7
conflict over social policy underlines both the importance of this policy realm and
the difficulty in having a single continuing measure: public order, religious
observance, and gender roles, for example, as well as their embodiment in
policies on crime and punishment, school prayer, and gay rights. Nevertheless,
the character of national life is in some sense the focus of them all, and the
problem for public policy is the question of governmental intervention.
(Scammon & Wattenberg 1970, Hunter 1991)
Measures for issue domains defined this way then had to have two fundamental
prerequisites of their own. First, they had to have face validity on their individual items.
That is, the surface content of all items had to reflect critical elements of the theoretical
dimension. And second, those individual items had to scale collectively. That is,
appropriate surface contents had to be correlated with each other, so that they were in fact
measuring at least facets of the same underlying phenomenon. Accordingly, for each
election year in the NES series, all items with clear policy implications were isolated and
assigned to one of six policy categories: welfare, foreign, race, social, other, and multiple.
Not all items that showed a clear policy implication belonged self-evidently to one of our
four major domains, hence the additional category ‘other’. By contrast, some items were
written in such a way as to tap more than one of our four great policy domains, hence the
category ‘multiple’.
Issue items that unambiguously referred to welfare, foreign, race, or social policy
were then subjected to separate factor analyses—principal component analyses—year by
year. (Kim and Mueller 1978a, Kim and Mueller 1978b; for greater detail on this
specific analysis, see Claggett and Shafer 2001.) Note that it was necessary to proceed
by individual policy domains because an analysis of all policy items together, especially
early in the series when there are so few items in total, could create components that
reflected nothing more than this paucity of items. On the other hand, there seemed no