Splitting Data : Transforms data from scale to ordinal

Discuss statistics related things

by bottlebrushpress » Tue Mar 23, 2021 6:30 am

Hi there,

I'm trying to compute a new variable that splits my data. When I try to split it using this : (FILTER(Cont_PostSum,Message=='Continuum')
it transforms my data from scale to ordinal, so I can't complete a one-way ANOVA to examine simple effects for my two-way repeated measures design.

Any tips?

Thanks :)

Anna
bottlebrushpress
 
Posts: 3
Joined: Thu Mar 18, 2021 1:15 am

by MAgojam » Tue Mar 23, 2021 9:54 am

Hi, @bottlebrushpress.

If you want to be able to change the Measure type for your new variable, use Data> Transform.
First, select your variable to Transform (Cont_PostSum) which will be recognized in the filter function as $source and proceed by entering in the "recode condition" box:
FILTER($source, Message == 'Continuum')
Before exiting you can change the Measure type if the one assigned automatically does not correspond to the one of interest.

Cheers,
Maurizio
User avatar
MAgojam
 
Posts: 145
Joined: Thu Jun 08, 2017 2:33 pm
Location: Parma (Italy)

by bottlebrushpress » Tue Mar 23, 2021 10:50 pm

Brilliant. Thanks :)
bottlebrushpress
 
Posts: 3
Joined: Thu Mar 18, 2021 1:15 am

by bottlebrushpress » Tue Mar 23, 2021 11:07 pm

Would you have any idea why I'm getting a message error saying 'factor consists of only one level'?

I'm trying to run a mixed ANOVA with a repeated measure on the second factor. I'm looking to examine whether opinions about schizophrenia change (and are stable over time pre-summary, post-summary, 3 weeks and 6 weeks), after participants have been presented with either a categorical definition of schizophrenia, or one that describes the condition as a continuum/spectrum. I selected a mixed ANOVA thinking it would be a 2 (message is either categorical or continuum) x 4 (time) with a repeated measure on the second factor. Within-subjects is time, between subjects is message. Does that sound about right?
The problem I'm having is that when I'm trying to split the data into 2 groups for simple effects/post hoc, it comes up with an error saying that my factor contains 'one level only', which I presume is the message factor. This factor should be continuum/symptoms (2 levels).
My concern is that maybe I'm running the wrong analysis? It wouldn't be a one-way analysis because there are two groups linked to the first factor, is that right?
bottlebrushpress
 
Posts: 3
Joined: Thu Mar 18, 2021 1:15 am

by jonathon » Fri Apr 02, 2021 10:32 pm

take a look at all the different combinations of factor levels ... what you'll probably find is that you have empty cells in your design.

cheers

jonathon
User avatar
jonathon
 
Posts: 1689
Joined: Fri Jan 27, 2017 10:04 am


Return to Statistics