The value of data mining in music education research and some findings from its application to a study of instrumental learning during childhood

Abstract

The use of data mining for the analysis of data collected in natural settings is increasingly recognized as a legitimate mode of enquiry. This rule-inductive paradigm is an effective means of discovering relationships within large datasets — especially in research that has limited experimental design — and for the subsequent formulation of predictions and rules. The method is dramatically under-used in education research in general, and is hardly represented in music education, if at all. The present article reports on several decision trees that emerged from mining for knowledge in datasets constructed from the musical journeys, experiences and abilities of 157 young people in Australia from the outset of instrumental tuition in primary school and for the following 12 years. This article illustrates the validity of knowledge discovery in databases for forecasting outcomes and behaviours in educational settings generally and, more specifically, it considers early predictors of students’ short- and long-term commitment to instrument learning. This machine-learnt knowledge provides music educators with useful information about the relationship between various attributes of student experience. Within months of beginning instrumental tuition, reported levels of self-regulation, practice on Sundays, parental reminders and self-efficacy beliefs emerge as potentially predictive of students’ ongoing musical engagement. Findings are discussed in relation to self-regulation and motivation theories.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 29 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.