Assessing goodness-of-fit measures for multidimensional personal epistemiology survey instruments: An alternate approach using multitrait-multimethod confirmatory factor analysis
Harbaugh, A.G. (2011) Assessing goodness-of-fit measures for multidimensional personal epistemiology survey instruments: An alternate approach using multitrait-multimethod confirmatory factor analysis. In: AARE 2011, 27 November - 1 December 2011, Hotel Grand Chancellor, Hobart
*No subscription required
Survey instruments measuring multidimensional attributes of individuals+C personal epistemology beliefs have not demonstrated strong goodness - of - fit between the hypothesized model and collected data. Despite this, the growing body of research in personal epistemology suggests the existence of multiple dimensions of epistemic beliefs. To date, previous surveys have been examined using confirmatory factor analysis (CFA) in which the dimensions of beliefs have been the levels of a single factor examined. The dimensional structure explored in this study includes certainty, structure, authority, innate ability and acquisition speed. Alternatively, the multitrait - multimethod (MTMM) CFA introduces a second factor into the model. In this context, the second factor was a 1Cmethod 1D effect introduced by the philosophical classification of the questions/items in the survey. This project demonstrated that the MTMM CFA is the better statistical tool to confirm the presence of five distinct dimensions in a traditional epistemic belief survey. This was accomplished by modelling the influence of the epistemic belief dimensions crossed with the philosophical nature of the items in the survey. Data were collected over two academic terms in various levels of mathematics courses i n a community college in the Pacific Northwest region of the USA. Items in the survey were selected to measure the five epistemic dimensions listed above. Items were then classified as axiomatic, ontologic, deontologic or procedural in nature (deontologica l items were excluded). Results statistically indicated that the MTMM CFA model was the strongest model compared to standard (single factor) CFA models or alternate conceptualizations of the (dual factor) MTMM CFA model. In addition, results indicate very strong goodness - of - fit measures for the MTMM CFA model (CFI = 0.94; RMSEA = 0.038). Epistemic belief profiles for students are various levels of mathematics instruction in the tertiary environment were also explored.
|Publication Type:||Conference Paper|
|Murdoch Affiliation:||School of Education|
|Item Control Page|
Downloads per month over past year