# spssforum.com

by SPSS users, for SPSS users
 It is currently Sat May 25, 2013 4:57 pm

 All times are UTC

 Page 1 of 1 [ 7 posts ]
 Print view Previous topic | Next topic
Author Message
 Post subject: which test should we usePosted: Tue Jun 19, 2007 1:32 pm
We are 2 students in the Netherlands.
We have SPSS 14.0 for windows.
We have the following variables.
Condition 1 - condition 2 - condition 3 - condition 4
The outcome of this variable is group1 - group2 or group3.
Also we have a variable called 'leglength' (outcome in cm)

We would like to know if the conditions are significantly different. Normally we would do an 'Repeated Measures' test (ANOVA) but because these condition variables are non-paremetric this isn't allowed.
Furthermore we would like to know if 'leglength' is a predictor for which group u are.

Which tests should we use?

Top

 Post subject: Posted: Tue Jun 19, 2007 5:49 pm

Joined: Tue Apr 10, 2007 8:17 am
Posts: 14
Hi Mendel & Leen,

What kind of variables do you have? Nominal, ordinal, interval, ratio? Do you have in total 8 variables (condition 1 to 4, group 1 to 3 and leglength)?

_________________
http://www.spssforum.com
http://www.spsslog.com

Top

 Post subject: Posted: Tue Jun 19, 2007 7:07 pm
 Moderator

Joined: Tue Jun 12, 2007 12:08 pm
Posts: 2074
Location: Florida, USA
As noted, please describe the variables; why do you say they are non-parametric?

FYI, ANOVA is robust so still might be viable but repeated measure design, not sure without more info.

_________________

Statman
Statistical Services
IBM/SPSS Beta Site

Top

 Post subject: Posted: Wed Jun 20, 2007 8:24 am

Joined: Wed Jun 20, 2007 7:54 am
Posts: 2
Hi statman & forum.

Thanks for ur fast replies.
Sorry for the short description.

This is the case.
We let people circumvent an obstacle in 4 different conditions, and observed the circumvention strategy.
We did see 3 strategies. Cross-step / semicross-step / side-step.
We have repeated every conditions untill we had 4 circumventions per condition.

So in total its for example (for 1 subject)
Condition 1: cross(C) C C sideway(SW) | Condition 2: C SW SW Semi-cross(SC) | Condition3: SW SW SW C | Condition 4: SW SW SC SC

we did this for 21 subjects total. So actually we have 16 variables. Condition 1-trial1, condition1-trial2, con3-trial3, con4-tr4 then con2-trial1, con2-trial2 etc.
We defined in SPSS C=1, SC=2, SW=3. as outcome of these variables.

Now we want to see if the conditions(con1 con2 con3 con4) make significant difference on the outcome (C/SC/SW or in spss 1-2-3)

Further we have sexe, age and measured for example leglength and like to know if the leglength explains the differences among subjects for their circumvention strategy choice.

So, the first described varible is nominal and we believe it's non-parametric too.
The leglength is ratio variable

Hopefully this information gives a clearer view about our question

Thanks again.

Top

 Post subject: Posted: Wed Jun 20, 2007 12:40 pm
 Moderator

Joined: Tue Jun 12, 2007 12:08 pm
Posts: 2074
Location: Florida, USA
I would look into multivariate ANOVA using circumventions ("4 time periods") as the covariate. You will probably have to restructure your data so that you have the 4 condition variables and then your cases as rows with a new variable being nominal from 1 to 4 the circumvention.

Yes, this methodology could be debatable but at least it might offer some insight into differences

If you want, send me your data and I will take a closer look

_________________

Statman
Statistical Services
IBM/SPSS Beta Site

Top

 Post subject: Posted: Tue Jul 03, 2007 2:28 pm

Joined: Wed Jun 20, 2007 7:54 am
Posts: 2

I think we have a solution.
We have the '4 time periods' per strategy as variable. The values of these variable are the frequency of the appearence of that variable.
So we have used cross-150 | cross-150DT | cross-300 | cross 300DT.
In this design we can just use the rep. measures. and see the differences between the different conditions.

For our other research questions we have restructured the data
in variables subject-trial(1-4)-DT(yes1 no=0)-delay(1=fast 2=slow)-strategy(1-2-3)-Force(value)
Then he used a Univariate anova with subject en trial as randomised factor
DT - Delay- strategy as fixed factor and Force as dependent variable.
This should give answer to the question if force is different over strategy and DT.

Hopefully this will work fine for us.

greets

Top

 Post subject: Posted: Sun Jul 08, 2007 7:55 pm

Joined: Fri Jun 22, 2007 7:00 am
Posts: 136
Location: Australia
Mendel,

Thank you for coming back to us on this. I assume your question is answered so I will close this topic.

_________________
http://www.spsslog.com
http://www.spssforum.com
From SPSS users, for SPSS users

Top

 Display posts from previous: All posts1 day7 days2 weeks1 month3 months6 months1 year Sort by AuthorPost timeSubject AscendingDescending
 Page 1 of 1 [ 7 posts ]

 All times are UTC

#### Who is online

Users browsing this forum: No registered users and 1 guest

 You cannot post new topics in this forumYou cannot reply to topics in this forumYou cannot edit your posts in this forumYou cannot delete your posts in this forumYou cannot post attachments in this forum

Search for:
 Jump to:  Select a forum ------------------ SPSS    General    Data input    Output    SPSS for beginners    Enhancing data    Modifying SPSS output    Coding SPSS syntax SPSS the software    Errors    Where is .. in SPSS Statistics    Statistics for beginners    Heavy statistics SPSS Jobs    SPSS Jobs SPSS Forum    Read this first    General feedback or questions about this forum    Library