Thread: Does the intercept p value matter?

1. Does the intercept p value matter?

Hi,

I have outputs of a simple linear model as follows:

y = 0.0339 + 0.27918X

The p value of the intercept is 0.206 and of the slope is 0.0343.
The t stat of the intercept is 1.50 and of the slope 3.15

I know the test of the p value is if < 0.05 reject the null hypothesis and conclude the variable is statistically significant.

Multiple R2 = 0.844
F test = 9.94

My question - Does this model seem good? Does the fact that the intercept has a p value > 0.05 mean the model is not significant?

Many thanks,

2. Re: Does the intercept p value matter?

Whether the intercept matters or not really depends on what your model is trying to assess and what the meaning at the intercept actually is. Often the intercept alone has very little meaning and I wouldn't abandon a model just on a non-significant intercept though (but I also wouldnt abandon any model just because variables were insignificant!).

Your R2 implies that your model describes whatever relationship you are looking at well. I wouldnt however really say that any model is 'significant' just that a model is significantly better than some other model which could be used.

3. Re: Does the intercept p value matter?

Thanks for this.
The model is to see the effect on job creation (Y) of GDP growth (X).
So you would imagine that as an economy grows the more jobs are going to be created

4. Re: Does the intercept p value matter?

Originally Posted by anglais428
Thanks for this.
The model is to see the effect on job creation (Y) of GDP growth (X).
So you would imagine that as an economy grows the more jobs are going to be created
Sorry for really late reply but just in case its still helpful.

In your problem the intercept is going to be the 'effect on job creation' when x=0, ie. when GDP is zero. I suspect that has little meaning in your work, so I really wouldnt worry about any p-value for it.

5. Re: Does the intercept p value matter?

Hi, what if my y-variable is operating expenses and x-variable is sales?
A company provides services, not selling any products.
My p-value for intercept is 0.5 and p-variable for sales < 0.05 so do you think my model is applicable? Thanks.

 Tweet