Post Your Answer

What could be the better way of implementing Wilcoxon signed rank test with covariate?

3 years ago in SPSS , Statistical Analysis By Akram

SPSS Statistical Analysis

I am performing a single group, double condition test in an experimental set-up for the instrumented gait analysis.

The first condition is measured after foot orthosis treatment
The second one is barefoot, without treatment analysis

Some subjects have received the treatment earlier as well for different time durations. My objective is to find out whether the time duration works as a covariate in these two conditions or not.

All Answers (3 Answers In All)

By Laura Answered 3 years ago

Hi, Akram. To answer your question, it would be easier if you tell what exactly is your question. Like what purpose is it that you need to fulfill by this analysis and what is the sample size? Have you used the distribution of the past use? If yes, then what was the response from that? Is it okay to use it as a variable?

Replied 3 years ago

By Akram

Hi Laura, thank you for your reply. The main question and purpose of the research is whether the treatment of foot orthosis is of any help in walking for the kids who walk on their toes. Considering that foot orthosis can have long-term impact and many kids have treatment earlier for different time durations ranging from 1 month to 2 years. That’s why to improve and strengthen the analysis, I am using the time duration of the treatment as variable to see how it affects the result. And as for the sample size, it is 21. To check the normal distribution, I conducted the shapiro-wilk test which showed that some variable didn’t have a normal distribution.

By Mohd. Hussain Answered 3 years ago

I don’t have much expertise and knowledge in the specifics of gait analysis but I read a few articles and studies. Which element(s) are you using as DV? Are they supported well with reasonings for using means and SD? These two questions become important as there are so many variables in this analysis.

Replied 3 years ago

By Akram

Thank you Mohd. Hussain. Actually, I am currently using spatio-temporal parameters like the step-length or the walking speed along with the kinematic and kinetic parameters like maximal angular excursion and the joint moment, respectively. If these variables provide normal distribution, then I’ll describe them with mean and SD. Otherwise, I will go with median and IQR.

Replied 3 years ago

By Mohd. Hussain

Thank you for clarifying these factors, Akram. However, I think that having multiple variables is just adding on to the original research question you have undertaken and is increasing the complexity as well. Also, I would suggest you to go through some previous studies as 21 is a small research sample and might not be able to deliver you the desired result. You can check out the article by Schulz and Grimes on multiplicity in randomized trials published in 2005. It might help you in sorting out things. I hope it helps, good luck!

By Akram Answered 3 years ago

Hi Laura,thank you for your reply. The main question and purpose of the research is whether the treatment of foot orthosis is of any help in walking for the kids who walk on their toes. Considering that foot orthosis can have long-term impact and many kids have treatment earlier for different time durations ranging from 1 month to 2 years. That’s why to improve and strengthen the analysis, I am using the time duration of the treatment as variable to see how it affects the result. And as for the sample size, it is 21. To check the normal distribution, I conducted the shapiro-wilk test which showed that some variable didn’t have a normal distribution.

View Related Questions