Over the half-way mark

I find myself staring at my computer screen often these days. In the past few months I've been entering into my computer all the survey and interview data that I collected late last year. Transcribing interviews is a tiring, long process but I have enjoyed hearing all the stories again and reflecting on the major themes emerging from them. Consequently I'm becoming excited about reaching a more creative point in the process, writing. That's where I get to weave together the rich tapestry of stories with the quantitative data collected from the surveys and make sense of it all. But before then I need to become closely acquainted with SPSS, the statistical package I'll be using to analyse my survey data. Unfortunately it won't be as simple as running an ANOVA and some t-tests.

My data is hierarchical in structure, with various levels that will influence each other unless the variance of each is properly accounted for. The levels in my data consist of three schools (A,B and C), each of which has three year groups (4,5, and 6), and each of those have two or three classes (x,y and sometimes z). You see, the children of one class might have quite different responses to the children of another class in the same year group because their teachers differ. The same goes for year level and school. So I need to use a multi-level approach to analysing my data in order to see clearly where differences lie, if there are any. Perhaps, as one might hope, the average attitude of children in each of the classes and year groups in a school that has adopted a whole-school integrated sustainability program will be very positive in response to a question such as "These environmental experiences are important to me". If so that's a great measure of how well the school is motivating and involving children in the school's sustainability initiatives, but I think it's fair to admit that such a finding might be unlikely. For example there may be one or two classes out of the six or so classes that were surveyed whose teachers just aren't that interested in environmental topics. As such one might expect that the average attitude of children in their class to the question stated above would be less than the average attitude of children in other classes (not to mention year groups). Of course, this is where the story will become more complex and this is where my qualitative data collected through interviews and open-answer questions in the surveys will come in handy :)

That's enough talk about data analysis for today I think. Instead it's time for me to go do some! ;)