Dear All,
I have been referred to this list by a colleague, he suggested that someone
may have an answer for me. I have completed a biomechanics study and have
collected knee flexion data over a set period of time at the same point in
the gait cycle for each subject, for four subjects across 5 conditions each
condition having nine trials. The conditions differ by the amount of load
carried by the subject, the load amount is set across the study and
therefore represents a different percentage of each subjects bodyweight.
I am endeavouring to work out the best method to assess the differing
condition effects within and across the subjects, as well as controlling
for the load as a percentage of the subjects bodyweight, i.e. use it as a
covariate (Cov). I am using statistica and have proposed the following
exprimental design.
|Trial 1 |Trial 2.......Trial 9
|subject|condition |CoV |time 1|time 2 ...time 50|
| | | | | |
| A | 1 | 20 |......|......
| .. | .. |.. |......|......
| A | 5 | 45 |......|......
| B | 1 | 20 |......|......
| .. | .. |.. |......|......
| B | 5 | 45 |......|......
| C | 1 | 20 |......|......
| .. | .. |.. |......|......
| C | 5 | 45 |......|......
| D | 1 | 20 |......|......
| .. | .. |.. |......|......
| D | 5 | 45 |......|......
I am therefore using the following test, 4 way ancova, with 1 covariate,
testing between subjects and conditions, testing within trials; with the 5
conditions nested within the subject. The plan would the be to carry out a
post hoc tukey test is any significances were found.
I am not sure that this is perhaps the most appropriate test because the
data is a time series, I am therefore wondering if anyone has any better /
more statistically valid methods of assessing my data. And if so please
could you let me know. I have seen cannonical correlation applied to this
scenario but not being a statistician the concept seemed difficult to
grasp. I would greatly appreciate any advice is layman's terms.
yours sincerely
David Davis
----------------------------------------------------------------------------
--------
Researcher in Biomechanics and Load Carriage
Department of Human Sciences
Loughborough University
Leics.
LE11 3TU
UK
Tel: (UK) +1509 223086
Fax: (UK) +1509 223941
Email: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|