nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Mismatch in Stepwise estimates and actual models

abra

Hello,

I worked with the stepwise platform to find a model in JMP 10.

My question deals with a comparison of the estimates that I get directly from the stepwise platform to those I receive after I choose run/make model.

If I select manually in the stepwise platform 2 variables of interest (x1, x2) to explain a certain y,

within the stepwise window the estimates are:

R^2 = 0.30,

x1 -> p = 0.002

x2 -> p = 0.92

I was interested in these variables and therefor I ran the model.

I got the following estimates in the fit least square window (after pressing "run")

R^2 = 0.69

x1 -> p < 0.0001

x2 -> p < 0.0001.

I checked the VIF, and both have the exact same VIF of 1.26 (same 7 digits after the dot)

In the past I saw differences between the estimates that stepwise offered and the actual model but they were never so big.

Right now I do not feel that I can work with the stepwise platform with this data set.

Can someone offer an explanation, or say what additional information I should provide?