Histories | Data Re-analysis | Eco-Systems
The re-analysis of existing large-scale data on participation
Our work on data re-analysis considers how survey methods and the ‘absences’ in cultural datasets affect our understanding of participation. We are demonstrating new ways to work with new and existing data to improve and broaden the understanding cultural practices in the UK.
Existing participation datasets do not directly carry many data on activities, tastes and lifestyles that fall outside of the established cultural canon. This is a telling reflection of how the survey method itself is implicated in our understandings of what it means to participate. By reducing the cultural field to a partial and specific set of measurable indicators, surveys both represent and help to reinforce particular ways of ‘seeing’ participation (Law 2009, Law et al 2011).
This is not to say that survey data have nothing to offer. On the contrary, it is important to examine the ways in which the ‘absences’ in cultural datasets frame our understandings of participation, but also to look for ways in which they might reveal more about informal and everyday practices than they set out to do. Here a broader range of more sophisticated quantitative methods needs to be brought into play to help unpack the findings from standard statistical approaches.
The aims of UEP’s Data re-analysis work is:
- to distil findings about pattern of participation in informal and everyday practices from existing surveys and datasets
- to explore what existing surveys of both formal and other cultural practices reveal about participation at the sub-national level
- to refresh, develop and refine survey approaches to collecting and analysing participation data
We have been pursuing these aims in several ways. Up to now our research in this Work Package has mostly focused on Taking Part. The National Survey of Culture, Leisure and Sport (TP). Here we have created a new approach to segmenting lifestyles using Multiple Factor Analysis, combining indicators across the complete range of culture and sport variables together with the Survey’s rarely analysed ‘free time’ fields.
This work has been being supplemented by (a) the construction of a literature review detailing studies and publications that have so far used the Taking Part data; (b) a study of process documentation and variable mapping; and (c), the development of new research questions and analytical/modeling approaches. In the latter respect, we have recently been developing new approaches to mapping the ‘space of lifestyles’ from the TP data, using a refinement of Multiple Correspondence Analysis known as Multiple Factor Analysis.
- Adrian’s explanation of Multiple Factor Analysis
- Multiple Factor Analysis presentation
Our attempts to distil representative local level patterns of participation in TP have been less successful. Although we were granted access to a ‘rolled up’ version of the data containing all waves carried out between the survey’s inception in 2005 and the beginning of 2012, we found that survey respondent numbers were still too small locally to pursue viable area level analysis. However we have been attempting to integrate findings from analysis as Government Region level with our case study work.
We are now moving on within our data re-analysis work to a third type of analysis, which looks into how types and patterns of participation are related to biographical time and the life course trajectories. For this we are using data from two of the British Cohort Studies: BCS70 and the (1958) National Child Development Study.
John Law (2009), ‘Seeing Like a Survey’, Cultural Sociology, 3 (2)
John Law, Evelyn Ruppert, Mike Savage (2011), ‘The Double Social Life of Methods’, CRESC Working Paper, No. 95
Image: From the film, Kitchen Stories (2003), directed by Bent Hamer. Photograoh by Erik Aavatsmark