A mixed methods approach to the construction and validation of a personality-centred model and measure of individual differences in ways of thinking
thesisposted on 01.03.2017, 05:50 by Roodenburg, Esther May
This dissertation reports the process of creating a measure of the individual differences in the ways people think, based on a personality-centered psycho-lexical model (Roodenburg, 2006) not previously instrumentalized. In developing the scale, the same six factors and facet structure of the a-priori student ways of thinking model were maintained, and the adjectival items used to generate sentences. Rather than continuing with a quantitative method as used in the model development, and traditionally used to develop self report measures, the current research proceeded with a mixed methods approach, in order to capitalize on the potential richness of data from a number of case studies, both to deepen the understanding of how individuals think, and to alternatively validate the measure. A range of views were gathered by semi-structured interviews with ten participants, selected as representing Holland’s occupational interests (Holland, 1996), and from a mature age range of 55 – 74 years, three males and seven females. The qualitative part of the process ran concurrently with data collection from the newly developed 120 item on-line questionnaire, with both quantitative and qualitative methods asking questions specifically about thinking. Phenomenological thematic analysis was completed before any analysis of the quantitative data. Principal components analysis (PCA) was initially used in an exploratory manner, and then with Procrustes targeted rotations for a confirmatory assessment of how items fitted in relation to the original model. After item weeding, 21 of the 22 original facets appear to offer sufficiently reliable constructs, seven were highly reliable, with Cronbach alphas and IRT marginal reliabilities greater than .8; fourteen with above .7 for IRT, though slightly lower on alphas on nine facets. Cluster analysis of respondents was carried out to gain person-centred profiles. Augmented by PCA innovatively using a transformed data matrix, the various resultant typological groupings were found to explain from 89% to 97% of variance. The results from both methodological approaches were then integrated, confirming the separately considered interpretation of results, thereby providing a mixed method validation of the ways of thinking (WOT) measure. Results demonstrated that when people, not variables, were considered both quantitatively and qualitatively, two distinct typologies emerged: the Realists and the Ideaists. Realists’ thinking suggested concern with ‘the here and now’, what can be seen and heard, being primarily concerned with practical issues, and with a preferred focus on things tangible rather than the theoretical or abstract. The second group, identified as Ideaists, enjoyed thinking for its own sake, looking for meaning and understanding, with ideas rather than practicalities uppermost, and who when inspired would engage in actions driven by their well thought ideas. The study suggested the remaining individuals were either ambidextrous, with a preference for being a Realist but clearly able to use both ways of thinking, or alternatively were individuals whose preferences had remained undifferentiated. The value of the qualitative insights that distinguished between such individuals also enabled additional interpretation of the quantitative data, though ongoing and further sophisticated person-centered analytical procedures are needed to clarify definitive explanations between those identified as ambidextrous from those undifferentiated through a lack of consistent distinguishing traits.