Comments and Ratings

Function Bugged ?
It returns negative SS values on some analyses and SS/F/p results differ significantly from RM ANOVA calculated using SPSS or Statistica.

The correct answer is 22225.0, which is correctly stated in the comment section of this code. However, the code does not compute this. Revision needed to compute PRESS value.

I just checked one-sample (T2Hot1.m) using an example 13.3 from Armitage et al. Statistical Methods in Medical Research, ed. 4. The results are nearly the same, which is very good:
From book: T2=7.439, F=1.24 and P=0.39
From T2Hot1.m: T2=7.4315, F=1.2386 and P=0.3873

Very good job!

Is it possible to couple this function to RMAOV1MS so that the returned p-value is corrected?
I want to use the adjustment in an univariate design with 2 groups measured across time, and I don't see how to apply this function to this data otherwise.
Thanks!

Claudia

Dear Michael,
Thanks for your interest of this m-file. As you know, we only generated the Matlab computational algorithm from the original published paper. If you have any mathematical or statistical fundamentals inquiry you must refer to the author(s). I give you the Dr. Norbert Henze's email address
henze@stoch.uni-karlsruhe.de
and/or
N.Henze@math.uni-karlsruhe.de

However, we don't need the lognormal cdf (logncdf) neither the mean and variance of the lognormal distribution (longnstat) functions. For the used mean (mu) and variance (si2), which are the Henze-Zirkler mean and variance, must, as the author establish, to be converted to the lognormal Henze-Zirkler mean and variance. As you can see, using the provied Iris data example. The mean(mu)=0.7635 and variance(si2)=0.0112. With a Henze-Zirkler lognormal mean: -0.279408 and Henze-Zirkler lognormal variance: 0.1379069. But, if you try to use te mu and si2 values by the longnstat function, you get the incorrect lognormal values of 2.1459 and 5.7768e-004, respecively.

Yours,

Prof. Antonio Trujillo-Ortiz

Thanks a lot for this useful function, when I was using the functions this error occurred:

When the answer is “The sampled population is not normally distributed” this error happens:

Error in ==> AnDartest at 111
switch nargin

??? Output argument "AnDartest" (and maybe others) not assigned during call to "…\MATLAB\AnDartest.m (AnDartest)".

it seems the only data it can take is the vector from the example inside the function.... ;(

works easily and well. Thanks!

There was a typo in the above comment:
the second 'if P>alpha' should be replaced with 'if P>alpha/2'

Congratulation !
Very good work !
R. Mansilla
UNAM

We thank Alex Karandreas from the Department of Acoustics, Aalborg University, Denmark, for encourage us to include a graeco-latin example. This has been done.

Antonio Trujillo-Ortiz

Again, learn how to use the function ERROR.
Secondly, avoid growing variables inside a loop. Try to pre-allocate X

Finally, this one-liner does the same
X = full(sparse(repmat(1:size(table,1),numel(lcv),1).' , bsxfun(@plus,table,[0 cumsum(lcv(1:end-1))]) , 1))

but this might be indeed to cryptic, so split it up into sub-units. Also add error checks, e.g. using for instance any(bsxfun(@gt,table,lcv))

Also take a look into SUB2IND.

In short, there is still room for improvement.

Dear Jos. Thanks for your valuable comments for improve this m-file script. They were taken into account.

Nice implemetation and sample data to test the file on. However, it runs a bit slow. I recompted the variable Djk by avoiding loops and it runs much faster:
Djk = - 2*Y' + diag(Y')*ones(1,n) + ones(n,1)*diag(Y')';
/J.D.

Thanks for implementing this function. it helped me a lot.

Good for my reserach

Michael, dispite your regular visits the 10 000 self dowloads continue.

(It is not an exagerated figure)
 

Thanks a lot. I have implemented it in my scripts and works nicely.

Hi Ahmed,
Suppose you have vectors on R^4, a=[1 2 5 7] and b=[5 0 3 4]. Then, c=4*a+2*b=[14 8 26 36]. Their norm, magnitud or length = norm(c)=47.2440

Antonio

Dear Jos,

I think that without greater preambles. The description and the explanation have always been incorporated in the twostone m-file. Statistically speaking, this function has nothing to do with the ttest. This is another Student's t-test approach.

Yours,

Antonio Trujillo-Ortiz

Dear Jos,

Thanks for your note. But, please don’t confuse to the users. Your comment doesn’t have any statistical fundamental. You is refering to the one-sample Student’s t-test (ttest), of course comparing one sample mean with one hypothetical mean (expected), where the t-statistics is computed by,

t = (mean(x) – m)/se

where se = sqrt(var(x)/n)

The here proposed twostone m-file deals with two-sample when one of the samples has only one observation. Where,

1) m isn’t a mean, but an observation, for

2) the estimate standard error (se) is

se = sqrt(var(x)[(n+1)/n])

besides

3) the numerator of the t-statistic is

o – mean(x)

where o is the only one observation of one of the samples

Thus, for the next example (given in the twostone m-file)

x = 4.02 3.88 3.34 3.87 3.18 (sample with 5 observations)

y = 3.02 (sample with only one observation)

For this case, if you incorrectly use the ttest you get

t = 3.8303, with a se = 0.1666, and you should reject Ho

Using the adequate twostone m-file you get

t = - 1.5637, with a se = 0.4080, and you should accept Ho

So, this is not a matter of ‘form’, but of the correct statistical procedure.

Yours,

Antonio Trujillo-Ortiz

thanks..
www.youtubeline.com

Hi al key,

What really you mean with...'could not handle the situation that there as perfect matched row and column. '?

Your comment is very ambiguous and your rate its very strict. In order to correct some bugs, I'm awaiting for a more professional inquiry.