Re: P-value in Excel
- From: "Linda" <devilish_child_5@xxxxxxxxxxxx>
- Date: 26 May 2006 06:48:19 -0700
Stephen J. Herschkorn wrote:
(snip)
tdist(x,d,2) gives the two-tailed area under the t density with d
degrees of freedom; the regions go from -infty to -x and from x to
infty. x must be nonnegative. Is this what you did not understand,
or do you not understand the statistics that tells us why we need this?
For this to be the correct p value, you must be running a test whose
test statistic has t distribution, and you must need a two-tailed test.
By the way, Microsoft designed the Excel functions for the t
distribution rather poorly. For the other distributions, the DIST and
INV functions (e.g., NORMSDIST and NORMSINV, CHIDIST and CHIINV, and
FDIST and FINV) are inverses of one another and refer to one-tailed
regions (i.e., the cumulative distribution function or its complement).
TDIST and TINV are different. TINV takes only two arguments (abscissa
and degrees of freedom) and always gives the *two-tailed* value. TDIST
takes a third argument which indicates whether you want a want a one-
or two-tailed value.
Yep, this is the answer that I was initially asking for. Thank you. All
input was much appreciated. Learnt a few things about Excel here!
.
- References:
- P-value in Excel
- From: Linda
- Re: P-value in Excel
- From: Stephen J. Herschkorn
- P-value in Excel
- Prev by Date: Re: Dynamics of Preditor-Prey Differential Equations
- Next by Date: Functional equation with delta
- Previous by thread: Re: P-value in Excel
- Next by thread: Corner Geometry
- Index(es):
Relevant Pages
|
