Some things to consider when analysing and writing up
Mood:  a-ok
Now Playing: Can't Think Straight - Gilbert O'Sullivan
Topic: Data Analysis
Well these are little juttings I made whilst doing some things - so what to make sure I keep them in consideration:

(i) I have to look up some plausible reasons for using a 5 point Likert scale versus a 7 point scale like Albritton et al. Hmm ... can't find any good papers to say why I shouldn't except one that says I should sort of use the 7 point - but they were actually trying to support their use of the 7 point scale (some people Wyrwich and Tandino). However, I could indicate that since the analysis was intended for chi-square and there were so few intended respondents, there was a higher likelihood of having empty cells and hence the reason to making it a 5 point scale. Also, we used 'fence sitting' to ensure a response and minimize the middle point as there is a tendency to use the middle point more in a 5 point scale.

(ii) Now we have a low response rate in the email questionnaire. Doug was saying that this is always true in comparison with a paper questionnaire but got to check this out. The problem with comparing with known data is that most people knew the list they were sending too - that the questionnaire had direct relevance to them - there was some uncertainty about relevance regarding my list. Anyway, found some literature that says the average response for email questionnaires is in the 20s to the 30s %. So, we'll see.

(iii) It might be also useful to compare level of courses with responses that I received for the ATI to check if there is any difference in the way the courses are delivered

(iv) I also have to consider what I would have done differently if I had to do it all over again (simple send it during the course term!!!)

(v) I'll also have to compare the distribution of the responses to the list I have to see if the absent responses were random - and how the distribution is towards discipline.

(vi) Persons might have seen the stuff on simplex algorithm, graphical solution etc. and thought that is what I meant by coverage and may have said they didn't cover the solution - not sure if that will influence the outcome.