Re: Solutions - Statistical Inference Casella & Berger
Bill
Atsuko
>You can buy solutions manual through www.amazon.co.uk
them. Do you know its ISBN or something more specific I can use?
Lauren
I can't find them either. This course is tough and a solutions guide
would be helpful. I have the first 5 chapters, but I need 6-10. If
you can find them, or anyone else, please let me know
Thanks.
audery
Please send me a copy of the solutions. My email address is:
Thanks a lot !
audery
yichun
my email is : yichu...@hotmail.com
Thanks a lot
sunflowers
Could you please send me a copy of the solution?
Tons of thanks!
my email: fyg...@yahoo.com
Fuygao
Bamboozled
I need solutions for statistical inference by casella and berger.
It'll be a great use for preparations for my qual exam.
Does anyone still give out solutions?
Please send me one to very...@hotmail.com
Very very appreciate it!!!!!
Thanks
saywhat
; )
Thanks -
saywhat
xjr...@yahoo.com
Thanks
Nick
statsbabe
I'd like a copy of the solutions to study for the qualifiers. Thanks!
lindsayrenfro at yahoo.com
Shweta
Can you please send me the solutions of Casella Berger. It will be a great help.
My email id is shwe...@yahoo.com
Thank you very much.
-Shweta
vahid
do any one help me.
if thre is please send me to this mail vmas...@iooc.co.ir
Erik Fransen
xiaoxin
aledinola
delighted if you could send me a copy at curzi...@hotmail.it Thanks a
lot!
statsbabe wrote:
> Hi Dr. Doofus,
> lindsayrenfro at yahoo.com
*** Sent From/Enviado desde: http://groups.expo.st ***
Alessandro
Could you send me the solutions manual for Casella Berger? My email address is curzi...@hotmail.it
I would be very thankful
Best regards
Alex
panjc
Can you kindly send me a copy of solution for statistical inference by Casella & Berger?
my email address:panj...@hotmail.com
Thanks!
Pan
Nuno
Can you please send me the solutions as well.
My email is nuno.miguel.cruz.neves AT gmail.com
Thank you.
shayan
If anyone can please send me the solution of casella and berger 2nd edition chapter 5 to onward, i will be very thankfull to him.
shayan
Howell
I got some questions from this book during my winter break, and I really need the solutions for those questions. However, it’s almost impossible for me to afford $24 for the solution manual on eBay. It will be nice if you can send to me. Thanks in advance.
marco piovesan
I'm looking for the solution manual for Casella Berger.
I know that someone has a pdf version. Can you help me?
Thanks,
marco
Nuno
My email is nuno.miguel.cruz.neves AT gmail.com
Thank you,
llpyher
Mirunix
fygao
Could you please send me a copy of the solution manual,
it would be definately a great help for my study.
My email: fyg...@yahoo.com
Tons of thanks in advance!
fygao
fygao
May I get a copy of the solution mannual of Statistical
Inferences by Casella & Berger, please?
My email: fyg...@yahoo.com
I appreciate your time and attention!
Fygao
Angie
I will be grateful to you.
my email is-
anju...@yahoo.com
Angie
could you please send me a copy of solution manual in this address-- anju...@yahoo.com
I appreciate your help.
Angie
Shavy
Thanks in advance
Angie
I found that you have solution manual for Casella and Berger (from
math forum). Could you please send me a copy of this, I am having hard
time preparing for my qual.
I appreciate your help.
you can send in
anju...@yahoo.com
or lami...@gmail.com
Rick
Could you please send me a copy of the solutions manual?
My Email is: tutu...@gmail.com.
I will really appreciate it.
Rick
Could you please send me a copy of the solutions manual?
My Email is: rickm...@hotmail.com.
tester
Can you kindly email me the copy of solutions manual?
My email is: stat_in...@yahoo.com
Thank you so much for all your help!!
tester
I need complete solution manual, not partial.
my email: stat_in...@yahoo.com
Really appreciated your kindness,
Ray Lien
I would really appreciate it.
My Email: zhang...@yahoo.com
eye1234
Casella and Berger.
I would really appreciate it.
My Email: abc...@hotmail.com
raymond1980
Optikal
pgoe...@gmail.com
Thanks...
alan turing
Could you please email me the solution for Statistical inference by casella?
My email address is jordanbhec...@hotmail.com
caicai
circinus
Thank you very much. my email: circinus...@dreamwiz.com
stmat
If you got that, could you please send me?
Regards
Thanks
email: stma...@yahoo.com
jolly_cool
Can you send me a copy of the solution manual of statistical inference book by casella and berger. I will be very grateful to you. My email id is: venky8...@yahoo.com
or do let me know how can I get the copy?
thanks.
jordanbhec...@hotmail.com
Could you please send me a copy of the solutions for chapters 9 and 10 of berger and cassela.
I would really appreciate it.
John
I would love a copy of the solutions for the
Casella & Berger book.
I am trying to work through it and the solutions
would be a great help.
My email is jmu...@mindspring.com
Thanks greatly,
- john
John
for the Casella & Berger book.
I am trying to work through it myself and the
solutions to the exercises would be a big help.
Send to
kel
Thanks.
frank
My email is iip...@hotmail.com,
Thanks-
Franklin
sando33a
Could you please send me the solutions manual to Statistical Inference by Casella & Berger.
My e-mail is sand...@yahoo.com.
Thanks
sando33a
Thanks
Don
farzin4049
Thanks a lot ...
Jen
I'm currently have a problem with a question in chapter 3 of Casella and Berger.
I have no idea how to do question 3.20 (b), can anyone help me please(any hint would be good)
Thanks alot
JuanFangio
I am in dire need of your assistance regarding Casella and berger. Could you send me the solutions manual. At least chapters 9 and 10?
Our professor does not give answers to problem sets, so there is no way to learn!
Ravi Karsha
I am just finishing up a year long sequence and just now noticed this site. I have heard that some problems are unsolved in the solutions manual. I am beginning to prepare for comprehensive exams and I was hoping someone had solutions to the problems not in the solution manual that they would be willing to share/sell. I would be willing to pay a fair price for them.
Ravi Karsha
Reef Fish
Ravi Karsha wrote:
> Hi,
>
> I am just finishing up a year long sequence and just now noticed this site. I have heard that some problems are unsolved in the solutions manual. I am beginning to prepare for comprehensive exams and I was hoping someone had solutions to the problems not in the solution manual that they would be willing to share/sell. I would be willing to pay a fair price for them.
Why don't you offer your "fair price" to Casella and Berger?
I had only seen a couple of the problems posed in their book.
I thought they were TERRIBLY stated. and ambiguously impractical.
No wonder that's the only book students are always scrambling
for "solutions", in this newsgroup, ever since I started reading here
over a year ago.
BURN the BOOK. Learn your statistics from other books. Perhaps
then, you'll find the solutions to the exercises in Casella and
Berger's
book a trivial exercise once the subject is understood.
-- Bob.
John
Do you mind sending me the solution manual for Statistical Inference Casella & Berger?
I have just purchased the book for self-study and would like to have the solutions. My email address is jsmit...@gmail.com
Thanks!
Anma
Thanking in advance
sf...@hotmail.com
Kristin Oskal
Kristin B. Oskal, kbo...@start.no
Erase Ego
Erase Ego
zoco
zocoxzoco2 AT hotmail.com
sheldon Du
Thank you very much!!
Sheldon Du
Francis
I really enjoyed your comments on education and I understand your pessimism. I also understand the pressures the students are under to do well and to compete in todays job market. They think that they can best do this by getting A's and a good overall GPA and that coveted piece of paper. What I think maybe missing is instituting qualifying exams throughout the college years and not just for us Phd students.
After graduating from Columbia one post grad asked me "why should one be awarded a degree just for doing ones homework?"
That made me think if I really deserved my Masters degree and if I had what it takes to go on formy Phd. Then I realized that I am just a goldfish living in the environment I am supplied. I did what was asked of me to the best of my ability and I was thereby playing by the rules. And as you say until they change the rules students will learn what they are taught and expected to learn and no more.
I would love to thoroughly understand the material but my experience has been that most professors dont really care. I was told, during a recent survival analysis class, "you dont pay my salary. My reserach grants do that". So what are students supposed to do when faced with professors such as these? They are left holding the bag. Some students do their best by themselves and others seek "solututions manuals" to just get by. If you asked students who cheat"would you rather not" I expect the answer would be of course.
I have filled in surveys and such at school and since my graduation found that they are now instituting qualifying exams at the end of the masters program in Biostatistics. To be honest I am a little dissappointed that I was not there for this as I would cherish my degree a little more. What I dont know is if the exams et diluted as well and follow suit with the current trend towards just letting students pass since they have paid their way.
Anyway Bob sorry to bore you with my ramblings suffice to say that I wish their where more professors like you around and its a shame thst you let the system grind you down. I always feel its the few good students that should make it all worthwhile for professors but I know thats a cliche.
Maybe teach privately and help change the future....
Francis
Reef Fish
> Hi Bob:
> I really enjoyed your comments on education and I understand your pessimism.
Hi Francis.
I must preface my response with the disclaimer that I am not able to
place who
you are (because there are almost as many Francis Pike as there as
Jacob
Cohen <Rodney Dangerfield> in a thread earlier today) and I hope I am
not
being too presumptious to assume that you may be addressing me (even
though I had counted at least 55 Bobs in sci.stat.math) in view of
your
"pessimism on education" characterization, because I DID have dozens
and
dozens of posts last year (much less this year) along that line.
In any event, if your enjoyment of comments was intended for some other
Bobs, I'll accept it for them and offer my own comments. :-)
>I also understand the pressures the students are under to do well and to compete in todays job market. They think that they can best do this by getting A's and a good overall GPA and that coveted piece of paper.
How true. The "devaluation" of GPA has gone to such absudity that in
a recent article I read about the GPA of local high school graduates,
out of a POSSIBLE GPA of 4 (straight As), almost every one of the top
graduates had GPAs greater than 4, ranging up to 4.57, and that student
was going to attend college in some school I had never heard of, while
one or two of the lowly 4.1s were admitted to Yale and Harvard.
When I left the university "out of disgust" of many undergrad expecting
"A"s for "Attendance", not a single graduation had a top student with
GPA less than 4,0 (at least that was the upper bound in college). I
also served on the committee of awarding teaching and assistantships to
graduate students, Almost every applicant had GPAs great than 3.8 so
that information was essentially disregarded. Grades and GPAs are now
nothing short of a FARCE,
>What I think maybe missing is instituting qualifying exams throughout the college years and not just for us Phd students.
That has its pros and cons take much to explain.
> After graduating from Columbia one post grad asked me "why should one be awarded a degree just for doing ones homework?"
Which Columbia was this? The one in NY or Missouri or South Carolina
or South America? :)
I don't think the "automatic graduation" has reached the college level
yet.
> That made me think if I really deserved my Masters degree and if I had what it takes to go on formy Phd. Then I realized that I am just a goldfish living in the environment I am supplied. I did what was asked of me to the best of my ability and I was thereby playing by the rules. And as you say until they change the rules students will learn what they are taught and expected to learn and no more.
And some considerably less than what they are expected to learn. At
the Ph.D. level, I think the course grades are looked upon more
seriously, but once one graduates, NOBODY ever looks at the grades
anymore -- which is good thing, although I might have held the record
for having had the highest grades at the Yale Graduate School over a
period of 4 years. At that point, one's own initiative, originality
and creativity takes over. That's why some Ph.D. student perished
academically as soon as they received their Ph.D., while others are
just beginning their fruitful research.
> I would love to thoroughly understand the material but my experience has been that most professors dont really care. I was told, during a recent survival analysis class, "you dont pay my salary. My reserach grants do that".
That's very true, especially in this day and age where funding of
Higher Education is not nearly as generous as it "used to be" (say a
few decades ago.). It's no longer the QUALITY of the research that
counts anymore. It's easy for the Dean to just count grants in DOLLAR
amounts, and research universities are even ranked by the $ amount they
receive in research funding. Bean counters prevail! How many
mathematicians or statisticians do you know that has a $5 million
grant? Or even a $1M research grant? The Engineering school, the
Computer Science, physics, chemistry, and other equipment-intensive
departments routinely get grants that are in MILLIONS of dollars (on
equipment alone). In one case at the university I left, the textie
Department got a $2M grant from the State's textile industry, for doing
essentially "consulting" work, exploiting the labor of grad students
and faculty members in that department, with no element of "real
research" involved. They got top billing over dozens of REAL
researchers in other departments who could scrape up $1M total to show.
:-)
The word "prostitution" had been heard often, about the faculty members
going after $$ for worthless research over no dollar for genuine
research because of the system of institutional awards for merit badly
misplaced.
>So what are students supposed to do when faced with professors such as these? They are left holding the bag. Some students do their best by themselves and others seek "solututions manuals" to just get by. If you asked students who cheat"would you rather not" I expect the answer would be of course.
>
> I have filled in surveys and such at school and since my graduation found that they are now instituting qualifying exams at the end of the masters program in Biostatistics. To be honest I am a little dissappointed that I was not there for this as I would cherish my degree a little more. What I dont know is if the exams et diluted as well and follow suit with the current trend towards just letting students pass since they have paid their way.
Having a qualifying exam for the Master's degree is rare, to the best
of my knowledge. When I was at Yale decades ago, the Ph.D. qualifying
exam (to admit students for the Ph.D. program) was (and may still be)
used as the qualification for the M.Phil (Master of Philosophy) degree
for those who failed. :-) They were told to pack up and leave after
the first or second year. So, the M.Phil was the consolation price.
But for universities in the top Ivy League schools, the ENTRANCE
qualification was much more stringent than other universities, and once
admitted, a Master's degree is almost automatic.
>
> Anyway Bob sorry to bore you with my ramblings suffice to say that I wish their where more professors like you around and its a shame thst you let the system grind you down.
Thank you for your compliment and sympathetic ear. Your post didn't
bore me at all. It's one of my favorite subjects. The system didn't
really grind me down -- no more so than a few in this newsgroup who are
MUCH worse than some the worst undergrads I've encountered there. I
simply held the line until I chose NOT to "sell my soul to the Devil"
and walked off proudly and voluntarily. It was the SYSTEM. The
SOCIETY. Those are the City Halls that are impossible to fight
successfully single-handedly. But I tried. :-)
> I always feel its the few good students that should make it all worthwhile for professors but I know thats a cliche.
That's not a cliche at all! The few good students I had at both the
undergrad and Ph.D. level made it all worthwhile to put up with all hte
abuses of Higher Education, orchestrated by SOME students and the
cooperating Aministrators euphemistically called "educators".
I told the story that I quit the Statistics profession in such a "cold
turkey" manner in 1999 that I didn't even know John Tukey had died when
I stopped by the Annual Statistics Meeting in 2004 to see FOUR of my
former Ph.D. students (a Dean, a Dept Chair, a Full Prof, and an Assoc.
Prof at major universities) all presenting papers at the same Meeting
in San Francisco, just to see THEM. When I saw the Memorial Lecture of
Tukey in the Annual Program, I asked Dave Hoaglin "When did Tukey pass
away?" He said with a straight face, "two years ago". :-) Dave
knew me and my Ph.D. students well to know that I had quit "cold
turkey" so that it was no surprise that I wasn't aware of a "cold
Tukey".
>
> Maybe teach privately and help change the future....
I started to ease back into a LITTLE bit of statistics by taking part
in sci.stat.math in 2005. The rest is history. The level of the
statistics "professionals" are so dismally low that no private teaching
will help IMHO, and all I am hoping to do in this forum is to raise the
LOWEST common denominator by an inch or two, so far with very little
success.
Richard Ulrich is the classic case of the FAILURE of our educational
system. No ifs and buts about it. Perhaps you should stay a little
while and try your hand at teaching him why a MODE is not an average,
or why it is not acceptable to say type-1 and type-2 errors are
numbers, or what a linear model is, and why he made major blunders and
errors in at least a dozen other ELEMENTARY STATISTICAL, while he is
here to TEACH OTHERS (while flaming me for correcting him) of his
Quackery.
Perhaps that will provide you with the first hand experience about
EDUCATION in general, and statistical education in particular, by
taking this forum as a microcosm of the statistics world at large.
Thanks for your sympathetic ear on the subject of EDUCATION.
That gave me a golden opportunity vent some air against some of my pet
peeves, with an actual CASE study of the FALIURE of our educational
system to point to.
Hope to see you around, Francis, if you can stand the heat and the
boredom. :-)
>
> Francis
Herman Rubin
Reef Fish <Large_Nass...@yahoo.com> wrote:
>Francis wrote:
>> Hi Bob:
>> I really enjoyed your comments on education and I understand your pessimism.
....................
>>I also understand the pressures the students are under to do well and to compete in todays job market. They think that they can best do this by getting A's and a good overall GPA and that coveted piece of paper.
This is unfortunate; it tends to get students to memorize
and regurgitate, and forget.
>How true. The "devaluation" of GPA has gone to such absudity that in
>a recent article I read about the GPA of local high school graduates,
>out of a POSSIBLE GPA of 4 (straight As), almost every one of the top
>graduates had GPAs greater than 4, ranging up to 4.57, and that student
>was going to attend college in some school I had never heard of, while
>one or two of the lowly 4.1s were admitted to Yale and Harvard.
This is the only way high schools have to get students to
take the "honors" courses, which are not up to the standards
of the old "college preparatory" courses. Good students
were taking the junk to raise their GPA, even though that
was not always the case. However, if teachers were "grading
on the curve", it would be.
>When I left the university "out of disgust" of many undergrad expecting
>"A"s for "Attendance", not a single graduation had a top student with
>GPA less than 4,0 (at least that was the upper bound in college). I
>also served on the committee of awarding teaching and assistantships to
>graduate students, Almost every applicant had GPAs great than 3.8 so
>that information was essentially disregarded. Grades and GPAs are now
>nothing short of a FARCE,
You are SOOOOOO right.
>>What I think maybe missing is instituting qualifying exams throughout the college years and not just for us Phd students.
>That has its pros and cons take much to explain.
There are two problems. The schools at each level will
have to state what they expect the students to be able
to do with their knowledge, and not just memorize. In
fact, many, if not most, parts of examinations should
be open book.
>> After graduating from Columbia one post grad asked me "why should one be awarded a degree just for doing ones homework?"
>Which Columbia was this? The one in NY or Missouri or South Carolina
>or South America? :)
>I don't think the "automatic graduation" has reached the college level
>yet.
Alas, it essentially has. Personally, I do not count
homework, except for take-home problems which are really
problems. Homework should be for learning, not grading.
.................
--
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, Department of Statistics, Purdue University
hru...@stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558
Old Mac User
I was under the impression that you were teaching at a major
university... didn't know you quit much less "cold turkey".
Just fyi I quit my job in industry "cold turkey", leaving an exaulted
position because I felt I had worn out my welcome
(after 40 years). That's a long story, but part of it was rooted in
the current attitude toward computers, software,
and Six-Sigma. Basically... it's about "don't think, just buy software
and learn which buttons to push." That,
combined with long boring "staph" meetings discussing things we had no
control over in the first place. Other
than getting paid for my efforts, all I expect of my clients is that
they listen, consider, and use what I have to teach them.
I'd still be at my old company if they had done the same.
If you think there are some lost souls at sci.stat.*** you should deal
with some Six-Sigma "Master Black Belts" and other
belts of many colors. Minimal knowledge... maximum egos.
Aside... the Ulrich "mode" thing was so awful I could not bear to
follow the thread. The awful thing is RU is offering
advice to innocent people.
The other awful thing I see on sci.stat.*** is that so many are trying
to use exotic methods they do not understand,
and just want someone to tell them how to use software they have
apparently just "discovered". That's what was
happening when I "walked" and said "to heck with it."
I think I'm telling you that the differences between
universities/students/industry is not all that great. After all,
industry hires those students.
Good luck in your current situation. OMU
Scott Seidman
news:1153417535.8...@i42g2000cwa.googlegroups.com:
> The other awful thing I see on sci.stat.*** is that so many are trying
> to use exotic methods they do not understand,
> and just want someone to tell them how to use software they have
> apparently just "discovered". That's what was
> happening when I "walked" and said "to heck with it."
>
This is no different from any other area in which software accessibility
makes it easy to carry out algorithms that users do not understand. All
the time, I see people thinking that they can carry out advanced image
analysis simply because they have Matlab and the Image Processing Toolbox,
without ever bothering to learn any image processing (or, in fact,
bothering to learn matlab or the associated toolbox).
--
Scott
Reverse name to reply
Old Mac User
understandable English... but written in MS Word. Not to mention
multicolor PowerPoint Presentations where the conclusions really don't
connect with the data. We are beset with computer literate idiots.
Reef Fish
Old Mac User wrote:
> Reef Fish:
>
> I was under the impression that you were teaching at a major
> university... didn't know you quit much less "cold turkey".
You meant you thought I was CURRENTLY teaching at a major
university, I presume, rather than thinking I taught only at minor
universities because of the "tree stumps" and the faculty and
administrators I talked about. I had always taught at major
universities, nothing short of the Top 75 Tier 1 Ph.D. grating
universities.
I don't think I could have survived more than a few months in
the 3rd and 4th rate universities or those diploma mills
euphemistically called "community colleges". Sorry is I
stepped on 1 or 2 of the good folks in those environments.
> If you think there are some lost souls at sci.stat.*** you should deal
> with some Six-Sigma "Master Black Belts" and other
> belts of many colors. Minimal knowledge... maximum egos.
I met plenty of those in other ngs. But meeting someone who calls
himself a staistician, and was the ALL TIME most frequent poster
since 1994 in sci.stat.math, making MAJOR BLUNDERS in just
about EVERY aspects of statistics, and not only is not ashamed
of his unequivocal errors, but continue to RE-HASH his errors,
while making misdirections and excuses to post his ad hominem
attack on ... my ability to read, my ability to write, etc. ...
THAT totality and combination of his attributes puts your
"master black belts" in other groups to a saint-hood status. :-)
>
> Aside... the Ulrich "mode" thing was so awful I could not bear to
> follow the thread.
That is not nearly as awful as many of his performance in OTHER
threads. He even defended his claim (the only one in the HISTORY
of statistics I've ever heard -- that Y = b X is a NONLINEAR model!
> The awful thing is RU is offering advice to innocent people.
Of course THAT is the reason I am stopping him from getting away
with murder on his Quackery and malpractice!
Luis A. Afonso is different. He is virtually harmless to OTHERS.
He only makes a fool of himself. Nobody ever takes him
seriously (you tried a time or two when you were new here and
caught on).
But Richard is quite good at his misdirection and obfuscation
techniques (such as bring up old dead horses such as the
errors he made in "1 - type-2 error" in the thread about the
MODE, and used that to rehash his ad-hominem attack.
When I started in this newsgroup in the early part of 2005,
Richard Ulrich had already attracted a "fan club" of the
innocent blind to hurl insult at me whenever I correct
Richard's SERIOUS errors about all aspect of applied
regression analysis -- and to this day he is STILL clinging
on to those errors.
I couldn't careless about Richard Ulrich's total ignorance
in Statistics -- as long as he doesn't palm them off, in
a Statistics group, as if his ERRORS and black magic
are valid statistical methodology or interpretation.
> The awful thing is RU is offering advice to innocent people.
And REPEATEDLY defend and re-offer his erroneous advice.
> The other awful thing I see on sci.stat.*** is that so many are trying
> to use exotic methods they do not understand,
> and just want someone to tell them how to use software they have
> apparently just "discovered".
The popularity of software is almost as much to blame as the
people who abuse the usage of such software, Garbage in,
Garbage Out. That's Richard Ulrich's "expertise". :)
I am much more sympathetic to that kind of ignorance, and there
are many receptive learners, once they learned what they do is
WRONG, such as drawing causal inference from un-controlled
regression data, misinterpreting the meaning of regression
coefficients, and so on. But Richard Ulrich comes back, over
and over again, to repeat his SAME errors and malpractice,
each time writing longer and longer ad hominem attacks on how
poorly Bob (and he always has to add Ling for flaming effect
because folks in other groups who flame me thought I was
trying to hide my real identity) reads, writes English and how
Bob Ling has difficulty in his comprehension of this and that.
I don't think many people are left in thses groups that are behind
Richard Ulrich's malpractice anymore -- at least I haven't heard
many defenders of his lately.
> That's what was happening when I "walked" and said "to heck with it."
> universities/students/industry is not all that great.
in some respects of human nature. But the difference is MUCH BIGGER
than you might think. ONE single major error could kill an
academician's
career in a university environment. That's why one keeps his mouth
SHUT unless he is on sure ground (and that's one trait and motto I have
carried over from my academic life). You don't see me entering in any
discussion on many topics which I know well, but not sufficiently
expert
to feel that my advice is needed.
In industry, as evidenced by many who work there, errors are seldom
recognized or noticed by the "boss". So, there are those who got away
with "murder". Greg Heath left himself a LONG TRAIL proving how
somebody like him could have lasted that long in industry palming his
erroneous theory and methods. To his credit, he has finally given
up definding the indefensible and I've even seen him giving OTHERS
some SOUND advice based on what he learned in this group.
On the question of tolerance on statistical errors that are COMMON
I am the hard-nose person who advocates ZERO TOLERANCE,
on Statistical crimes.
Kevin puts it best as a statistician to another, "If we don't do it,
who will?"
> After all,industry hires those students.
And we are witnessing the vicious cycle of down spiral of tolerated
ignorance and malpractice.
> Good luck in your current situation. OMU
Thanks. The current situation, as dismay as it may seem, is
actually quite promising, in terms of "a small step for sci.stat.math,
a giant step for statisticis".
More and more READERS are becoming INVOLVED in stemming
out some of the worse abuses. Richard Ulrich's MODE and his
misuse of hypothesis-testing terms and concepts, and Afonso's
"sample variance" to name just two recent examples.
-- Reef Fish Bob.
Old Mac User
You wrote: "You meant you thought I was CURRENTLY teaching at a major
university, I presume, "
Yes, that's what I thought.
Thanks for your post. I'll add a couple of general comments on working
industry. Over the years I worked with some outstanding,
honest, and thoroughly competent people who inspired others and who
innovated and contributed. Then there were those who
were in attendance nearly every day, and who retired with "1 year of
experience 40 times". In the left tail of the distribution were
those highly skilled in corporate politics... survivors who even got
promoted to rather high positions before they flamed out.
When I retired I realized that I had worked under 17 "bosses". I
learned a lot from about four of them. Most of the others were
the stumps you mentioned. One of them fancied himself as skilled in
statistical methods, and often tried to "help" us. Several
persisted in advertising our abilities, but often missed the target and
caused us to divert into arenas that we had no business
being in. I was fortunate to be promoted to a position where I had a
license to work all over the corporation, and not just in one
division. Collectively, over 115,000 people out there and a staff of
five statisticians in my shop. That brought access to the
CEO, and even opportunities to "teach" the CEO and his direct reports
about once a year... usually for an hour or two.
The purpose of that was to make them aware of our capabilities, and
help divisional managers understand what we could do
for them. My experience with the highest level managers was (1) very
bright; (2) fast learners; (3) real human beings.
I have no complaints. At the end of 40 years I was tired, and really
had worn out my welcome with a large number of
mid-level managers. It was time to leave, and I did.
You mentioned the hazards of making just one mistake in academia.
Please know that the same is true in indsutry. However,
since people tend to shuffle around rather rapidly in industry,
memories tend to be shorter and new faces tend to be more
accepting of past faults. In mid-career I was asked to dig into a
major R/D project that was going badly. It was at our site...
a large R/D facility. The project was loaded with politics and
intrigue. I started with a team of 3 chemists... two withered in
the heat and copped out. Working with the remaining chemist we finally
saw what was wrong... wrong with the data.
It was sneaky. We never resolved whether it was deliberate or just due
to ignorance. I think it was ignorance. In any event
I blew the whistle on it. That was the final straw... end of the
project with heavy consequences. Nobody was fired, but
several people were given "special assignments". My supervisor told me
to take my wife and leave town for awhile. I knew
then that my career had taken a sharp turn in the road... relationships
with local management would never be the same.
That was correct, but I found other significant work to do in other
locations and was well-rewarded for my efforts.
I mention this only to say that working in industry, even in the 1980s,
can be risky. I have a lot of sensitivity for statisticians
who "tell the unpopular truth" in ethical drug companies and other
places of that sort. The fact is if "it's not working" or if
"there are dangerous side effects" it's far better to get that out on
the table right now rather than end up sitting in a large
oak chair in a courtroom with lawyers squeezing it out of you.
My wife just announced dinner is ready.
I do hope things will continue to go well for you. Also, please keep
chugging away at sci.stat.math etc. I do believe
you have made a positive difference. OMU
Reef Fish
Old Mac User wrote:
> Reef Fish...
>
> You wrote: "You meant you thought I was CURRENTLY teaching at a major
> university, I presume, "
>
> Yes, that's what I thought.
If I were, I wouldn't have so much time teaching for free what some
students
paid $30,000 tuition a year to listen. :-)
> Thanks for your post. I'll add a couple of general comments on working
> industry.
Only a couple minor points of clarification on some points.
< snip >
> When I retired I realized that I had worked under 17 "bosses". I
> learned a lot from about four of them. Most of the others were
> the stumps you mentioned.
Remember the above when you applied it to this:
> You mentioned the hazards of making just one mistake in academia.
> Please know that the same is true in indsutry.
That's ESSENTIAL difference between academia and the industry.
In academia, there are too many watchful eyes for mistakes.
In the industry, because of the "stumps", people can make many
major blunders and get away. Example: Greg Heath and many
others in this newsgroup who work in the industry and make
many grave mistakes here.
Richard Ulrich reached the height of his academic career (at
about 58) with a master's degree and worked as an Assistant
Professor doing some clerical work in a Psychiatry Department,
and claim himself to be a statistician. I think he lost that job.
But if he WERE in the WORST of the Departments of
Statistics in which I had taught, he would not have been hired
in the first place and if hired, would not last more than a few
month before he becomes the laughing stock of the entire
department and the university. In the statistical profession?
There would be absolutely no place for him.
But in this newsgroup and in the industry? He even attracted
some praises for his foolishness! And he keeps on doing his
malpractice unchecked by his boss, if he is still employed.
THAT's the difference.
> CEO, and even opportunities to "teach" the CEO and his direct reports
> about once a year... usually for an hour or two.
Some of the brightest and best-motivated CEOs enroll in Executive
Programs in business schools. I taught a couple of those courses at
the Grad School of Business at Vanderbilt.
> My experience with the highest level managers was (1) very
> bright; (2) fast learners; (3) real human beings.
That's how they get to be top level managers and CEOs, but when
it comes to statistics, most of them are NOT fast learners or even
good learners. But they are the most appreciative and MATURE
human beings when they recognize it wasn't my fault (as most
immature students always blamed it when they get a bad grade)
that they had difficulty with the subject. That was the ONLY class
in which a student (who knew that his grade would not be better
than a C or D) thank me profusely in the "teacher evaluation" form
to be seen by me AFTER the course (so it wasn't done to try to
bribe a grade) and even wrote me a personal letter thanking my
effort after his grade of a D. Other Executive Program students,
regardless how well or how badly they learn the subject of
statistics, were always MATURE enough as a human being to
appreciate my teaching EFFORTS and when I had to teach, for
THEIR benefit.
Teaching the students in that Vanderbit Exec Program was one
of my most rewarding experiences. I also taught the SAME course
to their reguar MBA students.
This is the contrast. Some of the students in the regular MBA
class complained bitterly because of the grades they EARNED
(lower than what they thought they deserve). The Exec Program
class, as a whole, got worse grades, but they not only did not
complain, but showed their genuine appreciation. SEVEN years
later, in their reunion of their graduating class, while I was on
my sabbatical leave at Harvard, the class voted to invite me
to their reunion by paying all my travel expenses from
Cambridge to Nashville for their reunion which I enjoyed
immensely -- a highlight of my academic career, better than
any honor or award I had received.
In retrospect, I felt I might even have been too demanding on
that class, but they not only responded to the challenge, but
shook off their bad grades to show their genuine appreciation.
THAT's the kind of personal CHARACTER some individuals
show. They are all too rare in academia, and especially in
non-academic groups like this where some love to bite the hands
that feed them FOR FREE. :-)
>
> My wife just announced dinner is ready.
Bon appetit.
>
> I do hope things will continue to go well for you.
Thank you, OMU. I am a man of many interests (I call each
vastly different one a LIFE). I have already lived 11 LIVES.
This one may not be able to sustain my interest much longer.
But as long as time permits, I'll try to contribute positively to
the statistics profession and those in this group, notwithstanding
the lack of appreciation or the constant FLAMES from some.
Used to them. :-) Been there. Done that. As I always say.
Herman Rubin
Old Mac User <chendr...@juno.com> wrote:
>Putting it all together, I get technology reports written in barely
>understandable English... but written in MS Word. Not to mention
>multicolor PowerPoint Presentations where the conclusions really don't
>connect with the data. We are beset with computer literate idiots.
They are NOT computer literate. They know how to do tricks,
and proceed to do them. They are almost like the monkeys
at the typewriters, except they use words (sometimes) and
even manage to occasionally produce coherent sentences.
When it comes to statistics, almost all are trained, not
educated. To be educated in statistics, one has to
understand the concepts of probability, and that any
statistical method has to balance errors. Also, do
not just use a method you know; physicists have asked
me how many intervals to use for a chi-squared test
for bump hunting, and a chi-squared test has almost
no power compared to Kolmogorov-Smirnov or Kuiper.
The same is true in medical statistics. In my opinion,
few if any of the tests on which statins are being urged
are sound, and those doing the tests have no idea how
to use quantitative information.
>Scott Seidman wrote:
>> "Old Mac User" <chendr...@juno.com> wrote in
>> news:1153417535.8...@i42g2000cwa.googlegroups.com:
>> > The other awful thing I see on sci.stat.*** is that so many are trying
>> > to use exotic methods they do not understand,
>> > and just want someone to tell them how to use software they have
>> > apparently just "discovered". That's what was
>> > happening when I "walked" and said "to heck with it."
>> This is no different from any other area in which software accessibility
>> makes it easy to carry out algorithms that users do not understand. All
>> the time, I see people thinking that they can carry out advanced image
>> analysis simply because they have Matlab and the Image Processing Toolbox,
>> without ever bothering to learn any image processing (or, in fact,
>> bothering to learn matlab or the associated toolbox).
--
David Winsemius
$23...@odds.stat.purdue.edu:
> The same is true in medical statistics. In my opinion,
> few if any of the tests on which statins are being urged
> are sound, and those doing the tests have no idea how
> to use quantitative information.
>
I wonder if you would be willing to expand on that opinion? For years the
anti-cholesterol brigade pushed nostrum after nostrum. I was a skeptic.
Then the statins (HMG-CoA reductase inhibitors) came along. It was hard to
argue against randomized trial data where the primary endpoints were met.
But apparently you see flaws. I am open to education and persuasion.
--
David Winsemius
Herman Rubin
Not one of those trials considered the effect of such
things as HDL and CRP levels. I BELIEVE it will be found
that most of those with low HDL and high CRP, or some
combination of them, will have a greater than average
benefit from statins, but those I have mentioned will do
worse; this is on a guess from the studies which have
partial breakdowns, and other sources.
We have at least three quantitative variables here, and
the great majority of the studies are not quantitative
on any of the variables, at most using arbitrary breakpoints.
This is a problem of multivariate analysis, with at least
three causal variables not interacting linearly.
Old Mac User
"They are NOT computer literate. They know how to do tricks,
and proceed to do them. They are almost like the monkeys
at the typewriters, except they use words (sometimes) and
even manage to occasionally produce coherent sentences.
When it comes to statistics, almost all are trained, not
educated. To be educated in statistics, one has to
understand the concepts of probability, and that any
statistical method has to balance errors. Also, do
not just use a method you know; physicists have asked
me how many intervals to use for a chi-squared test
for bump hunting, and a chi-squared test has almost
no power compared to Kolmogorov-Smirnov or Kuiper.
The same is true in medical statistics. In my opinion,
few if any of the tests on which statins are being urged
are sound, and those doing the tests have no idea how
to use quantitative information."
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Well, in that sense you are right... they are NOT computer literate.
They move a lot of bits and bytes around, create a lot of files,
make a lot of graphics, and generally give the impression of
being highly productive. From my perspective one well-chosen
graphic beats presenting slide after slide... showing "the data" from
every angle. Consider this one. Shortly after we got desktop computers
(leaving punched cards behind) I came across a report in which certain
experimental data were presented in a table. There were 4 columns and
15 rows of data in that table.. a total of 60 numbers. These data were
presented in 66 graphics... apparently using every possible scale and
combination of views. Some graphics were actually the same as others,
but the axes had been rotated. Some graphics were the same as
others, but with bits and pieces of data omitted. This was not the only
example of graphics run amuck.
I like your chi-squared example. Well, I don't really LIKE it. It's
just
fascinating. The people who really drove me mad were those who
"created new statistical methods" and then published them in our
R/D technology reports. Their gambit was to entice others to join their
club and use their new "discoveries". Some of these were cleverly
done,
and worthy of publication in the Journal of Irreproducible Results.
The award-winning "best of breed" was a nasty piece of software
written by a chemist. It was "capable" of fitting a 5th-order
polynomial
with just four "points" ordered along the "X-axis". His purpose was to
"fit" certain physical property data, a common practice. This one was
presented
in a technical conference (I was not invited) and was touted as a
remarkable
way of "getting the work done with less data." Then there was the
idiot
that created an "equation" that had at least 56 "fitting constants"...
all
done with just ten rows of data (ten experiments"... truly a miracle
when
the R-sq was only 0.93. This one involved regression on principle
components.
Worse, he'd brought in a consultant from a big-name university
("Chemometrics"
was his thing) to bless it.
But now I come to the medical statistics matter. Since I take Lipitor
and it has greatly reduced my "bad cholesterol" (LDL currently running
near 67)
I'm curious to hear more. True, the linkage between heart attacks and
"bad cholesterol" has not been fully confirmed. But I thought that the
evidence for statins as cholesterol reducing agents was solid. By the
way, I also take a different drug (Niaspan... niacin) to increase the
"good cholesterol" which it has done very nicely (HDL = 42). I
recognize
the potential for liver and muscle damage, so tests are run for liver
function
etc. at least every six months. So far, so good. Anyhow, I'd like to
hear
more of your thoughts on this.
Since I do some work for several "medical device" companies and also
for
some bio-startups, I keep a watchful eye on claims for "success is at
hand"
in the media. The NY Times is guilty of publishing too many of these
prematurely, and I now have the ear of one of the editors there about
this.
I am keenly aware of some of the awful failings in the medical industry
(greed... greed... greed...). I blame the media for much of this.
Comments...?
OMU
jp
I would like to have a copy of the solution to Castella and Berger. jp...@virginia.edu
David Winsemius
Apotlogies for my delay in geting back. I probably should not have posted
just before a vacation.
A few questions for clarification of your meanings:
Q: In your earlier criticism stating "few if any of the tests on which
statins are being urged are sound" were you referring to the conduct of the
drug trials or to the manner in which decisions are being made clinically
by physicians in practice? I am going to proceed on the assumption that you
were criticising the former.
Q: When you say "those I have mentioned will do worse", which groups had
you "mentioned", (or was this an implicit contrast to the low HDL/high CRP
subgroup that you thought would do better)?
Q: Am I correct in thinking the three variables are a) HDL levels
(generally involving a test population with a truncated cholesterol
distribution by design), b) CRP levels (skewed by biology and probably
additionally skewed by the design cutoffs for lipid other risk factor
values), and c) statin administration?
Q: Are you implicitly criticising cardiologists' focus on LDL levels rather
than HDL levels?
Q: Which effect scale are you using when you suggest non-linear
interactions? Risk difference, risk ratio?
The whole point of a randomized trial is to create a situation where causal
factors (both measured and unmeasured) get randomly (most likely equally)
distributed. If there is evidence of unequal distribution of factors
between treatment groups, then multivariate statistics are available for
post hoc adjustment. We would both prefer to see a full predictive model
(including interaction terms) describing the probability of an event
distributed over time. I share your frustration that sometimes the authors
only report a risk difference or a risk ratio and then walk away with a big
smile and lots of invites for presentations. However, the unpleasant
reality is that they really do not have a huge responsibility to report
multivariate analyses unless there is sufficient evidence of a problematic
imbalance.
The question the trialists addressed was whether it is better to administer
a statin with a goal of preventing either mortality or a cardiac event in a
group at high (or perhaps medium) risk. Admittedly it would be good to
check whether the drug had differential effects on particular sub-groups,
and I do know that such post hoc analyses were done for women, older
persons, and diabetics. I had a hard time believing that no one has looked
at the homogeneity of effect of statin use over the range of HDL levels in
the randomized trials. I found one article(1) published a year after 4S
article(2) appeared.
----
Homogeneity of effect (on ratio scale at least) across quartiles of LDL:
1)http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?
db=pubmed&cmd=Retrieve&dopt=AbstractPlus&list_uids=7746058&query_hl=11
&itool=pubmed_docsum
2) 4S:
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?
db=pubmed&cmd=Retrieve&dopt=AbstractPlus&list_uids=7968073&query_hl=11
&itool=pubmed_docsum
----
I have an easier time believing that CRP levels may not have been examined,
but here you are apparently using your personal belief that CRP is an
established causal factor in determining either risk or response. The level
of evidence appears equivocal for that position. CRP has a lot of factors
affecting it (age, gender, some cancers, recent infections, diabetes,
systemic inflammatory conditions), so it is both more volatile than HDL
cholesterol and has the potential for being a confounder rather than a
causal ("state") marker. Some, but not all, studies have found an
association with cardiac mortality risk. It is generally considered an
investigational marker rather than a well-established risk marker.
So I remain unclear regarding the substance and the persons responsible for
that still appear for vague, non-specific faults.
-Appears you feel that insufficient effort has been made to study
heterogeneity of statin effects across combinations HDL levels and CRP
levels.
-Is your critique that post hoc analyses of lipid interaction with
treatment assignment used quartiles rather than fitting a curve (perhaps a
spline or other tractable curve)?
-Are you criticising the emphasis on LDL levels at the expense of attention
to HDL levels.
-(It does not sound as though you believe (as I originally thought you were
saying) that statins have no demonstrated value.)
-It also appears that you may think that an un-named group of
biostatisticians display evidence of inability to use quantitative
information. Is that "inability" a blanket indictment of all the
coordinating centers in the large trials: (Scandinavian Simvastatin
Survival Study = 4S; West of Scotland Coronary Prevention Study = WoSCoPS;
Cholesterol and Recurrent Events study =CARE); Air Force/Texas Coronary
Atherosclerosis Prevention Study = AFCAPS/TexCAPS; Lipid Research Clinics
Coronary Primary Prevention Trial = LRC-CPPT; Helsinki Heart Study = HHS)?
If you prefer to have any portion of the discussion off-group, I am open to
that mode as well. I will not repeat any material received in that manner
without permission. My email address is a simple concatenation of my first
initial and last name and my domain is comcast <dot> net.
--
David Winsemius
Greg Heath
> Old Mac User wrote:
> > Reef Fish:
> >
> In industry, as evidenced by many who work there, errors are seldom
> recognized or noticed by the "boss". So, there are those who got away
> with "murder". Greg Heath left himself a LONG TRAIL proving how
> somebody like him could have lasted that long in industry palming his
> erroneous theory and methods.
Here you go again.
1. You don't have the slightest idea of what I did or did not
accomplish
in industry. During my 27 year tenure at my last gig my 3 bosses were
exceptional Ph.D.s in Physical Chemistry, Engineering and Physics,
respectively. Significant errors never got past them and my eagle-eyed
colleagues.
2. For some reason you went ballistic at my suggestion of looking
at the all-variable correlation coefficient matrix when
multicollinearity was
indicated for a particular problem. From your rantings and ravings it
is
obvious that you misread that simple suggestion (which I still
recommend
in that situation with a small number of predictors) for a necessary
and
sufficient edict.
> To his credit, he has finally given
> up definding the indefensible and I've even seen him giving OTHERS
> some SOUND advice based on what he learned in this group.
To your credit, you have read my later posts more slowly and tried
to understand what I was posting.
Hope this helps.
Greg
Greg Heath
Reef Fish wrote:
> Old Mac User wrote:
> > You mentioned the hazards of making just one mistake in academia.
> > Please know that the same is true in indsutry.
>
> That's ESSENTIAL difference between academia and the industry.
> In academia, there are too many watchful eyes for mistakes.
> In the industry, because of the "stumps", people can make many
> major blunders and get away.
And then there are places like MIT Lincoln Laboratory where the
quality of work routinely exceeds that in academia.
> Example: Greg Heath and many
> others in this newsgroup who work in the industry and make
> many grave mistakes here.
The mistakes I've made here were neither many nor grave.
Try sticking to the truth (assuming you could possibly separate
your delusions from the truth).
Hope this helps.
Greg
P.S. If you're wondering why this reply was so delayed, I was
on vacation when this was written and just happened on it today.
Reef Fish
Greg Heath wrote:
> Reef Fish wrote:
> > Old Mac User wrote:
> > > Reef Fish:
> > >
> -----HUGE SNIP
> > In industry, as evidenced by many who work there, errors are seldom
> > recognized or noticed by the "boss". So, there are those who got away
> > with "murder". Greg Heath left himself a LONG TRAIL proving how
> > somebody like him could have lasted that long in industry palming his
> > erroneous theory and methods.
>
> Here you go again.
You've done much better lately. But the trail WAS long, the way you
insisted on using simple correlations for the job they can't do and
then
blamed it on your industry and your boss.
>
> During my 27 year tenure at my last gig my 3 bosses were
> exceptional Ph.D.s in Physical Chemistry, Engineering and Physics,
> respectively.
Many of those don't know anything about statistics.
> Significant errors never got past them and my eagle-eyed colleagues.
But how would they know the errors you made in STATISTICS?
>
> 2. For some reason you went ballistic at my suggestion of looking
> at the all-variable correlation coefficient matrix when
> multicollinearity was
> indicated for a particular problem. From your rantings and ravings it
> is
> obvious that you misread that simple suggestion (which I still
> recommend
> in that situation with a small number of predictors) for a necessary
> and
> sufficient edict.
That's exactly the old stuff. NOW you know better. So, there is
point to rehash.
>
> To your credit, you have read my later posts more slowly and tried
> to understand what I was posting.
That's because you no long make the same mistakes you had made
earlier -- which is GOOD! Some folks here like Richard Ulrich never
learned from all his mistakes in regression and just keep on making
them and arguing about them.
-- Reef Fish Bob.
Greg Heath
Reef Fish wrote:
> Greg Heath wrote:
> > Reef Fish wrote:
> > > Old Mac User wrote:
> > > > Reef Fish:
> > > >
> > -----HUGE SNIP
> > > In industry, as evidenced by many who work there, errors are seldom
> > > recognized or noticed by the "boss". So, there are those who got away
> > > with "murder". Greg Heath left himself a LONG TRAIL proving how
> > > somebody like him could have lasted that long in industry palming his
> > > erroneous theory and methods.
> >
> > Here you go again.
>
> You've done much better lately. But the trail WAS long, the way you
> insisted on using simple correlations for the job they can't do and
> then blamed it on your industry and your boss.
Ho-Hum (I used this instead of "Bull Pucky" because I'm a nice guy).
I did neither. Just two more fabrications substantiated only
by your delusional imagination.
> > During my 27 year tenure at my last gig my 3 bosses were
> > exceptional Ph.D.s in Physical Chemistry, Engineering and Physics,
> > respectively.
>
> Many of those don't know anything about statistics.
Obviously you didn't know these 3.
> > Significant errors never got past them and my eagle-eyed colleagues.
>
> But how would they know the errors you made in STATISTICS?
Errors are errors. Most in-house briefings to Group leaders were
attended by colleagues. The diversity of backgounds was usually
sufficient to cover ANY technical errors.
> > 2. For some reason you went ballistic at my suggestion of
> > looking at the all-variable correlation coefficient matrix when
> > multicollinearity was indicated for a particular problem.
> > From your rantings and ravings it is
> > obvious that you misread that simple suggestion (which I still
> > recommend in that situation with a small number of predictors)
> > for a necessary and sufficient edict.
>
> That's exactly the old stuff. NOW you know better.
Me? YOU are the one who misinterpreted what I wrote.
> So, there is point to rehash.
I assume you meant 'no' point.
I beg to differ. I don't appreciate you continually bringing up my
name w.r.t. statistical errors when the only errors made were
your misinterpretation of what I wrote.
> > To your credit, you have read my later posts more slowly and tried
> > to understand what I was posting.
>
> That's because you no long make the same mistakes you had made
> earlier -- which is GOOD! Some folks here like Richard Ulrich never
> learned from all his mistakes in regression and just keep on making
> them and arguing about them.
Again, the mistakes are in your head, not in the archives.
Greg
Reef Fish
Greg Heath wrote:
> Reef Fish wrote:
> > Greg Heath wrote:
> > > Reef Fish wrote:
> > > > Old Mac User wrote:
> > > > > Reef Fish:
> >
> > insisted on using simple correlations for the job they can't do and
> > then blamed it on your industry and your boss.
>
> Ho-Hum (I used this instead of "Bull Pucky" because I'm a nice guy).
When you make more new mistakes I'll let you know. Okay?
> I don't appreciate you continually bringing up my
> name w.r.t. statistical errors when the only errors made were
> your misinterpretation of what I wrote.
You should have let your dead dog lie. It's all in the archives.
>
> > > To your credit, you have read my later posts more slowly and tried
> > > to understand what I was posting.
> >
> > That's because you no long make the same mistakes you had made
> > earlier -- which is GOOD! Some folks here like Richard Ulrich never
> > learned from all his mistakes in regression and just keep on making
> > them and arguing about them.
You cannot undo your errors that have clearly been documented in the
archives by trying to talk yourself out of it as you TRIED (how you
tried,
to make all the excuses under the sun) to no avail.
Just be happy with your new learned knowledge, and get over your
past sins.
-- Reef Fish Bob.
Greg Heath
Keep my sins out of it. They have nothing to do with statistics.
I'm content to leave the story with the archives. I just don't
appreciate my name popping up now and then in your posts
just to get a rise out of me.
Greg
nick
~nick
nick
nick
nick
nick
nick
boilermci
I am a grad school student and currently studying the book by myself. So I would really appreciate it if you could send me a copy of the solution manual. My email is pgon...@yahoo.com.
Thanks a lot!
Regards,
Peng
riflerange
Thank you soooo much!