Multiple regression - minitab

I am getting back to you regarding the multiple regression analysis. I have found your suggested data set on the statistical website:
http://lib.stat.cmu.edu/datasets/bodyfat

and decided to calculate the body fat (so I am skipping the density - for my understanding - it is not for my project).

I have included the attachment with minitab.

1) is it correct?
2) When you look at the ind explanatory variables, it looks that only
knee
chest
have p-values less than .05 alpha
3) what represent a high riks or a fact of multicorreality ( i will fo the VIF test, and correlation coefficicient).

and when you look individually at t test for eac hvariable it looks that only:
neck
abdomen
forearm
wrist have value > than 1.96 (t/2 test for 5/2 alpha = .025).

So I should reduce the model but I am not sure what is the pefect one.

There are some unusual observations as well - I thought I could check them )Cook's values, DFits, H-leverage later)

-
Could you please send me the minitab attachment with the perfect model
and explain your decision?

I have a very detailed paper with lots of calculations, and I am planning to do everything (test for fits, multicorreality, stepwise, etc later). I would like to send you all my cal later and check and help me to correct mistakes.

I have 91% for A now, but still have a comprehensive exam ehead of me and this project. And I am actually not good in this subject, I really need help to excell here.

I would be very happy if you could help me to do it well.

For now, could you send me the
-improved model
and scatterplot of the best linear line to make sure it is ok.

Also could you help me to explain which variables I should eliminate.

Attachments