Quality of life: Validation of an instrument and analysis of relationships between domains

Abstract

Quality of life (QoL) is now widely recognised as a multidimensional concept. This study validates an instrument to measure multidimensional QoL, and investigates the relationships between the domains thereof. The domains analysed are: health, housing and infrastructure, socioeconomic status, social relationships, governance and safety. We utilise a rich household-level dataset collected by the GCRO on QoL in the Gauteng city-region of South Africa. To validate the instrument, we subject the different domains to exploratory factor analysis (EFA) to confirm the unidimensionality of each dimension, and then employ confirmatory factor analysis (CFA) to validate the instrument and to investigate the interrelationships between the domains of QoL. The results show that the items included to measure the different scales/domains of QoL are valid measures of the constructs. Furthermore, all domains are positively related, implying that policy measures implemented to improve a specific dimension may also positively influence the other domains of QoL.

Construction of a Reflective Quality of Life Index for Gauteng Province in South Africa

“…Having established their hypothesesbased on the prior literature or intuition-scientists deduce predictions that are testable against observable evidence. In this case, as we saw for Greyling and Tregenna (2020), the a priori configuration of six dimensions and their constituent indicators simultaneously provided the overall hypothesis and its testable implications. After the indicator assignments were separately tested for unidimensionality with EFA, the fit or otherwise of the overall CFA against the data, using standard criteria, provided the test of the configuration.…”

“…This seemingly minor conceptual attenuation has methodological and empirical implications. We find empirically that our approach, within a robust model, results in a seventh dimension, distinct from Greyling and Tregenna's (2020) Methodologically, this points to the worth of our taking full advantage of the exploratory capacity of EFA, to both identify and populate dimensions, while drawing on the full set of the relationships in the data-the painstakingly elicited responses of some 24,800 diverse Gauteng residents to 60 questions -revail on which indicators were to be selected, and how they congregate into dimensions. We believe that this is especially appropriate in very unevenly developed lower and middle income country contexts such as South Africa, where the patterns across multiple responses, and indeed the importance attached to the questions, may differ considerably in dense informal shack settlements or deep rural villages from those in formal suburban housing, for example.…”

“…In the context of South Africa, in the latest of their three successive papers Greyling and Tregenna (2020) retained the classical approach from their earlier two papers (Greyling, 2013;Greyling and Tregenna 2017): of defining their sets of dimensions a priori, by comparison with previous international quality of life studies. The dimensions in these studies are summarised in Table 1 in Greyling and Tregenna (2020). Their chosen dimensions are housing and infrastructure, social relationships, socio-economic status (SES), health, safety, and lastly 'governance'.…”

“…Given the restricted emphasis of Greyling and Tregenna (2020) on governance, in choosing five variables for this dimension they include only one variable that, abstractly, addresses bottom-up voice and engagement: 'Politics is not a waste of time'. This is an item on political alienation drawn from a survey by Everatt and Orkin (1993).…”

“…of reflective dimensions, to weight the aggregation of their final quality of life score. And in South Africa Greyling and Tregenna (2020), considered above, moved in their successive analyses of the GCRO QoL I data from their earlier use of principal components analysis, consonant with formative measurement models, to using exploratory oblique factor analysis, i.e. presuming reflective indicators, to check the unidimensionality of their hypothesised dimensions; plus subsequent confirmation of the validity of the reflective dimensions using confirmatory factor analysis (CFA).…”