[Bug]: Results differ from results of Minitab

[stataisr]

JASP Version

0.19.1

Commit ID

No response

JASP Module

Regression

What analysis are you seeing the problem on?

Linear Regression

What OS are you seeing the problem on?

Windows 10

Bug Description

The VIF values change after the nominal factor is added to the model

Expected Behaviour

Results from Minitab

Steps to Reproduce

Steps.zip

...
Gorech Regression.zip

Log (if any)

No response

More Debug Information

No response

Final Checklist

• I have included a screenshot showcasing the issue, if possible.
• I have included a JASP file (zipped) or data file that causes the crash/bug, if applicable.
• I have accurately described the bug, and steps to reproduce it.

[stataisr]

It's interesting, that in JASP module "Quality Control" - DOE- Analyze design the VIF values do not change when a discrete factor is added

[tomtomme]

yep, interesting. We should investigate why that is.
@JohnnyDoorn - any idea?

[stataisr]

And also in JASP module "Quality Control" - DOE- Analyze design the values of the determination coefficients differ from ones of the Regression module

I have no idea, I think there is mistake in Regression module

[stataisr]

Results fm DOE of Minitab

[JohnnyDoorn]

Hi @stataisr,

I can reproduce JASP's regression results in R:

dat <- read.csv("~/Downloads/GorechRegression.csv")

mylm <- lm(G ~ D + Z1 + Z2 + as.factor(R), data = dat)
summary(mylm)
car::vif(mylm)

Which gives the same as JASP - there is only a single VIF value for the whole categorical predictor, which JASP places in the first row/dummy variable in the coefficients table:

@JTPetter do you have an idea why the VIF is not shown in the quality control table? Maybe there is a difference in VIF and GVIF?

Cheers,
Johnny