What do I use
Institutional performance and educational decisions
Universities, polytechnics, colleges and school systems making decisions about programmes, staffing, admissions and quality.
Your question, and what answers it
Set it up once as a monitor and it re-runs itself. Most month-to-month movement in a pass rate is ordinary variation, and acting on it makes things less stable rather than more.
In ZStat: open your study, go to Analysis, and choose control chart.
With a post-hoc test, and with intake quality considered before anybody is held responsible for a difference.
In ZStat: open your study, go to Analysis, and choose one way anova.
With the base rate reported beside the accuracy, because a model that always predicts "yes" looks excellent when most students do graduate.
In ZStat: open your study, go to Analysis, and choose logistic regression.
The question every admissions policy assumes an answer to. Multicollinearity is checked, because entry score and prior grade usually carry the same information.
In ZStat: open your study, go to Analysis, and choose ols regression.
Yes. Treating students as independent when they share a lecturer inflates every statistic in the analysis.
In ZStat: open your study, go to Analysis, and choose multilevel mixed.
Time-to-completion, including the students who have not finished yet - dropping them is how completion time gets underestimated.
In ZStat: open your study, go to Analysis, and choose kaplan meier.
Which items discriminate between strong and weak students, and which everybody passes.
In ZStat: open your study, go to Analysis, and choose irt 2pl.
The same students measured before and after, with the honest caveat that anything else that changed in the same session is confounded with it.
In ZStat: open your study, go to Analysis, and choose repeated measures anova.
What goes wrong in institutional performance and educational decisions
- Ranking departments on raw outcomes without intake quality. The department admitting the weakest entrants will always appear worst, and the ranking then punishes the work you most want done.
- Reacting to one month above or below target. Set up a monitor and look at twelve points, not two.
- Analysing at the student level an intervention that was applied to whole classes.
- Treating student records collected for administration as free to reuse for research. The purpose has changed, which is precisely what the law is about.