Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: RE: joint significant


From   "Nick Cox" <[email protected]>
To   <[email protected]>
Subject   RE: st: RE: joint significant
Date   Wed, 17 Jan 2007 09:57:12 -0000

I would go a little further and warn you against 
mixing confidence and significance terminology. 

This practice presumably arises from noting instances 
in which a result significant at the 5% level (say) 
corresponds to one lying outside a 95% confidence interval,
but it does not really help anyone if one uses expressions 
such as `significant at the 95% confidence level' (I quote here
from a statistics for Earth scientists book put out by a major
publisher).  

There are several good reasons to avoid this usage completely.  

First, there are significance tests not associated with a 
confidence interval procedure. 

Second, it will seem at best sloppy and at worst wrong to those 
familiar with standard usage.  

Third, it is one step down a slippery slope. The same book 
makes frequent reference to significance at the 95% level 
and alternates between talking of the 5% confidence level 
and the 5% significance level. Competent statistical scientists 
can decode what the author really means, but they should not 
have to. Others can justifiably feel confused. 

Naturally these are sins of my unnamed author (I will supply 
the reference if anyone needs it!), but they are quoted to 
emphasise the dangers here. 

Nick 
[email protected] 

Rajesh Tharyan
 
> Let me rephrase what Justin has said...
> 
> 1. The null hypothesis is the hypothesis you are testing for possible
> rejection under the assumption that it is true.
> 
> 2. Therefore, you have to be concerned with wrongly rejecting 
> something that
> you assume is true.
> 
> 3. The maximum error you are willing to accept is say .05 i.e 
> 5 out of 100
> (assuming 95% confidence = 5% significance that is where the 
> .05 comes from)
> 
> 4. The P value tell you what the significance level is. In 
> your case it is
> .08 This is higher than what you are willing to accept.
> 
> 5. Therefore there is not evidence statistically to reject 
> what you assumed
> was true that is to repeat justins statement
> 
> Based on a confidence level of 95%, I would fail to reject the null
> hypothesis that the estimated coefficients are jointly equal to zero.

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index