8
In our assessment of these ideas using six conditions of our experiment, all three
models did a very good job accounting for the data (mean absolute values of P(s)
discrepancies, predicted – actual) range from .0228 (for the aggregation model) to .0459
(for the source model). Here, the fit of the aggregation model is improved a bit, and the
fit of the two variants is a bit weaker. However none of the models does a poor job,
lending confidence that, together, they give a good understanding of how individuals
process structural and interactional information in this type of situation.
6
Ten New Models.
Next, we formulate and assess a variety of models that incorporate a variety of
ideas about how individuals process status claims. The ideas are all plausible, and many
of them figure in important sociological writing. For all of these, we focus on effects of
overt claims, because these are the new elements in our theoretical and experimental
work. In all the new models (except #5), we presume structural status differences will
have effects, and we focus on effects of expressed status claims from different sorts of
individuals. Structural status effects are well supported by several decades of research,
and our models here presume that they will have the same sorts of effects as have been
widely reported in the past. Effects of claims, however, have only recently been subjects
of theory building, and we focus on those effects here.
4. Egalitarian Model. This model weights all status claims equally; it does not
give preference to those from a high status individual. This model captures a simple
version of the central ideas in Cooley’s “Looking Glass Self,” that individuals form their
self-concepts through noting opinions of others.
6
Models 1—3 have already been assessed using the first 6 conditions of the experiment. Here we use all 8
conditions, so the fit statistics differ somewhat from the earlier tests.
Convention