|
Science Forum Index » Statistics - Math Forum » The absence of evidence is not evidence of absence
Page 1 of 1
|
| Author |
Message |
| Luis A. Afonso |
 Posted: Sat Apr 12, 2008 6:46 am |
|
|
|
Guest
|
The absence of evidence is not evidence of absence
_____H0: The victim did commit suicide
_____H1: A murdering is occurred
The absence of evidence is not evidence of absence
The Victim was found at home, sitting on a chair, a hole in his head, holding the gun that fired Him. No strange fingerprints were found in the gun.
We cannot conclude that he committed suicide (H0): the murder using gloves put the gun in the Victim’s hand after to shoot Him.
THE ABSENCE OF EVIDENCE (that a murdering occurred) IS ____________NEVER_________THE EVIDENCE OF THE ABSENCE OF SUCH POSSIBILITY.
**************
A short and slight lesson
Do NOT say I accept the Null
If you want to be sorry free:
Not to reject is the true rule,
Pearson said, much before me.
It’s the right Theory, full.
Different thoughts, do you see,
Are wrong as teach in school
Do not ever admit: let be.
The Null be an equality
(´cause it’s the same to state)
It’s, true, an absurdity.
To fight silly buffoonery
We never, yes, should be late:
Reader friend, do believe me.
*****************************
FURTHERMORE
Jack Tomsky had not scientific papers of his own: I do have in Révue de Statistique Appliquée, 40 , 1 , 1992.
Precisely on a test based on Monte Carlo simulations, a procedure that is repeatedly denied by the ignorant one cited above because he thinks that without a real sample Test Hypotheses are just nonsense. He didn’t pay attention that since the Lillifors´s Test (early 60´s) for normal samples (from which the mean and standard deviation are sample estimated) the method is fully standard.
In Portugal there are a handful of Statisticians (younger than I am) that would full available to make a round of conferences in USA to teach all of you what the precise meaning and scope of the Hypotheses Tests in Statistics are, I’m pretty sure.
“. . . pois não se pode tirar o Certo do Aleatório. “
(J. Tiago de Oliveira)
:… (So) one cannot get Certainty from Casual Data.
Luis Amaral Afonso (The Moderator destroyer) |
|
|
| Back to top |
|
| John Smith |
| Posted: Sat Apr 12, 2008 11:32 am |
|
|
|
Guest
|
Adumbo you idiot, I missed your stupidity. Welcome back.
As to your nonsense subject, in probability theory, absence of evidence is always evidence of absence. If E is a binary event and P(H|E) > P(H), "seeing E increases the probability of H"; then P(H|~E) < P(H), "failure to observe E decreases the probability of H". P(H) is a weighted mix of P(H|E) and P(H|~E), and necessarily lies between the two.
-----
Can you fathom this?
John 7 Smith |
|
|
| Back to top |
|
| aruzinsky |
| Posted: Sun Apr 13, 2008 5:37 am |
|
|
|
Guest
|
On Apr 12, 10:46 am, "Luis A. Afonso" <lic...@hotmail.com> wrote:
Quote: The absence of evidence is not evidence of absence
_____H0: The victim did commit suicide
_____H1: A murdering is occurred
The absence of evidence is not evidence of absence
The Victim was found at home, sitting on a chair, a hole in his head, holding the gun that fired Him. No strange fingerprints were found in the gun.
We cannot conclude that he committed suicide (H0): the murder using gloves put the gun in the Victim's hand after to shoot Him.
THE ABSENCE OF EVIDENCE (that a murdering occurred) IS ____________NEVER_________THE EVIDENCE OF THE ABSENCE OF SUCH POSSIBILITY.
**************
A short and slight lesson
Do NOT say I accept the Null
If you want to be sorry free:
Not to reject is the true rule,
Pearson said, much before me.
It's the right Theory, full.
Different thoughts, do you see,
Are wrong as teach in school
Do not ever admit: let be.
The Null be an equality
(´cause it's the same to state)
It's, true, an absurdity.
To fight silly buffoonery
We never, yes, should be late:
Reader friend, do believe me.
I agree. I never understood the lack of symmetry in choosing null
hypothesis. The procedure seems to be nothing more than a cultural
delusion. It is my understanding that statistical decision theory
avoids this kind of nonsense. |
|
|
| Back to top |
|
| Luis A. Afonso |
| Posted: Tue Apr 15, 2008 12:46 pm |
|
|
|
Guest
|
Since the Hyper Stupid Jack Tomsky states that is able to ACCEPT THE NULL HYPOTHESES,
I´am asking
***********how many flips HE needs to perform TO CONCLUDE A COIN IS FAIR.*************************************
This is the simplest Test Statistics even invented.
Luis Amaral Afonso ( The Moderator´s Destroyer) |
|
|
| Back to top |
|
| Luis A. Afonso |
| Posted: Tue Apr 15, 2008 10:49 pm |
|
|
|
Guest
|
DO ANSWER Jack Tomsky
DO ANSWER Jack Tomsky (I didn´t hear you)
HOW MANY FLIPS YOU THINK ARE NECESSARY TO PERFORM IN ORDER TO STATE A COIN IS FAIR?
(Since you, stupidly, think that a Null Hypotheses can be ACCEPTED).
**************
********* Consequently
__One are unable to state that a parameter is ZERO
__Unable to state two parameters are EQUAL.
**** BUT, in preference:
__There are not sufficient evidence that p is different from zero; not sufficient evidence that two parameters are different one another.
Luis Amaral Afonso (The Moderator Destroyer) |
|
|
| Back to top |
|
| John Smith |
| Posted: Thu Apr 17, 2008 2:13 am |
|
|
|
Guest
|
The reader is reminded that, among other incredibly stupid things, Afonso does not know the difference between significance levels and p-values:
-------------------------------------
AFONSO WROTE:
As I found
__________1 – alpha =______0.99868 (equal
variances)
__________1 – alpha =______0.99832 (different)
I immediately got the SIGNIFICANT LEVELS:
0.00132, 0.00168 (0.13%, 0.17%) (***Brilliant
isn’t?***)
__________
Luis
TOMSKY WROTE:
These are p-values, not significance levels.
Jack (moderator of the Math Forum)
-------------------------------------
And he also claims to be a Monte Carlo expert, but he cannot answer a simple question:
Quote: Adumbo claims to be a Monte Carlo Expert, but if you ask him, "Are the percentiles of a monte carlo distribution statistics or parameters?" his answer is:
[Jan 28, 5:00 am]
" I DO NOT KNOW, and I care not: " |
|
|
| Back to top |
|
| Luis A. Afonso |
| Posted: Thu Apr 17, 2008 5:09 am |
|
|
|
Guest
|
Everyone knows that by Monte Carlo simulations one is able to find out rigorous p-values that is the greater value of the test statistics in accordance with the Null Hypotheses. (Readers should remember that the simulations are performed supposing H0 true).
This 50 years old method is employed everyday by Statisticians that, contrarily to Jack Tomsky - John Smith, had not opted to stay in formalin flasks.
Luis Amaral Afonso (The Moderator Destroyer) |
|
|
| Back to top |
|
| John Smith |
| Posted: Thu Apr 17, 2008 6:48 am |
|
|
|
Guest
|
Quote: Everyone knows that by Monte Carlo simulations one is
able to find out rigorous p-values ...
What's ironic is that Adumbo doesn't even know what a p-value is, as shown by The Moderator, Jack Tomsky. See my post above in this thread.
John |
|
|
| Back to top |
|
| Luis A. Afonso |
| Posted: Thu Apr 17, 2008 7:26 am |
|
|
|
Guest
|
Are you saying that throughout Monte Carlo one cannot obtain p-values.
If it´s the case you are completely ignorant about this point of Statistics.
I would like you say why I don´t know what p-value are.
Luis Amaral Afonso (The Moderator Destroyer) |
|
|
| Back to top |
|
| John Smith |
| Posted: Thu Apr 17, 2008 8:40 am |
|
|
|
Guest
|
Quote: I would like you say why I don´t know what p-value
are.
Luis Amaral Afonso (The Moderator Destroyer)
Luis,
Obviously you did not read my previous post. If you wish to comment, read the entire thread.
Here is something you wrote some time ago:
-------------------------------------
As I found
__________1 – alpha =______0.99868 (equal
variances)
__________1 – alpha =______0.99832 (different)
I immediately got the SIGNIFICANT LEVELS:
0.00132, 0.00168 (0.13%, 0.17%) (***Brilliant
isn’t?***)
__________
Luis
-------------------------------------------
You call those numbers "significance levels". But they are not significance levels, as Tomsky pointed out. They are p-values!!!! Your own program created those numbers. So you cannot recognize p-values when they come from your own program!!!! You call them significance levels!!!
John |
|
|
| Back to top |
|
| Luis A. Afonso |
| Posted: Thu Apr 17, 2008 9:20 am |
|
|
|
Guest
|
If you are unable to read, see a Doctor!
________ 1 – alpha = p
If p = 0.99868 then alpha=0.00132 = SIGNIFICANCE LEVEL
AS I WROTE:
But Jack Tomsky (drunk, or droped) switched the quantities.
************
Luis Amaral Afonso ( The Moderator Destroyer) |
|
|
| Back to top |
|
| Jack Tomsky |
| Posted: Thu Apr 17, 2008 9:50 am |
|
|
|
Guest
|
Quote: If you are unable to read, see a Doctor!
________ 1 – alpha = p
If p = 0.99868 then alpha=0.00132 = SIGNIFICANCE
LEVEL
AS I WROTE:
But Jack Tomsky (drunk, or droped) switched the
quantities.
************
Luis Amaral Afonso ( The Moderator Destroyer)
This is not correct. Alpha, the significance level, is a chosen constant. The p-value is random and varies according to the sample.
As moderator, I want to make sure that no one would be as confused as Afonso, who thinks that p is always equal to 1-alpha.
Jack (moderator) |
|
|
| Back to top |
|
| Luis A. Afonso |
| Posted: Thu Apr 17, 2008 10:32 am |
|
|
|
Guest
|
1) From real data (random samples) a sample statistics W0 is obtained. (suppose we are dealing with a ONE TAILED HYPOTHESES TEST),
2) If we by data simulation (Monte Carlo) we are able to obtain the set of N (say 10E7) values of W we can count what’s the proportion p0 such that W >W0.
3) Having choose previously alpha we simply compare
_____p0 >= alpha______fail to reject H0, the null hypotheses
_____p0 < alpha _______reject H0.
This procedure the last 40-50 has been repeatedly used and will be in future more and more as computers technology improves.
Luis Amaral Afonso (The Moderators Destroyer) |
|
|
| Back to top |
|
| John Smith |
| Posted: Thu Apr 17, 2008 4:12 pm |
|
|
|
Guest
|
Adumbo you moron, respond to Our Moderator's salient point.
You said:
Quote: ________ 1 – alpha = p
If p = 0.99868 then alpha=0.00132 = SIGNIFICANCE LEVEL
Obviously you think that alpha = probability of a Type I error is something that is deduced from the data!!! HA HA HA
But Our Moderator (be sure to capitalize!) Tomsky pointed out:
This is not correct. Alpha, the significance level, is a chosen constant. The p-value is random and varies according to the sample.
Adumbo, thanks for tomorrow's laugh at the water cooler!!!
John |
|
|
| Back to top |
|
| Luis A. Afonso |
| Posted: Fri Apr 18, 2008 12:50 am |
|
|
|
Guest
|
Jack Tomsky PROTOZOA is zero on simulative procedures
1) From real data (random samples) a sample statistics W0 is obtained. (suppose we are dealing with a ONE TAILED HYPOTHESES TEST),
2) If we by data simulation (Monte Carlo) we are able to obtain the set of N (say 10E7) values of W we can count what’s the proportion p0 such that W >W0.
3) Having choose previously alpha we simply compare
_____p0 >= alpha______fail to reject H0, the null hypotheses
_____p0 < alpha _______reject H0.
This procedure the last 40-50 YEARS has been repeatedly used and will be in future more and more as computers technology improves.
THOSE THAT are denying that this is liable procedure and THAT IDENTIFIES N AS A SAMPLE are besides stuoid COMPLETELY IGNORANT OF POST-CLASSICAL (Computer-aided) methodology.
Luis Amaral Afonso (The Moderators Destroyer) |
|
|
| Back to top |
|
| |
