Thursday, April 21, 2016

Taking on a new module #255

This past semester I had five classes:
  • Business Systems Analysis
  • Managerial Foundations of Information Systems
  • Business Data Analysis
  • Advanced Business Data Analysis (2 separate classes)
This is one module more than I normally am required to do and represents an average of 15 hours class contact per week, which is low by third-level standards outside the Universities (our normal "load" is 12 hours/week). The first three above I have been teaching for several years, and did not require much preparation on my part - this is a dream for many academics, but it is boring. However, the Advanced Business Data Analysis module was a new challenge and it took me more time than I have ever spent preparing classes to get ready for this module. A colleague who had taught it before kindly lent me his notes and module resources, but I much prefer to create my own notes so I set about doing this. 

I wanted to make the new module as practical as possible, this meant preparing examples in Excel, SPSS and in the R programming language. I am very used to Excel and SPSS, but had limited experience with R. Nevertheless, it was fun getting to grips with R and learning how to use its power. I got a great buzz when both I and my students in the class were able to carry out exercises on the same data with all three tools. 

Despite the word "Advanced" appearing in the tile of the module, most of the statistical tests used (ANOVA, Tukey, Mann-Whitney, Kruskal-Wallis, Chi, Shapiro-Wilks) were not that advanced. Given that this was the second term most students were able to pick up these tests quite easily. Some subjects such as power, confidence intervals, and effect size were a little more complicated. Factor Analysis is advanced and we finished up with this. 

It's hard to be 100% right at any time - even when teaching a module for several years it often happens that I discover an error in my notes or an exercise that should be different. We mostly operate at confidence levels of 95% in Statistics, so we're not perfect! When delivering a new module, there is obviously a higher risk of making an error, and of being unable to answer questions as they arise. Experience overcomes this, but that will now have to wait until next year.

No comments:

Post a Comment