[R] Inverse Student t-value

JLucke at ria.buffalo.edu JLucke at ria.buffalo.edu
Thu Oct 2 17:22:07 CEST 2014


Andre
For the last time, there is NO simple rational approximation to the 
quantiles of the t-distribution.

From  R help, qt = TINV is based on 
Hill, G. W. (1970) Algorithm 396: Student's t-quantiles. Communications of 
the ACM, 13(10), 619–620.  And
Hill, G. W. (1981) Remark on Algorithm 396, ACM Transactions on 
Mathematical Software, 7, 250–1. 

Both of these articles can be googled and the source code obtained from 
the ACM.  (I have done it.)
The code uses a rational approximation to the inverse for which you can 
compute values by hand.
Then you can see why we use qt().  :-)
Joe




Andre <geomodelers at gmail.com> 
10/02/2014 07:27 AM

To
JLucke at ria.buffalo.edu, 
cc
Duncan Murdoch <murdoch.duncan at gmail.com>, "r-help at r-project.org" 
<r-help at r-project.org>, r-help-bounces at r-project.org
Subject
Re: [R] Inverse Student t-value






Dear All,

The manual formula mean that how to calculate that value by hand for 
TINV(0.0000408831, 1221) and the resulted is 4.0891672

Appreciate your help in advance.

Cheers!


On Wed, Oct 1, 2014 at 9:15 PM, <JLucke at ria.buffalo.edu> wrote:
What do you mean by a "manual" formula? 



Andre <geomodelers at gmail.com> 
09/30/2014 11:54 PM 


To
JLucke at ria.buffalo.edu, 
cc
Duncan Murdoch <murdoch.duncan at gmail.com>, "r-help at r-project.org" <
r-help at r-project.org>, r-help-bounces at r-project.org 
Subject
Re: [R] Inverse Student t-value








Hi JLucke, 

Maybe the old excel function. TINV and T.INV.2T is same function for T
wo-Tailed Inverse of the student`s t-distribution but T.INV use for 
Left-Tailed inverse of the Student's t-distribution and can be use for 
 Inverse of the student`s t-distribution. 


I know automatic or functions any software but I just need a manual 
formula or compute formula (TINV or T.INV.2T) step by step presented by 
math for calculate until resulted. 

Thanks in advance. 


Cheers! 



On Wed, Oct 1, 2014 at 2:39 AM, <JLucke at ria.buffalo.edu> wrote: 
The website has your answer.  The t-distribution is a regularized 
incomplete beta function.  The incomplete beta function is given by R's 
pbeta function.  You regularize it with R's beta function.  Then you use 
R's uniroot function to find the inverse.  Good homework problem. 


Andre <geomodelers at gmail.com> 
Sent by: r-help-bounces at r-project.org 
09/30/2014 02:45 PM 


To
Duncan Murdoch <murdoch.duncan at gmail.com>, 
cc
"r-help at r-project.org" <r-help at r-project.org> 
Subject
Re: [R] Inverse Student t-value










Hi Duncan,

Let me explain again, I just need a manual expression for inverse student 
t
value.

You could go to web page
http://www.danielsoper.com/statcalc3/calc.aspx?id=10

That's inverse student t value calculator. Do you know a manual expression
use it.

Cheers!


On Wednesday, October 1, 2014, Duncan Murdoch <murdoch.duncan at gmail.com>
wrote:

> On 30/09/2014 2:26 PM, Andre wrote:
>
>> Hi Duncan,
>>
>> Actually, I am trying trace the formula for the "Critical value of Z" 
and
>> manual formula is =(I7-1)/SQRT(I7)*SQRT((TINV(0.
>> 05/I7,I7-2))^2/(I7-2+TINV(0.05/I7,I7-2)))
>>
>> So, I got new problem for TINV formula. I just need a manual equation 
for
>> TINV.
>>
>
> Sorry, can't help.  I'm not sure I understand what you want, but if it's 
a
> simple formula for quantiles of the t distribution, it doesn't exist.
>
> Duncan Murdoch
>
>
>> Hope solve this problem.
>>
>> Cheers!
>>
>>
>> On Wed, Oct 1, 2014 at 1:20 AM, Duncan Murdoch <
murdoch.duncan at gmail.com
>> <mailto:murdoch.duncan at gmail.com>> wrote:
>>
>>     On 30/09/2014 2:11 PM, Andre wrote:
>>
>>         Hi Duncan,
>>
>>         No, that's correct. Actually, I have data set below;
>>
>>
>>     Then it seems Excel is worse than I would have expected.  I
>>     confirmed R's value in two other pieces of software,
>>     OpenOffice and some software I wrote a long time ago based on an
>>     algorithm published in 1977 in Applied Statistics. (They are
>>     probably all using the same algorithm.  I wonder what Excel is 
doing?)
>>
>>         N= 1223
>>         alpha= 0.05
>>
>>         Then
>>         probability= 0.05/1223=0.0000408831
>>         degree of freedom= 1223-2= 1221
>>
>>         So, TINV(0.0000408831,1221) returns 4.0891672
>>
>>
>>         Could you show me more detail a manual equation. I really
>>         appreciate it if you may give more detail.
>>
>>
>>     I already gave you the expression:  abs(qt(0.0000408831/2,
>>     df=1221)). For more detail, I suppose you could look at the help
>>     page for the qt function, using help("qt").
>>
>>     Duncan Murdoch
>>
>>
>>         Cheers!
>>
>>
>>         On Wed, Oct 1, 2014 at 1:01 AM, Duncan Murdoch
>>         <murdoch.duncan at gmail.com <mailto:murdoch.duncan at gmail.com>
>>         <mailto:murdoch.duncan at gmail.com
>>         <mailto:murdoch.duncan at gmail.com 
>>> wrote: 

>>
>>             On 30/09/2014 1:31 PM, Andre wrote:
>>
>>                 Dear Sir/Madam,
>>
>>                 I am trying to use calculation for two-tailed inverse
>>         of the
>>                 student`s
>>                 t-distribution function presented by Excel functions 
like
>>                 =TINV(probability, deg_freedom).
>>
>>                 For instance: The Excel function
>>         =TINV(0.0000408831,1221) =         returns
>>                   4.0891672.
>>
>>                 Would you like to show me a manual calculation for 
this?
>>
>>                 Appreciate your helps in advance.
>>
>>
>>             That number looks pretty far off the true value. Have you
>>         got a
>>             typo in your example?
>>
>>             You can compute the answer to your question as
>>             abs(qt(0.0000408831/2, df=1221)), but you'll get 4.117.
>>
>>             Duncan Murdoch
>>
>>
>>
>>
>>
>>
>

                [[alternative HTML version deleted]]

______________________________________________
R-help at r-project.org mailing list 
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide 
http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.





	[[alternative HTML version deleted]]



More information about the R-help mailing list