This city level study forms part of an investigation of the geography of religious affiliation in Scotland at the advent of the twenty-first century. Using a multivariate statistical-cartographic approach the research seeks to identify the geographical distribution of the major religions in the Glasgow region, and explain the key socio-cultural processes underlying the revealed patterns. The discussion is organised into two parts. In the first a number of important methodological questions are addressed in relation to the mapping of geographies of religious affiliation in Scotland. In the second part univariate and multivariate statistical techniques are employed to map the geography of religious affiliation within the Glasgow city region. This empirical analysis employs univariate mapping of religious affiliation at the postcode sector level, complemented by multivariate analysis of religious affiliation in relation to a set of social, economic and demographic variables designed to aid explanation of the patterns revealed by univariate mapping. In combination the analysis illuminates the principal dimensions of the geography of religion in Glasgow at the advent of the twenty-first century. As Pacione explained, prior to the current research being undertaken a comprehensive national scale mapping of the geography of religion in Scotland was not possible due to the absence of reliable basic information on religious adherence. Until the 2001 Census of Population the UK Census Act of 1920 excluded questions on religion from the national Census. Without access to a regular national census of religion previous attempts to map the geography of religious affiliation in the country were based on church membership statistics, the accuracy of which was often compromised by doubts over reliability of enumeration, completeness of surveys, and the continuity and comparability of definitions. Thus, in enumerating its adherents the Catholic Church has normally employed figures based on the number of baptisms, while the Episcopal Church collects data on communicants. Other churches, such as the Brethren, are opposed to the concept of counting members. In addition to problems of data reliability, the validity of membership enumeration as a reflection of the strength and social significance of a religion is also open to question since these figures do not necessarily indicate active participation in a particular church. Clearly, a measure of 'membership' can span a spectrum of meaning from passive adherence to militant participation. Statistics of religious activity are, in theory, easier to collect than those on church membership and should provide a more reliable indicator of the contemporary distribution and significance of religion. In practice, however, the value of censuses of church attendance is reduced by inconsistencies in the bases of enumeration, (for example, some churches estimate participants by counting those attending on one particular Sunday of the year), and by a reliance on self-reporting by churches. One of the first studies to employ a multivariate statistical procedure as a basis for mapping the geography of religion was undertaken by Pillsbury who used the same data set as Zelinsky but employed principal components analysis to identify five 'primary religious areas' in Pennsylvania. The rotated factor solution scores for the counties were mapped individually and a composite map of all factors was constructed by assigning each county to the category (factor) on which it loaded most highly. Nevertheless, despite the statistical sophistication of the technique principal components analysis does not eliminate the need for researcher decisions. Interpretation of the factors is not unequivocal, even with a rotated factor solution. In addition, construction of the composite map of primary religious areas was not unambiguous, especially in counties where more than a single factor loaded highly. The empirical cartographic-statistical approach to mapping geographies of religion exemplifies the important distinction between inductive and deductive reasoning. While deduction is a method of reasoning from principles by Zelinsky induction reasons from the part to the whole. Principal components analysis, (and the related statistical techniques of factor analysis and cluster analysis), follows the principle of inductive reasoning. In Pillsbury's study the inductive method begins by observing and recording specific cases, (religious membership by county), and proceeds via the ordering and classification of the data to reveal regularities, (religious regions). Multivariate statistical techniques are more able to identify complex patterns that reflect multi-denominational affiliations within an area. As the following study of Glasgow illustrates, they can also incorporate and statistically measure the associations among religion and other social, economic, cultural and demographic characteristics, thereby aiding meaningful interpretation of revealed patterns. On the other hand, as we have seen, multivariate statistical analysis does not eliminate the need for researcher decisions. Clearly, researchers employing these techniques would wish to avoid producing the kind of "dismaying hodge-podge of almost illegible maps and mindless factor analysis" identified by Sopher. In this light it is instructive to compare Shortridge's religious regionalisation of the USA (based on 1971 data and employing a multivariate statistical procedure) with that of Zelinsky. Despite the methodological differences the maps are generally similar. This is reassuring as it suggests that most of the major religious regions identified are more than subjective constructs or statistical artifacts. Having established the key conceptual and methodological issues relating to the mapping of geographies of religion we can now turn to the detailed empirical study of the geography of religious affiliation in the particular context of Glasgow.
Please Rate this Article 5 out of 54 out of 53 out of 52 out of 51 out of 5
Not yet Rated
