Cronbach's alpha and research anxiety

As I was going through Luker (2008) today wondering what I might write about for my blog post, I hit upon a completely new concept: Cronbach's alpha, a coefficient of internal consistency. As Luker explains, once you've coded your data and created a coding book (a list of what codes relate to what themes and sub-themes), you ask someone who doesn't know your research to use it to code a sample of your data. Then, you run the test on the resulting analysis, along with your own analysis of the same data, and have a measure of the extent to which you have been consistent in coding.

Because I'm not a mathematician, I imagine this could be done with SPSS. In fact, in an FAQ document, SPSS shows the formula "for conceptual purposes" before giving more familiar (to me) screen shots of what this would look like in the program.

I liked Luker's coding by hand methodology, but wonder how long it would take to input the data for the blind-coded sample into SPSS in order to run the test. It might be easier to sit down with a "for dummies" book and learn how to do it myself.

What attracted me to this method is not so much, as Luker posits, that it would legitimize my research to canonical social scientists, but rather that it would keep me accountable and disciplined during the analysis process. Working towards a logical, clear coding book would keep my thoughts in order and seeing the coefficient (which we hope, would be high) might relieve my own anxieties about the research process.

In fact, I've been having some anxieties about my research, as I've been working on my research proposal. I started the blog post with a remedy for analysis anxiety, but there is also data collection anxiety. I have an exciting case study in mind. How to I ensure that I gain entry to my research participants? Walsham (2006), in a paper about interpretive research methods for the study of information systems, devotes a section to the social skills of researchers and places importance on being liked and respected by the participants. Similarly, Luker discusses the need to make your research relevant to the participants and showing reciprocity. When she gained access to an abortion clinic that she wanted to study, she donated blood in exchange. It sounds like a good idea, but I'm not sure I have enough blood for the number of interviews I would like to do. I imagine that what Walsham and Luker describe are simply normal social relations - why would our relationships and interactions with other human beings be any different because we are doing research? In any case, I can see why Luker discusses researcher anxiety near the end of her book. During my research journey, I will want to test my process many, many times, primarily to reassure myself that my work is sound.

No comments:

Post a Comment