Disadvantage of Non-parametric vs. Parametric Test
Dear All Could anyone kindly tell me a major disadvantage of using non-parametric test compared to parametric test? Your response will be appreciated. Regards, Boonlert W.
Re: probability question
Michael Cohen ([EMAIL PROTECTED]) wrote:
: : On Wed, 1 Dec 1999, Yonah Russ wrote:
: :
: : > how do you solve a problem like this one?
: : > thanks in advance
: : > ---
: : > X is a chance variable such that X(omega)={1,2,3...,n}
: : > and for every i in {1,2,3...n}, 4P(X=i+2)=5P(X=i+1)-P(X=i)
: : >
: : > find the breakdown of X.
: :
:
: Let's take n finite first. Let ^ denote "raised to power". Using difference
: equations, the solutions are constant and constant times (1/4)^i or linear
: combinations thereof: P(X=i) = a (1/4)^i + b where a and b are constants. The
: requirement that the P(X=i) sum to 1 forces a=3(1-bn)4^n/(4^n-1) and b is at most
: 1/n.
:
: If n is infinite, then b must be 0 (otherwise the sum of the P(X=i) is infinite).
: Then a=3. So P(X=i)=3 (1/4)^i.
:
One revision: At the end of the second to last paragraph, I should not have
required b be at most 1/n (a can be negative). We just need to be sure P(X=1)
and P(X=n) are non-negative --- the other values are in-between.
--
Michael P. Cohen home phone 202-232-4651
1615 Q Street NW #T-1 office phone 202-219-1917
Washington, DC 20009-6310 office fax 202-219-1736
[EMAIL PROTECTED]
Re: "Lambda" for discarding data
On Fri, 03 Dec 1999 18:18:21 -0800, [EMAIL PROTECTED] (PBL Powerbase 180) wrote: >First, I am a molecular biologist and have only taken intro stats > courses during my undergrad days. > > I am trying to find a criterion to use as an objective cutoff to discard > data from experiments. I use non-linear regression curvefitting, and > Prism, so I emailed graphpad tech spt. Here is the email, and response: < snip, Q and "never heard of it"; and noncentrality > > If anyone has any information how to compute "lambda", I would greatly > appreciate it (in case it helps, my boss was referring to the antiquated > (1975) program, BIOPROG, which gives "lambda" values.) 1) For lambda, I think you have to find someone who used that same program. It is *highly* unlikely that the lambda it used had anything to do with the noncentrality parameter that you found. 2) There is no such thing as a *valid*, generalizable, "objective cutoff to discard data from experiments." Some people might say, "discard" is a word that a scientist would never use. - I have seen the worth of extremes most neatly in examples from geological-surveys: where the paydirt, or motherlode, is among the extreme values which do not fit the others. One notion of cutoff which has validity might be the observation that a single point can be so extreme that it makes "non-sense" -- in one fashion or another -- of a particular *statistical analysis*. For instance, what is being analyzed in ANOVA is the sums of squares around means; one value can be so extreme that all the other values might as well be zero, for all their effect on the tests. However, you have to comment on the data-point when you set it aside from the ANOVA; you do not "discard data" from the experiment. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html
Re: Disadvantage of Non-parametric vs. Parametric Test
"boonlert" <[EMAIL PROTECTED]> writes: > Dear All > Could anyone kindly tell me a major disadvantage of using > non-parametric test compared to parametric test? > Your response will be appreciated. > Here's one. Lack of readily available software to produce confidence intervals. In some simple situations, confidence intervals for some population quantities are available through the order statistics, but I don't know of any readily available software that produces them.
Re: Disadvantage of Non-parametric vs. Parametric Test
Jerry Dallal wrote: > Here's one. > Lack of readily available software to produce confidence intervals. > In some simple situations, confidence intervals for some population > quantities are available through the order statistics, but I don't > know of any readily available software that produces them. MINITAB has commands WINT (signed-rank interval), SINT (sign interval), and automatically produces an interval for difference of medians with the MANN (-Whitney-Wilcoxon) command, as well as the hypothesis test. -Robert Dawson
Re: stats packages for Unix
try stata http://www.stata.com David Cross/Psych Dept/TCU <[EMAIL PROTECTED]> wrote: : Check out the following websites: : http://forrest.psych.unc.edu/ : http://www.linuxapps.com/ : Cheers, : David Cross : On Thu, 2 Dec 1999, Bob Hayden wrote: :> Any advice on basic stats. packages for Unix/Linux? We used Minitab :> on a DEC alpha box using DEC Unix. That was fine by me but the DEC :> was too expensive to maintain. So, we replaced it with a FreeBSD :> system runing on a pile of Intel boxes. But, Minitab does not run on :> that, so we started using the PC and iMac versions. However, the iMac :> lab is a disaster (local problem, not a Mac problem, I think) and we :> hate supporting three platforms. :> :> The software must be VERY easy to use and not cost much more than :> Minitab, which is quite cheap for a site license. We have a few :> hundred students per year using the software. Most of them have :> pretty limited math., computer and study skills. :> :> :> _ :> | | Robert W. Hayden :> | | Department of Mathematics :> / | Plymouth State College MSC#29 :>| | Plymouth, New Hampshire 03264 USA :>| * | Rural Route 1, Box 10 :> /| Ashland, NH 03217-9702 :> | ) (603) 968-9914 (home) :> L_/ [EMAIL PROTECTED] :> fax (603) 535-2943 (work) :>
Re: Disadvantage of Non-parametric vs. Parametric Test
Disadvantages of non-parametric tests: Losing precision: Edgington (1995) asserted that when more precise measurements are available, it is unwise to degrade the precision by transforming the measurements into ranked data. Low power: Generally speaking, the statistical power of non-parametric tests are lower than that of their parametric counterpart except on a few occasions (Hodges & Lehmann, 1956; Tanizaki, 1997). Inaccuracy in multiple violations: Non-parametric tests tend to produce biased results when multiple assumptions are violated (Glass, 1996; Zimmerman, 1998). Testing distributions only: Further, non-parametric tests are criticized for being incapable of answering the focused question. For example, the WMW procedure tests whether the two distributions are different in some way but does not show how they differ in mean, variance, or shape. Based on this limitation, Johnson (1995) preferred robust procedures and data transformation to non-parametric tests. Hope it helps. Chong-ho (Alex) Yu, Ph.D., CNE, MCSE Instruction and Research Support Information Technology Arizona State University Tempe AZ 85287-0101 Voice: (602)965-7402 Fax: (602)965-6317 Email: [EMAIL PROTECTED] URL:http://seamonkey.ed.asu.edu/~alex/
Re: Disadvantage of Non-parametric vs. Parametric Test
At 10:33 AM 12/6/99 -0700, Alex Yu wrote: >Disadvantages of non-parametric tests: seems to me that before one lists out "dis" advantages ... or for that matter "ad" vantages ... one needs to be very clear on what one wants to know about the target population ... now, in some cases ... there might be several approximately equal alternative parameters or pieces of population information that might be sufficient for your purposes and, in that case ... then using some technique with better power, etc. might be helpful. but, if we really are interested in some element in the population BUT, the technique for "inferencing" it happens to be non-parametric ... AND, it is that and ONLY that that we are interested in ... then we might have to "give up on some power" ... -- 208 Cedar Bldg., University Park, PA 16802 AC 814-863-2401Email mailto:[EMAIL PROTECTED] WWW: http://roberts.ed.psu.edu/users/droberts/drober~1.htm FAX: AC 814-863-1002
Mathematics and mind...
For those interested in mathematics and mind... I'm running a simple psychology experiment over the internet. It doesn't take much time, and very little mental effort. It is vaguely related to mathematics, and you might be interested in the outcome. If you take part in the experiment, you get shown some of the results. If you're interested, go to http://www.stanford.edu/~gruffydd/math/ and follow the instructions. There is a quick consent form to complete, and the experiment just involves responding to some simple questions. Having finished the task, you can spend the rest of the day being proud that you have furthered our knowledge of how the mind works.
Re: Mathematics and mind...
[EMAIL PROTECTED] (Thomas Lawlor Griffiths) writes: >For those interested in mathematics and mind... >I'm running a simple psychology experiment over the internet. It >doesn't take much time, and very little mental effort. It is vaguely >related to mathematics, and you might be interested in the outcome. >If you take part in the experiment, you get shown some of the >results. >If you're interested, go to >http://www.stanford.edu/~gruffydd/math/ >and follow the instructions. There is a quick consent form to >complete, and the experiment just involves responding to some >simple questions. >Having finished the task, you can spend the rest of the day being >proud that you have furthered our knowledge of how the mind works. It is actually rather painful (metaphorically speaking). After clicking "Begin experiment", the browser appears to hang. Then, after an almost interminable delay, it comes back with a message that the page contained no data.
Re: Disadvantage of Non-parametric vs. Parametric Test
[EMAIL PROTECTED] (Robert Dawson) writes: > Jerry Dallal wrote: > >> Here's one. >> Lack of readily available software to produce confidence intervals. >> In some simple situations, confidence intervals for some population >> quantities are available through the order statistics, but I don't >> know of any readily available software that produces them. > > MINITAB has commands WINT (signed-rank interval), SINT (sign interval), and > automatically produces an interval for difference of medians with the MANN > (-Whitney-Wilcoxon) command, as well as the hypothesis test. Great! Maybe it'll force other vendors to do likewise. Thanks for the heads-up.
Likelihood Ratio Test
I turns out that the difference between the constrained and unconstrained log likelihoods has a chi-square distribution (dergr. of freedom =restrictions) when it is multiplied by two. Where the two comes from is something I am not getting. (or -2, the information I have isn't completely consistent, but I guess that just depends on which way the difference between the likelihoods is taken). Where is this 2 (or -2) comming from? Sent via Deja.com http://www.deja.com/ Before you buy.
Re: Mathematics and mind...
it worked fine for me...perhaps your computer just coughed Neil W Rickert wrote: > > [EMAIL PROTECTED] (Thomas Lawlor Griffiths) writes: > > >For those interested in mathematics and mind... > > >I'm running a simple psychology experiment over the internet. It > >doesn't take much time, and very little mental effort. It is vaguely > >related to mathematics, and you might be interested in the outcome. > >If you take part in the experiment, you get shown some of the > >results. > > >If you're interested, go to > > >http://www.stanford.edu/~gruffydd/math/ > > >and follow the instructions. There is a quick consent form to > >complete, and the experiment just involves responding to some > >simple questions. > > >Having finished the task, you can spend the rest of the day being > >proud that you have furthered our knowledge of how the mind works. > > It is actually rather painful (metaphorically speaking). After > clicking "Begin experiment", the browser appears to hang. Then, > after an almost interminable delay, it comes back with a message that > the page contained no data. -- Are the stars and planets studying us? mailto:[EMAIL PROTECTED] http://www.ilos.net/~telefina/
Re: Likelihood Ratio Test
In article <82harm$1ae$[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote: >I turns out that the difference between the constrained and >unconstrained log likelihoods has a chi-square distribution (dergr. of >freedom =restrictions) when it is multiplied by two. >Where the two comes from is something I am not getting. (or -2, the >information I have isn't completely consistent, but I guess that just >depends on which way the difference between the likelihoods is taken). >Where is this 2 (or -2) comming from? The first paragraph sketchily explains why the problem can be reduced to one observation from a multivariate normal distribution with independent coordinates with variance 1. Then the calculation is made. The general case is asymptotically that of a multivariate normal distribution with known covariance matrix, and the particular covariance matrix is not important (algebra). As the mean is a sufficient statistic, we can assume sample size 1. This case also reduces to the null hypothesis that the entire mean is 0. Now the logarithm of the likelihood function is C - .5 * \sum (X_i - \mu_i)^2. The C is unimportant; for the unconstrained maximum, the sum is 0, and for the constrained maximum, it is chi-squared with the relevant number of degrees of freedom. But note the factor of -.5. -- This address is for information only. I do not claim that these views are those of the Statistics Department or of Purdue University. Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399 [EMAIL PROTECTED] Phone: (765)494-6054 FAX: (765)494-0558
