EVALUATIVE CRITERIA AND EPISTEMIC COMMUNITIES 93Altogether, the data retained in their diverse forms—what might be termed the genres of the
data—constitute the complex human meaning-making enterprise, which interpretive methods
are particularly suited to studying. Respecting the various genres of data enables interpretive
researchers to offer holistic, multifaceted understandings of human experience—not only of
the human love of numbers and words, but of the ways in which sight, sound, and embodiment
make possible and limit human activities, practices, and meaning making. To be clear, re-
searchers from the variables gestalt also study space, objects, documents, and human mean-
ings, but their approach is to “trans-form” data into numbers, whenever possible, in order to be
able to apply statistical analytic techniques in the ways and for the purposes illustrated in the
thought experiment.^5 It is for these reasons that the experience of reading interpretive empirical
research (in all its voluminous complexity, as attested to in this volume) is usually a very
different one from reading variables-based research. How, then, should interpretive research
be judged?
THE BEGINNINGS OF THE “CRITERIA LITERATURE”:
CLASSIC TEXTS
In the late 1970s and the early 1980s, scholars began to articulate and develop the criteria implicit
in qualitative-interpretive^6 research practices as distinct from those used in the variables gestalt.
Two texts are recognized as “classics” in the development of this literature (Creswell 1998b), and
they serve as useful benchmarks for understanding its evolution. Miles and Huberman’s Qualita-
tive Data Analysis: A Sourcebook of New Methods (1984) and Lincoln and Guba’s Naturalistic
Inquiry (1985) were among the first to present fully articulated responses to skepticism (much of
it coming from researchers influenced by positivist understandings of scientific practices) con-
cerning the scientific character of qualitative methods, and these texts are much cited in the litera-
ture on evaluative criteria.
Although Miles and Huberman did not add a full-fledged discussion of evaluative criteria until
their second edition (published in 1994), their 1984 text is notable because it focuses on qualita-
tive data analysis, the stage most maligned by methodological positivists as unreliable and sub-
ject to bias. Indeed, Miles and Huberman accept this critique and seek to develop “reliable and
valid” methods of analysis for “word data.” Their chapters on coding provide a variety of tech-
niques (e.g., writing memos; pattern coding; time, role, and effects matrices) for what they term
“data reduction”—by which they mean the process of “selecting, focusing, simplifying, abstract-
ing, and transforming” word data (1984, 21). As they emphasize, data reduction need not mean
quantification because “[q]ualitative data can be reduced and transformed in a variety of ways:
through sheer selection, through summary or paraphrase, through terms being subsumed in a
larger pattern and so on” (1984, 21). The Miles and Huberman text, then, is traditional in its
emphasis on coding, but it still deserves being labeled “qualitative-interpretive” because of the
authors’ healthy skepticism of quantification and their emphasis on qualitative data analysis as a
meaning-making enterprise built primarily on the analysis of word data retained in their contexts
of origin.^7
In contrast to Miles and Huberman, Lincoln and Guba (1985) chart what they term a
“postpositivist” approach, and, in their chapter on assessing the quality of interpretive research,
they explicitly reject the possibility of “universal criteria,” that is, criteria understood as applying
to the evaluation of the methods used in all research gestalts. The four evaluative criteria they
propose as appropriate to interpretive research are parallel to the four criteria rooted in method-
ological positivism with whose critique their analysis begins. So, for example, instead of the