help with my ANOVA

lgreg19 · Post by **lgreg19** » Mon Apr 17, 2023 9:03 am

I am currently in need of some guidance. I am performing a one-way ANOVA in order to find a difference/relationship between two different surveys and their scores. One of the scores have has continuous numbers and the other has 5 subscales considered nominal.

Whats weird is only 2 out of the 5 subscales will perform the ANOVA, the other 3 come up with 'not enough observations'. So i tried to total them up and do it that way, yet it still says not enough observations. I then tried to perform the non-parametric version and it still will not work. So, i'm trying to understand what i can do to combat this. Thank you:))

Bobafett · Post by **Bobafett** » Mon Apr 17, 2023 10:30 am

Can you show your data file here? It sounds as though you either want to run a correlation between your two surveys, or that you are using the second 'survey' as a grouping variable to see if these 'groups' have a significant difference in scores on the first survey. If you have too few participants (or none for specific groups) then the ANOVA should still run and exclude those groups from the analysis - hence why seeing your actual data (and output if you have it) would help clarify the issue...

lgreg19 · Post by **lgreg19** » Mon Apr 17, 2023 3:06 pm

Hi, Thanks for your response!

Im not sure I can upload to data set (just for ethical reasons)? However, all those subscales have the exact same amount of participants and all have responses, which is why it dosent make sense. Maybe ill try to run a correlation instead like you suggested!

reason180 · Post by **reason180** » Tue Apr 18, 2023 1:41 am

In an ANOVA, the dependent variable has to be "continuous" (i.e., sufficiently close approximation to being continuous). It cannot be "nominal."

Bobafett · Post by **Bobafett** » Tue Apr 18, 2023 9:54 am

The correlation could only really be done if your second variable was ordinal (not nominal) and so I would assume that you are indeed more interested in differences in mean 'first survey' scores based on the categories from your second variable (is this correct?). If so, I'm not clear on why you appear to be unable to run the ANOVA. It could be something really quite simple which is why a look at your data file would help. If you do not wish to share this, could you at least post a screenshot of it?

reason180 · Post by **reason180** » Tue Apr 18, 2023 11:18 am

@lgreg19

To get much assistance here, I think you'll need to create and post an omv file containing an attempted ANOVA (and/or correlation) on made-up data that's similar in structure to your actual data.

jamovi