Is using T-test correct for my approach?

#1
Hi everyone I am having some trouble as I am unsure which inferential statistic to use.

I have a table of data (excel) that have several columns as follows (all are in numeric values):
  • Subject id
  • Gender
  • Assessment mark (marks from their last exam/ assessment )
  • Hours usage (how many hours they use lms per week for the unit)
  • System benefits (the reliability of lms)
  • Satisfaction score (user attitude towards lms when they use it to learn for a particular unit)
  • Quality score (information that lms produces eg. Timetable, format)
  • Support quality (the amount of support provided for lms usage)
These columns are the key variables that were presented to a group of 200 students in the form of questionnaire that collates their use and percepons to the Learning Management system (LMS)

I am trying to do a compare or gather results on how these variables influence the success of this LMS system.

Not sure if I am on the right track, but I thought of using the Assessment Mark as the main variable against other variables such as system benefits, satisfaction score using inferential statistics t test. Will that be right?
 

Karabiner

TS Contributor
#2
If you want to analyse the relationship between assessment mark and characteristics of LMS use,
then t-test (or better Welch-t-test) can be used. In some cases, though, it is not clear whether
the variable might be ordinal rather than interval scaled (e.g. "satisfaction score"). If there's an
ordinal variable, you could use Mann-Whitney U-test for the analysis.

If you want to predict assessement mark by several LMS characteristics at the same time, then
mutiple binary logistic regression could be an option,

With kind regards

Karabiner
 
#3
If you want to analyse the relationship between assessment mark and characteristics of LMS use,
then t-test (or better Welch-t-test) can be used. In some cases, though, it is not clear whether
the variable might be ordinal rather than interval scaled (e.g. "satisfaction score"). If there's an
ordinal variable, you could use Mann-Whitney U-test for the analysis.

If you want to predict assessement mark by several LMS characteristics at the same time, then
mutiple binary logistic regression could be an option,

With kind regards

Karabiner
Hi thank you for getting back to me.

since I am not trying to predict assessment mark, I suppose the multiple logistic logical regression as you have mentioned is not for me then.

I suppose the variables are all ordinal values (see attached, I just started learning statistics and so, pardon me if I got this wrong)

PNG image.png