Towards a hierarchical and multifaceted model for the measurement of academic self-concept in science

G. Hardy

Research output: ThesisMaster's Thesis


Recent research into academic self-concept has included investigations into domain-specific self-concepts. Examples include Lau et al (1999) within English, Marsh et al. (1997) within physical education and Vispoel (2003) within Music. They have all indicated these subjects to be multidimensional in nature, consisting of distinct sub-domains. This is an important finding and helps teachers and researchers to understand how pupils feel about themselves as learners. In contrast there have been few, if any, studies about this within the subject area of science. Up to now self-concept in science has been conceived as a uni-dimensional construct.Using structural equation modeling this study explored the multidimensional and hierarchical nature of self-concept in science. The outcomes show that science self-concept of secondary aged pupils is heterogeneous in nature and presents a consistent, stable and valid set of measures for the ways in which school pupils feel about themselves when learning science. It argues that learners have a multidimensional self-concept ‘profile’ which represents their psychological response to being a learner of science. An instrument has been developed and validated for the measurement of science self-concept for secondary aged pupils (11 – 16 years). Carrying out model fit analysis using LISREL 8, the instrument has been shown to be extremely robust in measures of fit and construct validity, and has also shown itself to be invariant across sex and age subgroups
Original languageEnglish
Awarding Institution
  • University of Durham
Place of PublicationDurham, UK
Publication statusPublished - Apr 2007


  • Science education, academic self-concept, confirmatory factor analysis


Dive into the research topics of 'Towards a hierarchical and multifaceted model for the measurement of academic self-concept in science'. Together they form a unique fingerprint.

Cite this