i need analyze specific set of features related repeated events.
imagine following dataset having records flu during year.
- for each medical check taken on patient, doctor may find if patient flu (1) or not (0).
- multiple check have been taken on year multiple record related same patient present (and multiple flus can found too).
- the whole feature set composed static categorical value: age, sex, occupation, address etc doesn't change on year.
i want determine feature may candidate describe patient flu.
i'd use statistical test following problems:
- i discarded chi square because of repeated events on same patient , violating assumptions.
- i avoided mcnemar since of independent variables multilevel (e.g. occupation not dichotomous).
- i discarded cochran's q test trying set following setup stucked on geographical information: suppose patient share address information. in cochran's q test, if use address "treatment", patient "block" , can set 1 record (1 if specific patient got flu in year, 0 if person didn't flu in year).
with considered, seems me cannot infer using setup.
so there other statistics can use on repeated paired data categorical multilevel ivs , 1 categorical dichotomous dv?
thank you
Comments
Post a Comment