Hi,

I am currently doing data analysis and it has been suggested that I
look at meaningful inferences. I have read up on this area through a
number of Will Hopkins articles and downloaded the spreadsheet
(xcontrial), that uses raw data values, rather than the p value
spreadsheet.

However, I am finding it difficult to select a value for 'the
smallest practically important change'. Through some reading other
authors have either asked elite coaches, selected it through previous
research or used the mean divided by the between subject SD in the
pre-test.

I am looking at the difference between 4 groups (including a control
group) in the change of a number of variables over 4 days. Variables
include peak torque, total work, DOMS and CK.

I was wondering if anyone has any information on selecting the
smallest practically important change for the variables I have
mentioned or if I should use the latter way in the previous
paragraph? Should this value be greater than the error of
measurement? Further to this, has anyone used this data analysis for
so many groups and time-points?

Regards
Emma Cockburn