ABSTRACT

A novel approach to integrating Cluster Analysis (CA) within qualitative inquiry is presented, grounded in a large, unstructured dataset from open and rather unstructured interviews. This dataset was previously subjected to typical (theory sensitive) thematic analyses. Transformed into quantitative binary matrix structures, the CA offers robustness and transparency as it systematically exhausts the whole dataset in a replicable procedure. However, then the transformation becomes bi-directional, as resulting clusters provoke new qualitative interpretations and even further quantitative analyses. This approach led to theoretically interpretable results that significantly extended previous understandings of relations between ‘values’ and ‘learning approach’ relating to mathematics learner identity. This integrated methodology is evaluated for its significance to the substantive field, but is discussed more widely for social science research drawing on such interview datasets in general.

摘要

提出了一种在定性查询中整合聚类分析（CA）的新颖方法，其基础是来自大型的，非结构化的数据集，这些数据来自公开且非结构化的访谈。该数据集以前曾进行过典型的（对理论敏感的）主题分析。转换为定量二进制矩阵结构后，CA可以在可复制过程中系统地耗尽整个数据集，从而提供了鲁棒性和透明性。但是，由于生成的聚类激发了新的定性解释甚至是进一步的定量分析，因此转换变成了双向的。这种方法产生了理论上可以解释的结果，从而大大扩展了先前对与数学学习者身份有关的“价值”和“学习方法”之间关系的理解。