- For evaluators’ eyes only (21/07/2018)
- Reconciliation Week and findings from an Aboriginal health evaluation (04/06/2016)
- Evaluation amidst complexity: 8 questions evaluators should ask (04/12/2015)
- To count or not to count: Australian population data (20/02/2015)
- My pick of readings on scaling up health interventions amid complexity (12/12/2014)
Love your data
Tuesday, 24th January
The typical image of a scientist is working with test tubes and microscopes in a lab. Social scientists love data too. Delving into the data is what makes my fingers tingle and my heart race. When I first confront a dataset – whether of numbers or words, hyperventilate, hardly knowing what to look at first. Gradually I focus on a little piece and bit by bit I understand what the data has to say.
I get to know the data like I know my family. Only with familiarity do patterns start to make sense. These methods work equally well if you are starting with a theory or using a “grounded” approach of building your own theory.
If I have quantitative data I start by running frequencies of all of the major variables just so I know their distribution. This is a good time to “clean” variables. Recode categories which belong together like level of finishing school. If only 10 percent of your sample ever went to university, you do not need separate categories for started but didn’t finish, graduated as an undergraduate, master’s degree and PhD.
Once you know what your sample, you want to start to look at whether certain subgroups are different. Are clients from country areas from bigger or smaller companies than clients from cities? Are their differences by gender, nationality, age?
If you think of yourself as a “qualitative” person who prefers conducting long, open-ended interviews with a handful of people to short structured surveys with 100s or 1000s of respondents, you probably think that you cannot intimately know quantitative data. But you are wrong.
I know it is old fashion but I am a big believer in personally entering some responses of surveys myself. I quickly get a feel for what are the common responses and, more important, what are the common patterns of responses. It was through hand entering that I learned that some residents of the City of Greater Geraldton feel really passionate about the quality of footpaths (sidewalks), roads and garbage pick-up. Who would have thought? Those aren’t my top concerns. But by delving into the data I got a much greater insight. I had done half of the analysis and written the first paragraph of the conclusions from only 30 minutes of data entry.
If entering data is too 20th century for you, or just not feasible, try another method to delve into your data. Take a manageable number of cases, say 12-20 respondents and look really closely at each one. Look at all of their answers. Know their age, where they live, their attitudes, their health problems, their shopping behaviour – whatever the survey is about. Take one case at a time. Think if these people are like people you know. Can you understand what compels them to answer as they do? Is it the amount of money they have available? Is it their value system developed through their age and where they live? Use what they call “socioological imagination” to picture each respondent and what drives them.
Rejoice in your data. Seek inspiration from it. The more you really know your data before you start the sophisticated analysis processes your research will be more enjoyable for you and more meaningful for others.