The latter jack Tomsky comment !!!

*** The above is an example of statistical
illiteracy.

Jack (moderator)***

WHAT Jack Tomsky thinks is directed to? To imbeciles
or little boys to which can be said: DO NOT SAY THAT;
I WANT NOT.
(and one deserve any explanation!)
***********************

What I said (below) is the current way true
Statisticians use to draw conclusions from Hypotheses
Tests. Readers should be aware.

Luís Amaral Afonso [The moderator destroyer]


Wormtomsky said:

*** The null hypothesis is not a point in the
acceptance region. The null hypothesis describes a
subset of the parameter space. The acceptance region
is a subset of the sample space. As long as the
sample is in the acceptance region, the null
hypothesis is accepted.***

WHAT A MESH!!!
__1__The null hypotheses DOESN´t describes a subset
in the parameters space. WHERE YOU FOUND THIS?

__2__The acceptance region a subset of the sample
space? IMBECILITY . You do not know what the Sample
Space is!!! The sample space is the set of all
possible samples drawn from the Universe(s) in test.
They do not have directly anything to do with the
acceptance region !!! They merely include all
possible samples independently what one will do with
them!!!
Your statement
*** The acceptance region is a subset of the sample
space.*** worth to be included in a NON SENSE
ENCYCLOPEDIA.

__3__When you put (two-tails hypotheses test) H0: p=0
you
__or try to find if the parameter p is equal to 0, ,
which is you thesis, and are unable o prove so
__or the true Statisticians will find out if the data
is so that we cannot (scientific doubt) state it is
not compatible with H0, case in which we say that
there is no sufficient evidence to reject H0. The
state of the Universe change (if so) was so slightly
that one couldn’t realize that change. Could it
remain unchanged? Of course yes, but WE ARE UNABLE TO
SAY IF IT´S TRUE OR NOT TRUE:
In Science on is not allowed to peremptory state
things that are dubious. In Statistics all statements
issued from random sampling are liable to be wrong.
C´est la Vie!

WORMTOMSKY THEORY


*** As long as the sample is in the acceptance
region, the null hypothesis is accepted ***

My comment

Attention to the latest result in Statistics Theory
*** The samples can be in the accptance region or in
the rejetion region ***
____THE SAMPLES, BWAU; BWAU; BWAU; BWAW; BWAU
!!!!!!!!!!!!!!!!!!!!

Luís Amaral Afonso [The mderator (??????) destroyer]

*** As long as the sample is in the acceptance
region, the null hypothesis is accepted ***

WORMTOMSKY THEORY


*** As long as the sample is in the acceptance
region, the null hypothesis is accepted ***


My response
What I note is that WORMTOMSKY PUT SAMPLES IN THE
ACCEPTANCE REGION
(A pyramidal BRUNDLE because all of us knows that the
acceptance region is the region containing the values
of the test statistics corresponding to the decision:
fail to reject the null hypotheses.
It doesn´t could have noting to do with parameters
(GoFit) and samples are too wild to be inserted into
a region (or interval).
BRUNDLES and Jack Tomsky are TWIN BROTHERS.
THIS I AGREE.

Luis Amaral Afonso [the moderator destroyer]

Wormtomsky said:

*** Samples always fit into the sample space by
definition. They sometimes fit in the acceptance
region, in which case the null hypothesis is
accepted. If they fit in the rejection region, then
the alternative hypothesis is accepted.***

My response

__1__From you my concern is not to have my brains
completely in a chaos as you have,
__2__Only a drunk or druggy person do pretend that
*** samples *** could be included into acceptance
regions, which is an interval (One dimension or
several ones). TO GET THAT IT IS IMPOSSIBLE, enough
to take in consideration what samples are.

__3__Your complete despise from ethical norms, lead
you to say whatever, whenever, no matter the its
context.

Your way to reasoning (???) concerning Statistics is
very, very similar. to that of the Mexican Movies
Star, Cantinflas, Mário Moreno (famous several
decades ago).

NO, IMPOSSIBLE, INTERVALS CANNOT CONTAIN SAMPLES:
CONFIDENCE INTERVALS CAN CONTAIN OR NOT THE VALUE OF
THE TEST STATISCS WE ARE DEALING WITH.

one say (R.A. Fisher said) that we has not sufficient
evidence to reject H0, if NOT we reject the Null
Hypotheses because there are incompatibility between
the data and the extreme probability we are willing
to consider yet coming from sampling randomness.

Wormtomsky said:

*** Samples always fit into the sample space by
definition. They sometimes fit in the acceptance
region, in which case the null hypothesis is
accepted. If they fit in the rejection region, then
the alternative hypothesis is accepted.***

My response

__1__From you my concern is not to have my brains
completely in a chaos as you have,
__2__Only a drunk or druggy person do pretend that
*** samples *** could be included into acceptance
regions, which is an interval (One dimension or
several ones). TO GET THAT IT IS IMPOSSIBLE, enough
to take in consideration what samples are.




An acceptance region is not an interval, unless the sample size is one. The sample space is of dimension n, the sample size, and includes the space of all possible sample outcomes. The acceptance region is a subset of the sample space which includes all outcomes for which one would accept the null hypothesis.



__3__Your complete despise from ethical norms, lead
you to say whatever, whenever, no matter the its
context.



What I say is consistent with my knowledge of statistics. When you seemed to accept the acceptance region being non-null, I praised you. When you took that back, I then took back my praise.



Your way to reasoning (???) concerning Statistics is
very, very similar. to that of the Mexican Movies
Star, Cantinflas, Mário Moreno (famous several
decades ago).

NO, IMPOSSIBLE, INTERVALS CANNOT CONTAIN SAMPLES:
CONFIDENCE INTERVALS CAN CONTAIN OR NOT THE VALUE OF
THE TEST STATISCS WE ARE DEALING WITH.




The acceptance regions are not intervals, except for the special case when the sample size is one. Confidence intervals are not intervals on test statistics, but on parameters.



If contains
one say (R.A. Fisher said) that we has not sufficient
evidence to reject H0, if NOT we reject the Null
Hypotheses because there are incompatibility between
the data and the extreme probability we are willing
to consider yet coming from sampling randomness.




Fisher's formulation of tests of significance had only a single hypothesis and a calculated p-value. There was no alternative hypothesis. Since there was no alternative hypothesis that could be accepted, it didn't make sense to be able to accept the null hypothesis. The NP framework had a null and alternative hypothesis, either of which cold be accepted.

Jack (moderator)

Wormtomsky said

[0] *** An acceptance region is not an interval,
unless the sample size is one. The sample space is of
dimension n, the sample size, and includes the space
of all possible sample outcomes. The acceptance
region is a subset of the sample space which includes
all outcomes for which one would accept the null
hypothesis. ***


From the WEB

__1__Definition: The acceptance region occurs in the
context of hypothesis testing. Let T be a test
statistic. Possible values of T can be divided into
two regions, the acceptance region and the rejection
region. If the value of T comes out to be in the
acceptance region, the null hypothesis being tested
is not rejected. If T falls in the rejection region,
the null hypothesis is rejected.

__2__acceptance region
–noun Statistics. the set of values of a test
statistic for which the null hypothesis is accepted.

__3__the set of values in a test statistic for which
the null hypothesis can be accepted

ENOUGH
Jack Tomsky is WRONG by three (!!!) reasons: THREE
ERRORS IN A ROW,

__a__The region is defined based on the test
statistics, because he omitted this reference one
stays his *definition* doesn’t lead ANYWHERE:

__b__Size n=1 is ABSURD and impracticable to test
WHATEVER: the test statistics variance is impossible
to be estimated.

__c__The null hypotheses is not accepted (see__1__)
but simply ONE FAIL TO REJECT IT.

Luis Amaral Afonso [The moderator destroyer]

Jack,

you patiently answered to the same nonsense several
times and patiently
corrected everything. Credit for that. I think it is
difficult to stay
the course, especially because the author not only
lacks basic knowledge
but almost always gets out of line and starts
ts screaming.

Since you apparently volunteered for the job you may
make your life
easier by compiling all your answers to some kind of
FAQ and then refer
to this? Following the last few days we encountered a
lot of basic
stuff, for example 1) what is a space, 2) what is a
sample space, 3)
Fisher's null hypothesis testing vs.
4)Neyman-Pearson, 5)statistical
inference vs. 6)proof

and so on.

The problem is of course that you can't post a whole
textbook. But I
remember that some other guy (I think it was Rich
Ulrich) did something
similar and summarized several discussions about the
interpretation of
coefficients in a regression with multiple
interaction terms.

perry

The team that do NO ACCEPT THE NULL HYPOTHESES

___Ronald A, Fisher
___Herman Rubi
___Luis Amaral Afonso

preferring to say that (if so)
there is no sufficient evidence to reject H0 if the
test value lies out the critical region.

Luis Amaral Afonso [the moderator destroyer]

The team that do NO ACCEPT THE NULL HYPOTHESES

___Ronald A, Fisher
___Herman Rubi
___Luis Amaral Afonso

preferring to say that (if so)
there is no sufficient evidence to reject H0 if the
test value lies out the critical region.

Luis Amaral Afonso [the moderator destroyer]

I never heard of Herman Rubi, but here is what Professor Herman Rubin wrote on November 26, 2005.

http://www.math.yorku.ca/Who/Faculty/Monette/Ed-stat/0241.html

"I have yet to see a classical approach which considers the balance.  But if one assumes that the decision maker is consistent, then it is an easy result that action must consider a linear combination of the probabilities of the resulting outcomes in the various states of nature. Looked at otherwise, this corresponds to minimizing the integral of some consequence function with respect to some measure on the states of nature. This is the behavioristic Bayesian approach. But it is not the philosophical one; only the product of loss and prior is even considered by this. To take a simple example, suppose one is testing
that the mean of a normal distribution is exactly 0. Then the probability of rejecting the null hypothesis if it is true must be balanced against an integral of the probability of accepting this hypothesis over the non-zero values of the mean with respect to some measure. It is not that hard to proceed from this type of approach."

What is this business about "the probability of accepting this (null) hypothesis"?

Jack (moderator)

On Jul 13, 9:52 pm, Jack Tomsky
jtom... at (no spam) ix.netcom.com> wrote:
The team that do NO ACCEPT THE NULL HYPOTHESES

___Ronald A, Fisher
___Herman Rubi
___Luis Amaral Afonso

preferring to say that (if so)
there is no sufficient evidence to reject H0 if
the
test value lies out the critical region.

Luis Amaral Afonso [the moderator destroyer]

I never heard of Herman Rubi, but here is what
Professor Herman Rubin wrote on November 26, 2005.


http://www.math.yorku.ca/Who/Faculty/Monette/Ed-stat/0
241.html

"I have yet to see a classical approach which
considers the balance. But if one assumes that the
decision maker is consistent, then it is an easy
result that action must consider a linear combination
of the probabilities of the resulting outcomes in the
various states of nature. Looked at otherwise, this
corresponds to minimizing the integral of some
consequence function with respect to some measure on
the states of nature. This is the behavioristic
Bayesian approach. But it is not the philosophical
one; only the product of loss and prior is even
considered by this. To take a simple example, suppose
one is testing
that the mean of a normal distribution is exactly
0. Then the probability of rejecting the null
hypothesis if it is true must be balanced against an
integral of the probability of accepting this
hypothesis over the non-zero values of the mean with
respect to some measure. It is not that hard to
proceed from this type of approach."

What is this business about "the probability of
accepting this (null) hypothesis"?

Jack (moderator)

My comment

YES I made an error: the guy is Herman Rubin.
HOWEVER you called Prof. TIAZ when the ecat name is
TIAGO (de
Oliveiea) THEN you had 1 letter absent (O) and 1
wrong (Z by
G).Conclusion: YOU ARE WORSE THAN I AM.
BY THE WAY : why you don´t ask him if the Null
Hypotheses can be proved tue, then acceptable?
LET US KNOW WHAT HIS OPINION IS!!!!

Luis Amaral Aoso [the moderator destroyer]