108 MEANING AND METHODOLOGY
Again, there are interconnections between negative case analysis and other techniques and crite-
ria already reviewed. Negative case analysis, like informant feedback/member checks, is aided by
reflexivity and, particularly, by the use of a reflexive journal to examine self-other relations as well
as the researcher’s evolving understandings of purpose and theory (see Soss, chapter 6, this volume,
pp. 136–37). An interpretive understanding of triangulation, one of the first-order criteria, similarly
pushes the researcher to look for and analyze inconsistencies and contradictions, and the admoni-
tion to “thickly describe”—another first-order criterion—heightens researchers’ observational skills
and the possibilities for detecting the ambiguities, missteps, and covert disagreements among actors
in a field setting. In sum, when asked the question, How do I know that you didn’t only look for
confirmatory evidence?, interpretive researchers who follow these procedures can lay out for a
reviewer or other reader a number of different techniques employed.
The extent to which these techniques are effective and persuasive depends not only on re-
searchers’ careful preparations but also on the character of the audience. The use of any of the
techniques reviewed here does not guarantee that readers will be satisfied with the researchers’
answers to the three questions commonly put to them. As discussed in chapter 1 (see pp. 20–21 on
reader-response theory), interpretive researchers appreciate that they are contributing to a con-
versation, that there will be another “interpretive moment” because readers bring their own back-
grounds and experiences, their own meaning making, to researchers’ texts.BUT WHAT ABOUT CAUSALITY AND GENERALIZABILITY?As causality and generalizability are a constitutive part of the variables gestalt and are seen, along
with reliability and validity, as hallmarks of “scientific” work, a word is necessary concerning
their meaning from an interpretive perspective. In their 1985 book Lincoln and Guba discussed
causality in terms of “internal validity”—the phrase popularized by Campbell and Stanley’s (1963)
canonical treatise, Experimental and Quasi-Experimental Designs for Research, which has been
particularly influential in the fields of education and psychology and has been widely used as the
basis for discussion of research design in social science methods texts. Campbell and Stanley held
up the true experiment as the preferred method for determining causal relationships.^29 They show
how the true experiment controls for eight “threats” to internal validity, that is, threats to the claim
that a change in the independent or “treatment” variable was responsible for a change in the
dependent variable. This mode of thinking about causality is still the modus operandi in a variety
of fields, notably mainstream evaluation research, psychology, and medical research. It is an
understanding of causality tied to a search for general, predictive laws of human behavior.
Given the interpretivist rejection of the possibility of ahistorical, acultural laws, it is not sur-
prising that some have imputed to interpretive scholars a disinterest in causality (see Gerring’s
[2003b, 27] question to Clifford Geertz). But, as Geertz’s reply clarifies, what is at issue is the
understanding of causality. The “causal laws” perspective emphasized by Campbell and Stanley
(and hegemonic in methods texts) is only one way to think about the concept. Causality is receiv-
ing renewed attention in a variety of fields,^30 and interpretive researchers are coming to better
understand the ways in which they can reclaim this powerful term on their own ground.
Two interrelated moves are possible. One is to reorient the understanding of causality from
general laws to specific cases. Telling quite specific causal stories about how a policy was imple-
mented (as done, e.g., by Pressman and Wildavsky 1973) is not inconsistent with interpretive
presuppositions; far from it. This is “Sherlock Holmes causality”—a careful mapping of clues in
context, a tracing of connections among events. Interpretive methods lend themselves very well
to demonstrating such causality.