A conceptual framework for the use of the generalised latent variable model in psychological research

Through the 20th century to today, latent variable modelling remains predominant as a psychometric technique employed in psychology research practices. Through time it has also become increasingly complex, and dependent upon pre-specified parameters and assumptions coded in computer software, which may remain unreviewed by a researcher. With these trends, there has been a loss of focus on important mathematical and statistical questions for the latent variable model. Further, we remain without a unified solution for presentation of extra-statistical assumptions necessary for use of the model, where it remains the case that decisions researchers must make when using the latent variable model remain unrecorded and perhaps, unreconciled to anything more than researcher intuition. The primary aims of this thesis were to firstly, to document a cognitive-historical analysis of the methodological problems for the generalised latent variable model; and secondly, to use this analysis as a guide to develop a conceptual framework for the use of the latent variable model in psychological research. Contemporaneous developments in philosophy of mathematics and logic were utilised to ensure the proposed structure for the framework was levelled on grounds that facilitate logical consistency and unification, given that diverse methodology accompanies the assembly of data, variables, relations and models of phenomena, in latent variable modelling. Such unification was shown to facilitate assessment of quality uncertainty in psychology study outcomes that make use of latent variable models; facilitating for perhaps the first-time true study comparability and replication assessment, as well as facilitating robust integration in meta-analyses. The principal finding of this thesis was the essential role for a realist constructivist conceptual framework for research projects that make use of the generalised latent variable model, in psychological research.