Help! Econnometrics Multicollinearity (MC) Question

Watch
Announcements
#1
Hi so i know how to check MC and know the basic ins and outs, theres a screenshot attached of Eviews where i used VIF to check for MC and I can obviously see an issue exists between Age and Age Square I was wondering if theres a typical cut off point for MC that i should write about in my theory such as I've seen some people note anything above 0.7 or below -0.7 denotes an issue is this limit typical? How do i find my own personal limit like for example how would i write out why theres multicollinearity, obviously because they're highly correlated but all things are correlated in one way or another so how much correlation is too much if that makes sense? Also I've chosen to remove age square to solve the issue, what economic theory could i cite to attribute this decision too? The question relates to the factors that affect wages.
Thank you for any help in advance
0
3 years ago
#2
Factors affecting wages is a well-explored topic, and it's accepted that age-squared is an important factor. Both are important terms. As you get older, you get a higher wage (positive age coefficient), but this is a diminishing effect (negative age-squared coefficient).

When you include both a variable and it's square, it's likely that there will be multicollinearity - one is dependent on the other after all. This isn't a problem in this case, you have specified the model to be this way.

To take away from this - you shouldn't blindly follow rules on how to interpret numbers, you have to really think about what the numbers mean. For another good example of this, see p-hacking.
0
#3
(Original post by Speckle)
Factors affecting wages is a well-explored topic, and it's accepted that age-squared is an important factor. Both are important terms. As you get older, you get a higher wage (positive age coefficient), but this is a diminishing effect (negative age-squared coefficient).

When you include both a variable and it's square, it's likely that there will be multicollinearity - one is dependent on the other after all. This isn't a problem in this case, you have specified the model to be this way.

To take away from this - you shouldn't blindly follow rules on how to interpret numbers, you have to really think about what the numbers mean. For another good example of this, see p-hacking.
Hi we were told to remove one to solve the issue of multicollinearity, so most people picked to remove age square but nobody has actually given a reason as to why they've removed age square as opposed to age I was wondering if you could provide me a reason as to why removing one variable is more effective or useful in our regression than the other
0
X

new posts
Back
to top
Latest
My Feed

Oops, nobody has postedin the last few hours.

Why not re-start the conversation?

see more

See more of what you like onThe Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

Poll

Join the discussion

I'm worried it will negatively impact me getting into university/college (205)
43.9%
I'm worried that I’m not academically prepared for the next stage in my educational journey (52)
11.13%
I'm worried it will impact my future career (36)
7.71%
I'm worried that my grades will be seen as ‘lesser’ because I didn’t take exams (99)
21.2%
I don’t think that receiving these grades will impact my future (48)
10.28%
I think that receiving these grades will affect me in another way (let us know in the discussion!) (27)
5.78%