Abstract
This article presents a typology of designs for social research in the built environment. Currently there is no such typology, while the notion of ‘research design’ is less known in the built environment compared to the social sciences. Twenty-five subtypes are identified and clustered into 10 prototypical designs, namely: (1) surveys, (2) experiments, (3) modelling, simulation, mapping and visualisation, (4) textual and narrative studies, (5) field studies, (6) case studies, (7) intervention research, (8) evaluation research, (9) participatory action research and (10) meta-research. After determining the extent to which these designs feature in actual studies, the designs are classified according to six design considerations, including research – context, aim and purpose, methodological – paradigm and approach, and source of data. The typology contributes towards greater clarity in terms of ‘research design’, improved teaching of research methodology and greater methodological coherence in projects.
Notes
1. Blaikie (Citation1993, p. 7) defines ‘research methodology’ as ‘the analysis of how research should or does proceed. It includes discussions of how theories are generated and tested – what kind of logic is used, what criteria they have to satisfy’, thus, the meta-study of designs and methods. On a practical level, ‘research design’ refers to the overall plan for conducting research (a more detailed definition follows later). ‘Research methods’ refers to techniques or procedures within a research design to collect, analyse, and interpret data (Bryman & Teevan, Citation2005, p. 24; Creswell, Citation2009, p. 15; De Vaus, Citation2001, p. 9).
2. See Du Toit (Citation2010a, pp. 125–128) for a more detailed outline of the 10 prototypical designs with their subtypes, specialised subtypes and areas of application in built environment research and practice.
3. See Du Toit (Citation2010a, pp. 131–137) for a more detailed description of the selection procedure.
4. This time span was chosen for three reasons: (1) JUD was first published in 1996 only; (2) the 10-year span ensured an adequate data-set for analytical purposes and (3) there was a sufficient gap between 2005 and the time of conducting the review in 2009 to examine the scholarly impact (citations) of articles published as late as 2005.
5. Because earlier articles have had more time to be cited compared to more recent articles, citations of more recent articles were weighed incrementally over time to make them comparable to citations of articles in a base period. See Du Toit (Citation2010a, pp. 144–145, 166–170) for a more detailed description of the weighing procedure.
6. Statistical significance was calculated at the 95% confidence level. See Du Toit (Citation2010a, pp. 185–194) for detailed results from Chi-square tests and measures of association.