of postures support and reinforce one duces biases (believing is seeing). In all
another in extremely synergistic ways. Each events, theory is more powerful when it
is, in a sense, a raison d'etre for the others, arises from the data rather than being im-
and to compromise on any of them is to posed on them. It is better to find a theory to
considerably weaken the collective power explain the facts than to look for facts that
of all. accord with a theory. Again, there is some-
thing to be said for each point of view.
Preferred Methods. We have already noted Surely rationalists would not wish to devise
that the rationalistic and naturalistic theory that was never shown to have any
paradigms are often treated as though the relation to facts, nor would naturalists insist
major differentiating characteristic is their that each inquiry had to establish its own
relative preference for quantitative or theory de novo. Yet naturalists, using
qualitative methods. It is likely that, among themselves as instruments, building on
the six postures that will be briefly de- their tacit as well as propositional knowl-
scribed here, the quantitative-qualitative edge and unrolling the inquiry design as the
distinction is the one that can be most easily study proceeds, would find a priori theory
and sensibly compromised. Cook and uncongenial, preferring to develop the
Reichardt (1979) have referred to the distinc- theory as the collection of facts grew and
tion as "unhelpful," and have called for insights into their possible meanings ma-
more widespread utilization of both types tured.
of methods, a call with which we can in the
main agree. Each approach has advantages Knowledge Types Used. Rationalists confine
 NATURALISTICINQUIRY 245

the types of knowledge admissible in any perfect an instrument is for some use, the
inquiry to propositional knowledge less adaptable it is to others. The human
(Polanyi, 1966), that is, knowledge that can instrument, although admittedly imperfect,
be cast into language forms (i.e., sen- is nevertheless exquisitely adaptable. For
tences). In view of their insistence on a naturalists, with their propensity for
priori theory and their interest in shaping grounded theory and emerging design, the
inquiry preordinately around certain ques- human instrument is the ideal choice.
tions and hypotheses derived from the
theory, such a tendency is not surprising. Design. The rationalist insists on a preordi-
The naturalist, intent upon the use of the nate design; indeed, it is sometimes as-
human as the prime data collection instru- serted that a good design specifies in
ment and wishing to utilize the capabilities "dummy" form the very tables that will ul-
of that instrument to the fullest, also admits timately be found in the report, so well exp-
and builds upon tacit knowledge--intui- licated can the hypotheses and the proce-
tions, apprehensions, "vibes"--which, al- dures be. Naturalists, entering the field
though not expressible at any given mo- largely without a priori theory or hypothe-
ment, nevertheless occur to inquirers by vir- ses, literally are unable to specify a design
tue of their training and, especially, their (except in the broadest process sense) in
experience. 7 Of course naturalists seek to advance. Instead, they anticipate that the
recast their tacit knowledge into propo- design will emerge as the inquiry proceeds,
sitional form as soon as possible, since, with each day's work being heavily depen-
without so doing, they cannot communi- dent on what has gone before. Given their
cate with others, and probably not even other postures, naturalists have no choice
with themselves, about their findings. Yet but to opt for an emergent (rolling, cascad-
to confine the inquiry itself only to those ing, unfolding) design. Of course there is no
things that can be stated propositionally be- reason why naturalists should not be as
fore the fact is unduly and insensibly limit- specific as they can, without constraining
ing from the naturalist's viewpoint, since it their options.
eliminates to a large extent the predominant
characteristic warranting the use of the Setting. Finally, the rationalist prefers to
human-as-instrument. 8 conduct studies under laboratory (i.e., con-
trived, controlled, manipulable) conditions
Instruments. The rationalist prefers nonhu- to exclude from the inquiry any influences
man devices as data collection instruments other than those at which the inquiry is
for reasons such as they appear to be more aimed (that is, to exclude all confounding
cost efficient, they have a patina of objectiv- variables). The naturalist, on the other
ity, and they produce information that can hand, prefers natural settings, arguing that
be systematically aggregated. The naturalist only in such settings can the naturalist ar-
prefers humans-as-instruments for reasons rive at reasonable formulations and in-
such as their greater insightfulness, their terpretations. If theory is to be properly
flexibility, their responsiveness, the holis- grounded, the naturalist must observe the
tic emphasis they can provide, their ability facts as they normally occur, not as they are
to utilize tacit knowledge, and their ability contrived in an artificial context; without
to process and ascribe meaning to data si- meeting those conditions, the resulting
multaneously with their acquisition. Just as theory would be suspect indeed. In all
a "smart" bomb need not be dropped accu- events the findings would be transferable
rately on target to find its way unerringly to only to another contrived, artificial context.
it, just so the "smart" human instrument
need not begin with a precise problem We hope it is now clear why we asserted,
statement, theory, hypothesis, or method at the beginning of this section, that the
to find its way unerringly to what is most postures described above constitute a
salient in a situation. As Hofstadter (1979) synergistic set. Compromises are of course
pointed out, there is a precise tradeoff be- possible on each posture: one can use both
tween perfection and adaptability; the more quantitative and qualitative methods; one
246 ECTJ WINTER 1982

can use theory grounded in one study in a confidence in the "truth" of the findings of a
priori fashion for another study; one can particular inquiry for the respondents with
build on tacit knowledge but translate it into which and the context in which the inquiry
propositional knowledge as quickly as pos- was carried out?
sible; one can use both human and nonhu- 2. Applicability. How can one determine
man instruments depending on the pur- the degree to which the findings of a par-
pose of the inquiry and where one stands ticular inquiry may have applicability in
with respect to grounding and design other contexts or with other respondents?
emergence; one can keep design options 3. Consistency. How can one determine
open but make design decisions as ex- whether the findings of an inquiry would be
peditiously and completely as circum- consistently repeated if the inquiry were re-
stances allow; one can use both natural and plicated with the same (or similar) respon-
contrived settings depending upon the dents in the same (or a similar) context?
questions to be asked. But each posture 4. Neutrality. How can one establish the
supports the others; we could not argue for degree to which the findings of an inquiry
the naturalist's preferences on any one are a function solely of respondents and of
posture without invoking preferences on the conditions of the inquiry and not of the
other postures as well. It is difficult to imag- biases, motivations, interests, perspectives,
ine a naturalist at work who could be con- and so on, of the inquirer?
tent with a "mix-and-match" strategy, The terms typically utilized within the
however desirable that might be from the rationalistic paradigm in relation to the four
point of view of achieving a rapproche- questions are, respectively, internal validity,
ment. external validity, reliability, and objectivity.
Guba (1981) and Guba and Lincoln (1981)
THE TRUSTWORTHINESSOF propose four analogous terms within the
NATURALISTICINQUIRY naturalistic paradigm to supplant these
Naturalistic inquiry is, as we have tried to rationalistic terms: credibility, transferability,
demonstrate, very different in form and in- dependability, and confirmability.
tent from rationalistic inquiry. At present, The "translation" of the conventional
after some two centuries of experience with terms into these four naturalistic terms re-
rationalistic inquiry, several criteria of im- quires some justification. As provided by
portance have been identified for judging Guba (1981) and Guba and Lincoln (1981),
the trustworthiness of findings that emerge the justification is as follows:
from it. It is not unreasonable to ask 1. Credibility. Internal validity is best
whether naturalistic inquiry can also meet demonstrated through an isomorphism or
those criteria or, in the event that the criteria verisimilitude between the data of an in-
are deemed inappropriate, meet some new quiry and the phenomena those data repre-
criteria that are more appropriate and of sent. Although such isomorphism cannot
approximately equal power in differentiat- be directly represented in either paradigm,
ing good from bad, inadequate, or un- naturalists do have at least indirect access to
trustworthy research. Clearly such criteria the multiple realities they deal with: since
have importance for designing, monitoring, these realities are in the minds of people (as
and judging an inquiry, whether from the idiosyncratic constructions that vary from
perspective of the inquirer, a monitor (for individual to individual), naturalists can ask
example, a sponsor, an administrator, or a those people whether their realities have
dissertation committee), or an editor who been represented appropriately. Thus, the
might be asked to publish the results of such crucial question for the naturalist becomes,
research. "Do the data sources (most often humans)
Guba and Lincoln (1981) have sum- find the inquirer's analysis, formulation,
marized the four major traditional criteria and interpretations to be credible (believa-
into four questions, as follows, to which ble)?"
they suggest the naturalist has an equal ob- 2. Transferability. In the rationalistic
ligation to attend: paradigm, generalizability (external valid-
1. Truth value. How can one establish ity) is demonstrated by showing that the
 NATURALISTIC INQUIRY 247

data have been collected from a sample that However, Guba (1981) has attempted what
is in some way (randomized, stratified, etc.) he himself characterized as a primitive ef-
representative of the population to which fort. His formulations will be summarized
generalization is sought. The naturalist, dis- here by way of illustration of the proposi-
counting generalizability (see Axiom 3) and tion that methods of dealing with the
the assumption of context-free laws which trustworthiness of naturalistic inquiries are
have enduring truth value (especially as forthcoming and can be expected to be ex-
they might relate to human and sociobehav- panded on in the near future (Lincoln &
ioral phenomena), nevertheless believes Guba, in preparation).
that some degree of transferability is possi- With respect to credibility, Guba suggests
ble under certain circumstances. Those cir- the following as means either to safeguard
cumstances exist if enough "thick descrip- against loss of credibility or to continually
tion" is available about both "sending" and test for it:
"receiving" contexts to make a reasoned 1. Prolonged engagement at a site, to over-
judgment about the degree of transferabil- come distortions introduced by the in-
ity possible. quirer's presence, to test for ethnocentrism
3. Dependability. In the rationalist (Lincoln & Guba, 1981), to test biases and
paradigm, reliability is a matter of replica- perceptions of both inquirer and respon-
bility; a study ought to be repeatable under dents, and to provide time to identify sa-
the same circumstances in another place lient characteristics of both the context and
and time. If there are discrepancies or devia- the problem.
tions between two repetitions of the same 2. Persistent observation, to gain a high de-
study, the difference is charged off to unre- gree of acquaintance with and understand-
liability (error). The naturalist cannot be so ing of "pervasive" qualities and salient
cavalier, however; first, designs are emer- characteristics as well as to appreciate atypi-
gent so that changes are built in with con- cal but critical characteristics and to elimi-
scious intent, and second, emergent design nate those that are irrelevant.
prevents an exact replication of a study in 3. Peer debriefing, to keep inquirers "hon-
any event (since a second inquirer might est," to provide them with the opportunity
choose a different path from the same data). to test their growing insights against those
The naturalist defines the concept of depen- of uninvolved peers, to receive advice about
dability to mean stability after discounting important methodological steps in the
such conscious and unpredictable (but ra- emergent design, to leave an audit trail (see
tional and logical) changes. below), and to discharge personal feelings,
4. Confirmability. As Scriven (1971) has anxieties and stresses that otherwise might
noted, the rationalistic concept of objectiv- affect the inquiry adversely.
ity is based on a quantitative notion of in- 4. Triangulation, whereby a variety of
tersubjective agreement. But clearly, 50 data sources, different perspectives or
million French people can be and have been theories, and/or different methods are pit-
wrong; what is important is not that there ted against one another to cross-check data
be quantitative agreement but qualitative and interpretation (Denzin, 1978).
confirmability. The onus of objectivity 5. Referential adequacy materials, that is,
ought, therefore, to be removed from the documents, films, videotapes, audio re-
inquirer and placed on data; it is not the cordings, pictures, and other "raw" or
inquirer's certifiability we are interested in slice-of-life materials are collected during
but the confirmability of the data. 9 The nat- the study and archived without analysis;
uralist finds this a powerful formulation. these materials can later be utilized by the
It is premature to expect that adherents inquirer or others (especially an auditor; see
of the naturalistic paradigm would have below) to test interpretations made from
evolved as sophisticated a methodology for other analyzed data.
dealing with rigor and trustworthiness 6. Member checks, whereby data and in-
questions as have their rationalistic coun- terpretations are continually checked with
terparts, especially since the latter have had members of various groups from which
literally centuries to work on refinements. data are solicited; done on a continuous
248 ECTJ WINTER 1982

basis throughout the study and again at the cal assumptions, reasons for formulating
end when the full report is assembled, the study in a particular way, and implicit
using either (or both) the same members assumptions, biases, or prejudices about
from whom the data were originally col- the context or problem; the most appropri-
lected or other surrogates from the same ate means for this exploration and presenta-
groups. tion takes the form of a reflexive joumal,
kept in the field.
With respect to transferability, Guba has
suggested that the inquirer engage in or 3. The confirmability audit, a counterpart
to the dependability audit, in which the au-
provide the following:
ditor takes the additional step of verifying
1. Theoretical~purposive sampling, that is,
(or warranting) that each finding can be ap-
sampling intended to maximize the range of
propriately traced back through analysis
information collected and to provide most
steps to original data, and that interpreta-
stringent conditions for theory grounding.
2. Thick description, by which is meant tions of data clusters are reasonable and
meaningful, in much the same way that a
providing enough information about a con-
fiscal auditor would verify at least a sample
text, first, to impart a vicarious experience
of it, and, second, to facilitate judgments of entries in a bookkeeping journal to be
about the extent to which working hypoth- certain that each represented a "real" trans-
action and that the "bottom line" accurately
eses from that context might be transferable
to a second and similar context. represented the actual financial situation.
The criteria we have posed, although not
With respect to dependability, Guba has
theoretically elegant formulations, do have
suggested the following:
utility at several stages of the inquiry pro-
1. Use of overlap methods, one kind of
cess, as Guba (1981) has pointed out. They
triangulation process, which, although
aid and abet in making a priori judgments
usually advocated in support of validity,
about the quality of proposed research, as in
also undergirds claims of reliability to the
the case of peer reviews of proposals or
extent that they produce complementary
results. committee review of proposed dissertation
research. A naturalistic design ought at
2. Stepwise replication, a kind of "split-
least to propose what the inquirer will do to
halves" approach to which inquirers and
satisfy each of the above criteria and
data sources are split into two roughly equal
provide for trustworthiness. The criteria aid
halves to be investigated independently,
inquirers in monitoring themselves and in
provided, however, that there is frequent
guiding activities in the field, as a way of
exchange between the two teams to allow
determining whether or not various stages
for the proper development and unfolding
in the research are meeting standards for
of an emergent design.
quality and rigor. Finally, the same criteria
3. The dependability audit, modeled on the
may be used to render ex post facto judg-
fiscal audit, but limited to that part of the
ments on the products of research, includ-
auditor's role which deals with process. In a
ing reports, case studies, or proposed pub-
fiscal audit, the first concern of an auditor is
lications. The final reports ought at the very
whether the accounts were lcept in one of
least to include, as do rationalistic paradigm
the several modes that constitute "good,"
reports, statements about what the inquirer
or acceptable, professional practice; to reach
actually did to satisfy each of the four sets of
that judgment the auditor must of course be
criteria, and reports from dependability and
supplied with an "audit trail" which de-
confirmability auditors concerning their
lineates all methodological steps and deci-
verification of the inquirer's processes and
sion points and which provides access to all
conclusions.
data in their several raw and process stages.
Guba has made the point that to carry out
With respect to confirmability, Guba has even all of these steps (usually not logisti-
proffered the following: cally or fiscally possible in an actual inquiry)
1. Triangulation, as described above. will not guarantee the trustworthiness of a
2. Practicing reflexivity, that is, attempting naturalistic study. These criteria, unlike
to uncover one's underlying epistemologi- those of the rationalistic paradigm, do not
 NATURALISTICINQUIRY 249

provide an unassailable defense against ac- ence. We believe that paradigm to be the
cusafions of nonrigorousness, but they will naturalistic.
contribute greatly toward persuading a The naturalistic paradigm has emerged,
reader and consumer of their meaningful- in part, out of concern that research and
ness. He has noted too the special risks that evaluation studies reflect what is, rather
are posed for the naturalistic practitioner than what some researcher thinks ought to
who operates within a rationalistic culture. be, and, in part, from intense scrutiny of the
Funding sources, sponsors, and disserta- assumptions and epistemological axioms
tion committee members who are asked to which undergird rationalistic inquiry. It is
make judgments about the utility of such to those concerns about which the preced-
inquiry do not have at their disposal a com- ing arguments have been addressed, and,
plete design about which they may feel therefore, we have tried to make explicit the
great certainty. Because of the open-ended, nature of the epistemological assumptions
emergent nature of the design, they may which undergird the two paradigms,
feel the inquirer is asking them to extend rationalistic and naturalistic, and have ad-
approval to an unknown quantity, and may dressed persistent criticism that the latter is
therefore not wish to put a stamp of legiti- "soft," "nonrigorous," and attentive to re-
macy on what might turn out to be sloppy levance over rigor.
research. The barriers and constraints to Although it is true that rationalistic in-
carrying out such research are, therefore, quirers do not accept the axioms we have
many, but it is hoped that the criteria set imputed to them here without reservation,
forth may stimulate those who have not we have tried to deal with them in their
considered carrying out or sponsoring such purest form, as they can be traced through
research to take a chance. the philosophy of science and scientific
Finally, we would like to warn against the writers. By doing so, we believe, the reader
adoption of these procedures as another or- is able to see the sharpest of contrasts and to
thodoxy. The suffocating grip which the understand better why it is maintained that
rationalistic paradigm and experimental there can be no compromise on axiomatic
methods have exerted on the social and be- assumptions, although there may be com-
havioral sciences for so many years should promise on various "postures" which are
serve as warning that no one set of proce- typically ascribed to the two paradigms.
dures may be taken as gospel or represent Thus, we have accounted for five major
prescriptions for how inquiry must be done. differences between the two paradigms: the
Rather, any group of procedures represents nature of reality, the nature of the
merely a set of guidelines that may facilitate inquirer-object (or respondent) relation-
the development of this new mode of in- ship, the nature of truth statements, as-
quiry. sumptions about causal relationships, and
the role of values within disciplined in-
quiry. We have asserted, for instance, the
SUMMARY following: (a) reality, as it is lived by the
We have tried to argue here that we are in subjects of research, is not fragmentable
the midst of a paradigmatic revolution into variables and processes, but is rather
(Kuhn, 1962). That revolution centers about experienced holistically and mediated heav-
the growing concern that the paradigm ily by values, attitudes, beliefs, and the
which we have typically utilized for scien- meanings which persons ascribe to their
tific (hard and life sciences) inquiry has ill- experiences, and as a result, inquirers must
served us when applied to the social and approach h u m a n subjects and h u m a n
behavioral sciences, and, further, that at- p h e n o m e n a holistically rather than in
tention to rigor within this paradigm has piecemeal fashion; (b) inquirer and subject
meant increasingly that results of social invariably interact; it is not possible to main-
studies are less and less applicable and less tain a discrete and inviolable distance be-
and less "life-like" as time has passed. It is tween the inquirer and the subjects of the
now time for a new paradigm, one which research; and the appropriate response of
takes account of the nature of social experi- inquirers to this natural and unavoidable
250 EGTJ WINTER 1982

interactivity is to both "track" it and exploit before the person can express them in ver-
the insights it lends by becoming "smarter" bal (or propositional) form. The possibility
instruments themselves; (c) truth state- of utilizing both types of knowledge, in
m e n t s expressed as g e n e r a l i z a t i o n s - - turn, abrogates the necessity of specifica-
enduring, context-free laws, particularly tion of design early on, or prior to the re-
about human behavior--are not possible search, so that designs can now be consid-
except in the hard sciences, and human be- ered emergent; that is, they can unfold as
havior, bounded as it is by time and context, the h u m a n i n s t r u m e n t discovers new
is best described by means of "working knowledge and reshapes inquiry to fit with
hypotheses," temporary assertions about the context. Thus, although naturalists are
context-specific situations; (d) the search for not compelled to adopt one posture or
direct, highly "tied" and systematic cause- another, initial choices about theory and in-
effect relationships in human affairs is of struments may lead them to other decisions
little utility when human beings are caught that appear to be choices but are almost
up in interactive webs and patterns of fac- compelled by the nature of the problem, the
tors, events, processes, and ascribed mean- paradigm, and the context of the inquiry.
ings, so that it is more important to search Finally, we have argued that although
for factor patternings than for "if-then" several centuries of rationalistic inquiry have
causal chains; and, finally, (e) inquiry is al- allowed the development of rather strict
ways value bound and is never in the social and inviolable canons of rigor, the naturalis-
sciences (and rarely in the physical sciences) tic school is only beginning to develop an
value free; rather the choice of a problem, a arsenal of weapons against the charge of
paradigm to guide the investigation of the nonrigor or untrustworthiness. We have
problem, a substantive theory and proce- demonstrated that it is possible to consider
dures to guide collection and analysis of the the questions of internal validity, external
data, and the context in which to conduct validity, reliability, and objectivity within
the research are all arenas where values the framework of naturalism, to reject those
enter into empirical inquiry. Along those labels as inappropriate, to argue for con-
assumptions, we have argued, there can be cepts that are more germane, i.e., credibil-
no compromise between paradigms. The ity, transferability, dependability, and con-
inquirer must choose one set of assump- firmability, and to propose criteria by which
tions (axioms) or another to undergird his or external reviewers of naturalistic research
her inquiry. might judge the trustworthiness of those
Along certain other dimensions called studies. Although those criteria do not as
postures, however, compromise may be yet provide unassailable defenses against
possible, although we would argue that, charges of u n t r u s t w o r t h i n e s s , they
like dominoes, one choice may impel the nevertheless assure the consumer of such
inquirer to come to make other choices that research that any and all appropriate steps
traditionally have characterized naturalistic have been taken to assure that data from
inquiry. That set of associational decisions human sources and contexts are meaning-
has probably inadvertently given rise to the ful, trackable, verifiable, and grounded in
current "unhelpful distinction" between the real-life situations from which they
rationalistic versus naturalistic paradigms were derived.
as they are confused with quantitative ver-
sus qualitative methods. Nevertheless, the
decision to utilize the h u m a n - a s - NOTES
instrument, for example, probably militates 1. In previous writing (Guba, 1978, 1979; Guba
heavily in favor of using largely qualitative & Lincoln, 1981) we referred to what we here call
methods, since those are the methods at the rationalistic paradigm as the scientistic or the
which the human is most skilled. Qualita- scientific paradigm. The use of even the less
tive methods, in turn, allow the instrument pejorative of these latter two terms now seems to
us inappropriate on two counts. First, readers
of choice to utilize both propositional and have tended to view the naturalistic paradigm as
tacit knowledge, building on and testing less scientific (or even as nonscientific) and have,
"hunches," intuitions, and insights, long therefore, denigrated it as less valid, probably
 NATURALISTICINQUIRY 251

because of the enormous legitimation accorded to by the naturalistic. We are dealing, however,
anything scientific in our culture. Second, several with the large majority of studies that could be
critics have accused us of setting up a straw man, undertaken by psychologists, sociologists, an-
on the grounds that vanguard scientific thinkers thropologists, educational researchers, and
have moved beyond the nineteenth century logi- evaluators, including evaluations of other social
cal positivism of which our descriptions are at process fields such as social work, law enforce-
times reminiscent. It is undoubtedly true that ment, or health services delivery.
many scientists now think differently, but that
change does not characterize, in our opinion, the 5. We find the use of the term object of inquiry
large majority of "scientists" who engage in in- when applied to a human pejorative; we prefer
quiries in either the hard or soft sciences. There, the term respondent, which carries the connota-
the old culture still dominates. It is to that level of tions of interaction and equality.
practice that our criticisms are directed, and it is 6. It is ironic that the naturalist does permit the
of that moribund culture that our descriptions are data to "speak for themselves" in the sense of
apt. However, to avoid the unintended meanings grounding theory in them, a use never con-
that some readers have drawn from our work, we templated by the rationalists fond of using that
have shifted to the term rationalistic to describe phrase as an assertion of their "objectivity.'"
the paradigm that guides so much conventional
inquiry. 7. The distinction made here is similar to that
between a connoisseur and a critic of art. The
2. The propensity of the rationalistic model to connoisseur need only "feel" a painting to appre-
avoid process considerations is not based upon a ciate it; the critic must cast feeling into lan-
lack of interest in process but an essential inabil- guage to convey a critique. Connoisseurship is a
ity to deal with it. In the early decades of this private art, but criticism is a public art (Eisner,
century, for example, physicists were obsessed 1979).
with modeling the atom, and a variety of models
were proposed, ranging from the quaint "plum- 8. It seems quite clear that the rationalist de-
pudding" model to Bohr's sophisticated orbiting pends on intuition as much as does the naturalist;
electron model. But all of these models proved however, the reconstructed logic (Kaplan, 1964)
unsatisfactory. Moreover, means were not (nor of rationalism militates against making such de-
are) available to "see" inside the atom in any pendency public. The rationalist's unforgivable
event. Atoms came to be regarded as "black sin is to own up to humanness.
boxes" which could be manipulated from the 9. Once again, we would point out that al-
outside and which would produce reactions, but though the rationalists say that "the data speak
the process by which the stimuli were reacted to for themselves," in fact the assertion is, more to
(inside the black box) remained a mystery. The the point, that the methods speak for themselves,
inability of physicists to deal with process and the since the methods, if they conform to standard
invention of the "black box" idea came to be and accepted practice, serve to insulate
viewed, in an interesting reversal, as the proper rationalists from the subjects of inquiry and to
way to do research--stimulate, wait for reaction, provide for isomorphism and generalizability.
observe reaction, and never mind how stimulus Naturalists, however, recognize that when deal-
came to be translated into reaction. The familiar ing with human beings as data sources, they can-
pretest-treatment-posttest design was of course not, and probably should not, be "isolated" from
admirably suited to this posture, and came to be them. Rather, the data can and must surely speak
viewed as the methodological alternative of for themselves; although both inquirer and
choice. But what physicist would forego looking methods must be publicly inspectable, ultimately
inside the atom if able to do so? And if, in other it is the data which must stand or fall on rigorous
areas, process can be examined, even if only in- audit.
trospectively, why persist in the use of a model
that ignores that possibility?
REFERENCES
3. However, the naturalist would not wish to
eliminate interactivity with a respondent even if Bahm, A. J. Science is not value-free. Policy Sci-
possible. See the discussion of this axiom in the ences, 1981, 2, 391-396.
following section. Campbell, D. T., & Stanley, J. C. Experimental
and quasi-experimental designs for research
4. An appreciation of the constraints which on teaching. In N. L. Gage (Ed.), Handbook of
this limitation places on the subsequent discus- research on teaching. Chicago: Rand, McNally,
sion is crucial to an understanding of the point we 1963.
will make. We are not dealing with tangible ob-
jects, events, or processes as would the physicist, Cook, T . D . , & Campbell, D . T . Quasi-
chemist, or biologist. Nor do we mean to include experimentation: Design and analysis issues for field
those aspects of human studies that can be settings. Chicago: Rand, McNally, 1979.
labeled as genetically or d e v e l o p m e n t a l l y Cook, T. D., & Reichardt, C. I. (Eds.) Qualitative
mediated. Study of such matters is undoubtedly and quantitative methods in evaluation research.
better guided by the rationalistic paradigm than Beverly Hills, Calif.: Sage Publications, 1979.
252 ECTJ WINTER 1982

Cronbach, L. J. Beyond the two disciplines of tion. Educational Evaluation and Policy Analysis,
scientific psychology. American Psychologist, 1980, 2, 37-45.
1975, 30, 116-127. Kuhn, T. S. The structure of scientific revolutions.
Cronbach, L. J., & Suppes, P. Research for tomor- Chicago: University of Chicago Press, 1962.
row's schools: Disciplined inquiry in education. Lincoln, Y. S., & Guba, E. G. Do evaluators wear
New York: Macmillan, 1969. grass skirts? Going native and ethnocentrism as
Denzin, N. K. The logic of naturalistic inquiry. In problems of utilization in evaluation. Paper pre-
N. K. Denzin (Ed.), Sociological methods: A sented at the joint annual meeting of the Eval-
sourcebook. New York: McGraw-Hill, 1978. uation Network and the Evaluation Research
Eisner, E. The educational imagination. New York: Society, Austin, Texas, September 1981.
Macmillan, 1979. Lincoln, Y. S., & Guba, E. G. Issues in naturalistic
Filstead, W. J. Qualitative methods: A needed inquiry. Manuscript in preparation.
perspective in evaluation research. In T. D. Mill, J. S. A system of logic. New York: Longmans,
Cook & C. I. Reichardt (Eds.), Qualitative and Green, 1947. (Originally published, 1843.)
quantitative methods in evaluation research. Be- Perrow, C. Disintegrating social sciences. NYU
verly Hills, Calif.: Sage Publications, 1979. Education Quarterly, Winter 1981, 530-541.
Glaser, B. G., & Strauss, A. L. The discovery of Polanyi, M. The tacit dimension. New York:
grounded theory. Chicago: Aldine, 1967. Doubleday, 1966.
Guba, E. G. Toward a methodology of naturalistic Rice, J. M. The futility of the spelling grind.
inquiry in educational evaluation (Monograph Forum, 1894, 163-172; 410-419.
Series No. 8). Los Angeles: University of Scriven, M. Objectivity and subjectivity in educa-
California, Center for the Study of Evaluation, tional research. In L. G. Thomas (Ed.), Philo-
1978. sophical redirection of educational research (71st
Guba, E. G. Criteria for assessing the trustwor- Yearbook of the National Society for the Study
thiness of naturalistic inquiries. Educational of Education, Pt. 1). Chicago: University of
Communication and Technology Journal, 1981, 29, Chicago Press, 1972.
75-92. Tranel, D. D. A lesson from the physicists. Per-
Guba, E. G., & Lincoln, Y. S. Effective evaluation. sonnel and Guidance Journal, 1981, 59, 425-428.
San Francisco: Jossey-Bass, 1981. Webb, E. J., Campbell, D. T., Schwartz, R. D., &
Hofstadter, D. R.G6del, Escher, Bach: An eternal Sechrest, L. Unobtrusive measures. Chicago:
golden braid. New York: Basic Books, 1979. Rand, McNally, 1966.
Homans, G. A. What kind of a myth is the myth Weir, E. Types of explanation in educational evalua-
of a value free social science? Social Science tion (Paper and Report Series, No. 34, Research
Quarterly, 1978, 58, 530-541. on Evaluation Project). Portland, Ore.: North-
Kaplan, A. The conduct of inquiry. San Francisco: west Regional Educational Laboratory, 1980.
Chandler, 1964. Wiles, D. D. The logic of Y = f(X) in the study of
Kelman, H. C. A time to speak: On human values educational politics. Educational Evaluation and
and social research. San Francisco: Jossey-Bass, Policy Analysis, 1981, 3, 67-74.
1969. Zukav, G. The dancing Wu Li masters. New York:
Krathwohl, D. R. The myth of value-free evalua- Bantam Books, 1980.