네트워크 분석을 활용한 작문 연구 동향 분석

Abstract

This paper aims to use network analysis on the intellectual structure of domestic writing research trends. Existing studies on writing research trends have made an interpretation by researchers’ dispositional judgement. This study attempts to offer a more objective, quantitative analysis. For this purpose, 293 articles on writing research from 4 journals from 2009 to 2014 were collected and their keywords extracted to conduct a frequency analysis and a co-word analysis. Their relationships were then analyzed by the algorithms of Pathfinder Networks (PFNets) and Parallel Nearest Neighbor Clustering (PNNC). A comparison between Triangle Betweenness Centrality and Nearest Neighbor Centrality also analyzed their global and local influences. Our proposals for domestic writing research in terms of the analysis and its result are as follows: first, “writing evaluation” has to be combined with a greater variety of writing research subdomains; second, various levels of “curriculum” apart from the national level need to be more considered in the domain of writing research; third, despite quantitative and qualitative growth, subdomains of writing research require more consideration with regard to the conceptual flow of the whole map as they tend to be disconnected with each other.