Bringing together multiple qualitative data sets a breadthanddepth

Bringing together multiple qualitative data sets: a breadth-and-depth approach to analysis Susie Weller*, Emma Davidson**, Lynn Jamieson**, and Rosalind Edwards* * University of Southampton ** University of Edinburgh

STEP 1 Overviewing the archived qualitative data and constructing a corpus Stages 1. Data audit 2. Data management 3. Data storage

STEP 1 Audit

STEP 1 Management – file harmonisation

STEP 1 Management – file harmonisation

STEP 1 Management – file reorganisation

STEP 2 surface mapping Assembled data sets/corpus already informed by interest in: • the substantive topic; • honouring principles of qualitative methods; • advancing theory and strengthen possibilities for generalising; • research questions and rationale for comparison

STEP 2 surface mapping • Basic so-called ‘text mining’ starter techniques – – – Bags of words, bags of tokens Counting frequencies Concordance Co-location, proximity, clusters Measuring ‘keyness’

Some of many softwares • Ant. Conc http: //www. laurenceanthony. net/software/antconc • Wordsmith http: //www. lexically. net/wordsmith/ • Wmatrix http: //ucrel. lancs. ac. uk/wmatrix/; http: //ucrel. lancs. ac. uk/usas/ • open source programing language R and Python IRAMUTEQ. http: //www. iramuteq. org/ • Leximancer https: //info. leximancer. com/

Keyword Analysis love All other words Total Male 414 1714029 1714443 Female 1214 2592238 2593452 Total 1628 4306267 4307895 • Expected frequency: Row total times Column total divided by the total number of words in the corpus. • Plug into this equation: • Comparison between the observed frequency and expected frequency of the word in question and all other words.

STEP 3 Preliminary analysis ‘theme’ Thematic mapping ‘theme’ Where to ‘dig deeper’? • Select ‘theme’ for further examination • Read sample extracts containing theme of interest • Aim of justifying further analysis Our approach • Mindfulness of context • Distracted by detail • Sampling logic • Recursive process

STEP 3 Preliminary analysis – keyword ‘love’ Pre-1950 s Men (total cases 10) Sample where keyword ‘love’ identified: 7 Pre 1950 s_Men\P 7_W 2_ASINT 2 - 1 reference coded I So what’s it like this area that you live in, is it a friendly area? BS Its alright. AS They all mind their own= BS As you know we came from <Southern town>, I love <Southern Town>, but I have never settled here, not really. Its erm, it’s the estate I don’t particularly like. They all keep themselves to themselves really, you don’t see them anytime. But they are quite friendly, very friendly some of them. We go over on a Monday and make their tea, well Mary makes their tea, because when they’re coming frae their work because Julie is at the college, but this is a break or something, ken, but she has got a job with BT where her mum works, and erm - when they come home Mary has their tea all ready, and we go over Thursday and again she makes their tea, and they love it. They love her coming in, because she cleans up the house and makes their beds.

STEP 4 Depth Analysis In-depth qualitative analysis • Immersion in data that is sensitive to context and multi-layered complexity • Focus on rich detail to represent intricate social realities and produce nuanced social explanations Examples • • • Thematic analysis – recurrent themes in the data Framework analysis – commonalities and differences across and within cases Grounded analysis – inducting meaning Narrative analysis – construction and sequencing of stories Conversation analysis – procedures used to communicate Discourse analysis – mapping ways of knowing

CONCLUSION

Publication from the project • Davidson, E. , R. Edwards, L. Jamieson & S. Weller (2018) Big data, qualitative style: a breadth-and-depth method for working with large amounts of secondary qualitative data. Quality & Quantity, 1 -14.