Thursday, February 26, 2015

Assigning numerical values

Despite setting a person record in walking yesterday evening (5.24km, 6.84 km/h, 435 calories), I didn't sleep very well last night. One might think that the exercise tires me out which would enable me to sleep better, but I believe that the fact that I walked so fast and that I didn't sleep well are connected to a third factor - my mind was working overtime. This helps when I'm walking as I can ignore certain physical aspects and so walk faster, but it doesn't help when I'm trying to relax.

Over the past few days, I have begun to receive questionnaires back from the company where the pilot study is taking place. These questionnaires are very useful as I can see a few problems which still have to be addressed. Aside from that, I have also been working on the program to store each recipient's answers in a form suitable for future analysis. Once I completed that task, I started on another task which I have purposely ignored until now: assigning numerical values to the various sections.

For some sections (e.g. age, department, gender), this is trivial. Similarly, for the sections which are built on Likert scales (I strongly disagree/I disagree/I neither agree nor disagree/I agree/I strongly agree), calculating a value is simple. The section on Priority and spreadsheet usage is critical but I think that I can manage this, by assigning weights to questions (that is, one question may be important that another so I assign it a higher weight). I have already changed the order of one question's options in order to improve the calculation and I may have to change the option order in other questions as well. Otherwise I will have to assign weights to the various options which seems (at the moment) like overkill.

There may or may not be a problem with the section on spreadsheet competency. My original conception was that six questions would be presented; each question has five options, of which one is correct, three are incorrect and one is "I don't know". The competency score would simply be the number of correct answers. Whilst this still stands, I wondered whether I should differentiate between supplying a wrong answer and admitting that one doesn't know. I tried finding some information about this via Google Scholar last night but got nowhere. I have asked my advisor about this point.

At the moment, I need to spend a few concentrated hours going over the questionnaire. Hopefully I will find that time on Saturday.

[SO: 3846; 3,15,36]

No comments: