|
|||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||
party will be low on reflection, and one might consequently expect it to be high on concentration. However, it will also be low on concentration because of its 50:50 division of Catholic and Protestant members. Furthermore, another party in the same country whose members are all Catholic would be high on reflection and very high (i.e., perfect) on concentration. These examples illustrate the point that the nature and distribution of social cleavages and groupings within a country can impact in quite reasonable but varied ways on the measures of social attraction, concentration, and reflection. Each of these measures thus seems to tap a different perspective from which to view social support, and the three measures can be regarded as alternative indicators of "diversity" in social support for four of the six cleavage dimensions, producing scale reliabilities of .83 or greater in all instances. But, for socioeconomic status and education (which might be regarded by many as another indicator of social status, especially in Third World countries), the measure of social concentration plays a different role, and the scale reliabilities of about .70 in each occur in spite of social concentration rather than because of it. For these cleavage dimensions, it appears that the social diversity--captured in the attraction and reflection measures--is not an empirical trade-off on social homogeneity--as represented by the concentration measure. Hence, the concentration measures for socioeconomic status and education cannot be counted as performing completely to predictions in the validation process, although all other sixteen measures of social support do behave as expected. Issue Orientation. Parties are frequently classified according to "issue orientation," with the most common basis for classification being some type of left-right or perhaps liberal-conservative dimension. All too often the conceptualizations of these dimensions are vague or even unstated, and party positions on a wide range of specific issues are rarely linked to the classification in a clearly explicable fashion. The approach employed in the ICPP project in measuring issue orientation was to test the unidimensionality of the left-right or liberal- conservative conceptualization of issue orientation by analyzing the parties' positions on thirteen issues selected for their relevance to cross national politics. In keeping with conventional political discourse, parties were judged in varying degrees as being "for" or "against" each issue. To help our test of unidimensionality, we attached a left-right interpretation to the parties' positions at the time of scoring, adopting the convention that positive scores were associated with leftist positions--usually interpreted as favoring greater governmental activity in the issue area. Negative scores obviously signaled rightist positions in the sense of less government activity. When degree of "government activity" was inapplicable to the issue, popular notions of left and right positions were used instead. The thirteen issues involved in the scoring and the positive or leftist positions are as follows:
The empirical findings for the intercorrelations among these issues reveal clear patterns of relationships, but not the simplified unidimensionality that served as a foil for scoring purposes. In this instance, factor analysis is especially useful for revealing the underlying relation- |
|||||||||||||||||||||||||||