# Thread: All my coefficients are statistically significant

1. ## Re: All my coefficients are statistically significant

Originally Posted by victorxstc
.... and still doesn't know the importance of power.
R A Fisher did not know that either. He never used “power”. Never type 2 error and no alternative hypothesis.

I wonder if Bayesians at all use power. (But maybe something correspondingly.) Surly no type 2 errors.

Let us make the hypothesis that Noetsys all teachers were hard core Fisherians! @Noetsy, at what university were you?

(Or Noetsi, we can confess that we have all been sleeping sometimes during lectures.)

Originally Posted by royalstatus
I am running a fixed effects model in stata, but all my independent variables are statistically significant. I am wondering if I am doing something wrong because I have never seen this happen before. Has anyone ever encountered this problem before?
Jag have been running regression models and found that none of the independent variables were significant. And it has happened hat I have thought of that as a problem.

2. ## The Following User Says Thank You to GretaGarbo For This Useful Post:

victorxstc (09-25-2013)

3. ## Re: All my coefficients are statistically significant

From previously cited Lewis-Beck...

"As sample size increases, a given coefficient is more likely to be found significant (36)." From section called "Reasons why a Parameter Estimate May Not Be Significant." Berry's book is about regression assumptions, while sample size matters for that discussion (perhaps) it certainly doesn't need to cover the interpretation of coefficients in the context of power. Although, I suspect Berry too mentions power (haven't read that book for five years).

It's hard for me to imagine a (legitimate) statistics course not teaching students that larger samples recover smaller and smaller effects (although the language may very).

Regarding the original poster, he/she just didn't provide enough information.

4. ## Re: All my coefficients are statistically significant

All I can say is that we did not. And I took a variety of statistics courses at several graduate programs. We did talk about it in my graduate program in statistics.

I listed the Berry book because it was one of the two books used in a graduate regression course.

5. ## Re: All my coefficients are statistically significant

You are right. It is a very large sample of data