Quality & Quantity 37: 21–41, 2003.
© 2003 Kluwer Academic Publishers. Printed in the Netherlands.
Class Analysis and Earnings Inequality: Nested and
Non-Nested Comparisons of Two Class Models in
ILAN TALMUD, VERED KRAUS and YUVAL YONAY
Department of Sociology, University of Haifa, Haifa, 31905 Israel
Abstract. This paper demonstrates how nesting and non-nesting analytical strategies provide differ-
ent answers regarding the comparative utility of theoretical models. This paper demonstrates this
incompatibility by testing the empirical efﬁcacy of Goldthorpe’s and Wright’s class schemes in
explaining earnings inequality in Israel. These models are non-nested, because while they partially
overlap each other conceptually and empirically, neither can be written as a parametric restriction of
the other. As they are non-nested, we cannot test each model against the other by using the conven-
tional sociological approach to hypotheses testing. For the sake of demonstration, however, we show
results obtained from the conventional Ordinary Least Squares regression models with conventional
Baysian Information Coefﬁcient statistic, serving as criterion for a decision rule. Wright’s model was
found to be more signiﬁcant in explaining earnings variations in Israeli society. Yet when we used
two models of non-nested speciﬁcation tests (the Cox–Pesaran model and the J test) to examine each
model’s unique contribution, neither of these models were able to reject the rival hypothesis.
It is customary in sociological research to look for the “best plausible model” by
comparing two or more theoretical explanations, which are not mutually exclusive.
A conventional analytical strategy to achieve this goal is to derive two competing
hypotheses from the rival explanations and to contrast each hypothesis against the
null hypothesis. Weighing the results of both tests against each other, and selecting
the one that accounts for more variability then chooses the “best” explaining model.
The problem with this practice is its methodological assumptions. Formally,
employing this conventional strategy means that one model can be re-written as a
parametric restriction of the other one. This results from two possibilities: a com-
mon assumption is that the theoretical models are mutually exclusive.
that each model is the logical complementary of the other one. It is possible, there-
fore, to combine the two models into a joint artiﬁcial model, comprising of both
mutually exclusive, rival hypotheses and to test it against the null hypothesis. An-
other workable assumption is that the two models are hierarchically nested within
In other words, like other widely used testing methods, this method
employs nested tests, although the theoretical overlap of the tested theories requires