Linear Regression - Influence of experience on wage

#1
Hello Talk-Stats-Members,

i am new to the forum and already have a (in my eyes) complex theoretical problem that i hope you can help me with.

I am analysing a linear regression: The influence of job experience (indipendent variable) on the worker wage (dependent variable) in the USA in the year 1980.

In the first run, with other variables like education and family state, the job experience has a positive influence on the recieved wage.
But if i analyse the isolated influence of the experience on the wage, it has a negative influence.

My thoughts were that maybe in the year 1980, if you had a lot of working experience, you were less educated an therefore receive less wage. Also the standard deviation is relativly high, so maybe there is no correct statement you can conclude...

Anyways, i would like to know what possible reasons for this negative influence you could think of. Please let me know if you need any additional informations or real figures to the case.

Thank you in advance
Michael
 
#3
Thank you for your reply.

There is actually collinearity between some of the variables, another good point i can add to my work.

Is there a way to conclude something out of this fact or does it just mean that the model is not very precise?
 

Karabiner

TS Contributor
#4
Anyways, i would like to know what possible reasons for this negative influence you could think of.
IMHO this is not a question about statistics in the first place, but
some which requires substantial knowledge about the subject.
It seems to be a question for an economic science forum or social
science forum?

Generally speaking, you could read a bit about suppressor effects,
and think about factors which are known to influence wage, whether
they could possibly act as a suppressor here.

With kind regards

K.