170 ACCESSING AND GENERATING DATA
that people’s attitudes toward self-disclosure are universal and identical to those of U.S. survey
makers. Only a close knowledge of respondents would make apparent that this is not always
the case. McNabb’s (1990) research, for example, found that large-scale preference surveys in
several native Alaskan communities were highly problematic. The native Alaskan respondents
were loath to give strong answers and tended to cluster them toward the middle numbers of the
Likert scale. Restraint, McNabb found, was viewed as a sign of maturity and rationality. Also,
the very structure of the survey was alien; data collection in the form of storytelling, while
providing far fewer responses, would have been far more nuanced, and those responses could
have set the stage for some type of larger-scale questioning using native conceptual categories
rather than the etic ones used by the U.S. government workers. Unlike many social scientists
who believe that large-scale surveys are valid because of their sample size and replicability,
McNabb argues that verification and validity are questionable because there are too many
sociolinguistic and cultural factors in survey responses affecting researchers’ interpretations of
respondents’ replies.
Other ethnographic studies provide similar cautionary tales and useful lessons about assuming
a monolithic stance cross culturally (in the broadest sense, including across traits such as rural/
urban divides, regional urban differences among different cities within a single country, differ-
ences across ethnic groups, and other types of difference demarcating interpretive communities).
Often, just reading an ethnographic case study provides unanticipated insights that even the eth-
nographer might not have noticed, but that make sense to readers from different perspectives who
might be asking different questions than the ethnographer. In a sense, every ethnography is a
living, moving animal with tentacles drawing in many questions and answers from places previ-
ously unknown, or known as part of a different context with a different meaning than the
researcher’s previous expectations would lead her to expect.ETHNOGRAPHIC SENSIBILITY AND POLICY ANALYSISParticipatory, experiential, and interpretive approaches to daily behaviors provide a good starting
point for understanding various rationales underlying past and present social policies. These ap-
proaches start with the assumption that a more profound understanding of the relation between
everyday behavior and the success of policies can have significant social and policy implications.
Many research stories illustrate ways in which participant-observation makes the difference be-
tween accurate interpretations with strong explanatory powers and inaccurate analyses. The latter
often perpetuate, even exasperate, ineffectual programs because they start from misconstrued,
and often negative, externally derived stereotypes rather than from familiarity. Analyses that start
from inaccurate premises or from an ill-informed base are less likely to arrive at truly useful and
appropriate programmatic solutions. Too many social policies start from policy analysts’ own,
emic expectation that others will respond to given situations as they themselves would. Too many
analyses start from a commonly unspoken, well-intentioned, but culturally constrained platform
of “I want others to have what I want for my own family.”^7 As a result, many decision makers end
up working from an implicit, individualistic, and inaccurately stereotyped perspective (assuming,
for example, that people can succeed if they try hard enough or if they can be taught to respect
work over baby making and drug dealing).
The studies related here present only a small selection of cases in which participant-observation
demonstrated its efficacy for analyzing and designing policy. In each, had policy analysts or
policy makers asked questions from the beginning, rather than making assumptions that all of the
stakeholders were, or should have been, more like the politically and economically dominant