ANCOVA graphic?

[Miro]

Hello everyone,

I write my thesis and I am currently working on the data analysis. I did a factorial 2x2x2 ANOVA to test my main hypothesis. After that I included a covariate in the analysis and the effects on the dependent variable changed.

I want to show the effects of the factors BEFORE I controlled the influence of the covariate (the results from 2x2x2 ANOVA) and AFTER I controlled the covatiate (the resulta from ANCOVA). Is it possible to create a graphic on that?

[technocrat]

Hi, @Miro.

I'm fairly confident that it's possible, given assignment of the before and after results to separate objects.

For example, the ggpubr package has short-cut anova plots. The single page documentation is pretty sketchy, but it seems well covered in a post

Likewise, ggplot2 has long been used to plot ancova objects, for example here

A great resource to look for examples is the custom pre-Google filtered for R search engine rseek.org.

[Miro]

Mr. Richard Careaga, thank you for your response! I noticed that you left a comment on one of my previous posts, so I want to ask you something.

I would like, instead of ANCOVA to make a mixed ANOVA, therefor I need to switch to long data format (currently I am on ‚wide‘). My design is a 2x2x2x2 mixed ANOVA, as 3 factors are between- subjects and the last one is within-subjects.
I have many variables in my data frame (each item from my questionnaire is coded as variable).
When I try to do the data frame from wide to long data format using the command:
melt(data, id=c(„Faktor1“, „Faktor2“, „Faktor3“), measured=c(„Time1“, „Time2“))

it goes wrong – I get 5 columns (that is okay) but in the column “variable”, where the within variables are supposed to be, are all of the variables that I used to have in my initial data frame (incl. every item etc.). That's why and the scores in the last column („value“) are wrong.

I need to have the 5 variables (isolated) in long format to test my hypothesis with mixed ANOVA.
How do I manage to solve this, is there any way because I'm not sure where my mistake is.
Please, let me know if you have any question.

[technocrat]

Hi @Miro

Please see the FAQ: What's a reproducible example (`reprex`) and how do I do one? Using a reprex, complete with representative data will attract quicker and more answers. A question like this is very hard to answer in the abstract without reverse engineering the problem. Although that is sometimes possible it takes away from the time I spend volunteering (as much as I would like this to be paid, it isn't!) helping others.