Siegling, AB;
(2015)
Facet Benchmarking: A psychometric method for refining multi-faceted assessment instruments.
Doctoral thesis , UCL (University College London).
Preview |
Text
Siegling, AB PhD.pdf Download (2MB) | Preview |
Abstract
Driven by the challenge of defining and measuring psychological attributes, this thesis advances an instrument refinement method aimed at identifying “problem” facets detrimental to construct validity. The method, labelled “Facet Benchmarking” (FB), integrates theoretical and empirical steps and is intended to supplement established scale construction approaches; it is part of the wider construct validation paradigm. FB seeks to detect redundant and extraneous facets based on their inability to occupy a unique part of the variance attributed to a given construct. An alternative, more objectively derived representation of the construct is used to assess if the hypothetical facets of a given measure fulfil this general criterion. That representation is a composite extracted from systematically selected criteria, or outcomes, of the construct. In this thesis, FB is examined across three investigations (three chapters) of increasing rigour, each involving a different construct and data from multiple samples. The first application of FB (Chapter 2) was based on existing data, gathered in previous validation studies of the Trait Emotional Intelligence Questionnaire. Chapter 3 lays the psychometric groundwork for the second application of FB, addressing issues of homogeneity and dimensionality of the relevant measures. These were measures of dispositional mindfulness, the construct to which FB is applied in Chapter 4. The third application of FB focuses on the General Factor of Motivation, a re-conceptualisation of motivation proposed and validated in Chapter 5 (two measures were developed and used for this purpose). The purpose of this final investigation (Chapter 6) was to assess plausible alternative explanations for the method’s efficacy: domain underrepresentation and common-method variance between facets and criteria. The results from all three investigations of FB supported the efficacy and integrity of FB. The implications are, therefore, discussed in detail in Chapter 7, along with considerations for the method’s application and future development.
Archive Staff Only
 |
View Item |
