Making sense of numbers and words: Statistical methods

Peter Grimbeek

Site Notes Contact details


On this page, I provide links to a collection of papers where in each case, as one of the authors, I set out to do things differently, methodologically speaking.

I've applied SPSS Optimal Scaling to survey data gathered from Chinese migrants to Australia to demonstrate how one might emulate the French sociologist, Pierre Bourdieu, by using demographic information to good effect.

I've given an example of the use of z-scores to map the outcomes from a larger study that serves as a baseline to outcomes from a much smaller sample. In doing so, I've considered the influence of context (parents' views about their child with a disability vs. teacher views about a classroom that includes children with disabilities) on participant responses to a questionnaire about the extent to which students with disabilities (autism) benefit from regular classroom settings.

I've used the automated text analysis software, Leximancer (now up to vsn. 4), to analyse abstracts of Conference papers over a three-year period, as a way to identify shifts in topics of interest to those submitting papers.

In conjunction with a QUT student, Jocelyn Leigh, I've used an experimental design to design and administer a survey that examined the extent to which males and females agree about the likelihood of a list of life events being part of the future of a baby boy or girl (Gender in the future).

I've used Rasch analysis and factor analysis to re-examine the outcome of administering the Cognitive Holding Power (CHP) instrument in a high school setting. This instrument measures two levels of cognitive functioning: Procedural knowledge vs. problem-solving. A methodological interest here is in comparing the outcome of appplying the Rasch analytic aim of minimising disorder in the rated difficulty of responses (e.g., where agreement is rated as a more 'difficult' response than strong agreement) by collapsing across response categories vs. the parametric aim of enhancing reliability by maximising the number of response categories.

In conjunction with a colleague, Steven Nisbet (now retired), who developed and administered the survey, I've also used Rasch, exploratory, and Confirmatory factor analysis to examine patterns in teacher responses related to the benefits of state-based numeracy testing (Year 3, 5, 7).

Peter Grimbeek