Konvergencija sociološke i statističke metodologije - mogućnost unapređenja empirijskog istraživanja u sociologiji

Abstract

Ovim radom se nastoje sumirati prednosti koje korišćenje razvijene statističke metodologije u sociološkim istraživanjima može doneti na naučno-saznajnom planu. Osnovanost upotrebe nekih metoda multivarijacione analize pronalazi se u dugoj istoriji konvergencije dve metodologije, sociološke i statističke, koja počiva na trajnoj potrebi za društvenim istraživanjima i potrebi da se ta istraživanja sprovedu na precizan i pouzdan način. Od početka ovog veka, kada statistika i sociologija dobijaju svoj savremeni izgled, kontakti između dve nauke postaju sve intenzivniji, da bi u drugoj polovini ovog veka razvoj njihovih metodologija postao interaktivan. Ovaj napredak se u polju simultanog tretmana višestrukih merenja najjasnije ispoljava kod analize kategorijalnih varijabli bazirane na višesmernim tabelama kontigencije. Razvoj računarske tehnologije je doveo do razvoja loglinearnih i logit modela i njihove rasprostranjene upotrebe u sociološkim istraživanjima. Primena multivarijacione analize u empirijskom istraživanju uzajamnih, a posebno uzročnih povezanosti društvenih pojava dovodi do napretka sociologije na ukupnom saznajnom planu. Na ovaj način dolazi do ujednačavanja merenja i klasifikovanja i do konstituisanja formalnih standarda komunikacije na području istraživanja. Osim toga, postiže se bolja proverljivost i objektivnost novih saznanja, kao i veća preciznost. Konačno, primenom ovih metoda su olakšani dublje poniranje u suštinu pojave, ispod manifestne ravni i sinteza pojmova na višem nivou apstrakcije.

In this article, an attempt has been made to summarize the advantages the use of advanced statistical methodology in sociological research could bring on epistemological level. The foundation for the use of some multivariate methods could be found in long history of convergence between sociological and statistical methodology, which was based on permanent need for social researches as well as on need to conduct this researches in reliable and precise way. Since the beginning of this century, when both sociology and statistics came to their contemporary outlook, contact between two sciences has become more and more intensive, leading to an interactive development of their methodologies in second half of this century. In the field of simultaneous treatment of multiple measurements this progress is most obvious on the categorical variables analysis based on multiway tables of contingency. Development of computer technology led to the development of log-linear and logit models and their wide spread use in sociological research. The use of multivariate analysis in empirical research of mutual, and especially causal, relations between social phenomena brings sociology to the more developed position on general epistemological level. This way measurements and classifications could be unified, and formal standards of communication in the field of research could be constituted. Also, better verification and objectivity of new knowledge, as well as higher precision, could be achieved. Finally, deeper penetration under the manifest level, to the essence of the phenomenon, and conceptual synthesis on a higher level of abstraction could be made easier.