Should Values Be Written Out of the
Social Scientist's Lexicon?*
MICHAEL HECHTER
Rational choice theory' is often faulted for treating values, preferences, and other internal
states as exogenous to its explanations. Calls for an endogenous theory of values have
been gaining momentum among critics of rational choice for the last decade.2 They make
it seem as if including values in social theories is merely a matter of will. Yet these critics
seldom have even defined values, let alone showed how they can be incorporated fruitfully
into behavioral explanations. Nor have they proposed any endogenous theories of values;3
indeed, there is a striking absence of discussion about what such a theory might consist
of. In general, there has been too little appreciation of the difficulty of incorporating
values in theoretical models
Before we assess the prospects for an endogenous theory of values, it would be wise
to recognize one cautionary fact. Whereas 30 years ago values occupied a central place
in all of the social science disciplines (even in economics),5 use of the concept has declined
precipitously in each discipline (for sociology, see Spates 1983; for political science, see
Barry 1970, p. 180; for anthropology see Barth 1993).6 Despite the strong intuition that
evidence about the full range of values is needed in order to explain behavior fully, Stigler
and Becker's (1977) claim that no significant behavior has ever been illuminated by
assumptions of differences in values (or what they refer to as tastes) probably echoes the
current wisdom in positive social science.7
This paper is not yet another plea for the importance of values. Instead, I want to focus
on the problem of how values may be studied so that ultimately they can be made
endogenous in our theories. In the first part, I discuss some of the reasons why the concept
has largely been abandoned in social scientific discourse. Progress in scholarly research
on values has been slow because of several major impediments: in all their forms, values
are unobservable; theory in sociology, economics, and psychology provides little guidance
for understanding how values shape behavior; simply postulating values is unconvincing
when the processes responsible for generating them are unknown; and, finally, measurement
problems abound. Social researchers will pay heed to values only if these impediments
are appreciated and given their due. In the second part, I suggest that the key to
achieving progressi n understandingt he genesis and efficacy of values lies in findingb etter
measures of them, and propose that the study of advance directives for medical treatment
offers a research site that is well suited to explore these issues with respect to one highly
salient value, that of life itself
FOUR IMPEDIMENTS TO THE STUDY OF VALUES
1. Values can take many forms, but all of these are unobservable
2. Existing theoretical traditions provide little guidance for understanding how values
shape behavior.
3. Postulating values in behavioral explanations is unconvincing when the processes
that generate them are unknown.
4. Measurement problems abound.