# Which stats method to use? HELP stats rookie who just wants to graduate

#### bulletproofpineapples

##### New Member
i am writing a proposal. i just need to mention my stats methods in a few lines but i'm not sure which is right.

it is a study with 4 groups + 1 control categorised by different stages of a disease. all 4 groups are given one kind of intervention (excluding the control).
at baseline, each group had 3 outcome measures eg (pain, distance and speed ) assessed and will be reevaluated at end of the 8 weeks of treatment.

what do i use?
t-test
anova
manova

#### Karabiner

##### TS Contributor
You did not mention sample size and you did not explain the measurement level of the outcome measures.
Supposing that sample size is large enough and that outcome is measured on an interval scale, a mixed
analysis of variance (repeated-measures analysis of variance wth an additional grouping factor) could be
an option.

With kind regards

Karabiner

#### bulletproofpineapples

##### New Member
You did not mention sample size and you did not explain the measurement level of the outcome measures.
Supposing that sample size is large enough and that outcome is measured on an interval scale, a mixed
analysis of variance (repeated-measures analysis of variance wth an additional grouping factor) could be
an option.

With kind regards

Karabiner

sample size aims to be 60, with 10 per group including drop outs. measurement level of 2 of the outcome measures are ratio and 1 of the outcome measure is ordinal

#### Karabiner

##### TS Contributor
For the ordinal scaled variable, a simple solution could be the calculation of a new
variable (improved/unchanged/unimproved) and use a Chi2 test, but n=10 per goup
is too small for that. Maybe there's an exact test for this. Another approach would be
ordinal logistic regression of post-score on group and baseline score.

With kind regards

Karabiner