Re: Statistics in Psychology?


Brett Magill wrote:
On Mon, 19 Jun 2006 10:53:29 -0700, Reef Fish wrote:


Reef Fish,

Isn't it also true that there are various levels of users of statistics?
Is it really necessary to have a PhD in Statistics to apply statistics to
discipline specific problems?

Of course not! I cited the Harvard Program as an "extreme example"
of the requirement for a top-notch researcher in medicine using
statistical methods. I also gave what I considered to be the
minimal requirement of "a good UNDERGRADUATE training in
statistics" which is not met by many posters in these sci.stat. ngs.

RF> It can change ONLY when people in those disciplines take
RF> Statistics much more seriously in both their INSTRUCTION
RF> and PRACTICE -- as Harvard being the extreme example.


You often use MDs as an analogy for the
training required by a compotent statistician.

Not exactly. See my original post or the paragraphs above.

Though my memory is by no means photographic, I usually do pretty well at
concepts. The quote that follows is from the thread below.

http://groups.google.com/group/sci.stat.math/browse_thread/thread/3e6eca59991a84dc/140ad288671ae844?lnk=st&q=%22Reef+Fish%22+Medical+Doctor&rnum=8&hl=en#140ad288671ae844

From that thread, Reef Fish said:
That is one of the reasons WHY a competent USER of statistical
methods require as much specialized training as a medical doctor
or a neuro-surgeon. That is also the reason WHY there is a
joint-program at Harvard for MEDICAL researchers that make
serious use of statistics to have BOTH a Harvard M.D. degree
AND a Harvard Ph.D. degree in Statistics. I had one Harvard
M.D. in that program in my Data Analysis course for the
Statistics Ph.D. program when I taught there. The M.D. prepares
the research for the medical knowledge; while the Ph.D. prepares
the same researcher for doing the statistical analysis properly
to leave little room for the "Garbage in; Garbage out" sydrome.

I don't think it is the only time you expressed that sentiment, regarding
a competent USER.

You cited me correctly and also correctly interpreted my sentiment
that Statistics require as much training as a medical doctor or a
neuro-surgeon (I have also used "brain surgeon" at other times).

The only thing to be quibbled there is the word USER. I had already
clarified in the posts you cited before the preceding paragraph that
the Harvard joint PhD.-M.D. USER is the extreme of the medical
RESEARCHER (which is a USER).

I had also given what I considered to be other lesser USERS (but
those conducting research on their own nevertheless) to require
minimally the equivalent of a GOOD undergrad degree in statistics,
covering the BASIC and CORE material of common statistical
practice.

Then there is the lowest level of USERS, such as nurses and
some physician assistants -- whose only required training are
those minimal tasks to get certain routine statistical chores or
record keeping done.

So, a USER covers an entire spectrum of CONSUMERS of
statistics -- from the top notch researcher to the low level go-fers.
As for being a good RESEARCHER in statistics, I would go as
far as saying, as I believe Kevin Thorpe said in this thread, that
a Ph.D. in statistics is neither necessary NOR sufficient!

I actually had ONE student in the Harvard Data Analysis course
who was a "star" student in their Ph.D. Program in Statistics --
an all theory and no application type who was Herman Chernoff's
student, with a huge Fellowship. But he had never seen DATA.
And when he was required to find SOME data (any data) for
a regression project, he couldn't think of ANY, and came up with
some data from a textbook that was completely INAPPROPRIATE
for regression. I am sorry I have to report this piece of dirty
laundry to make the point that a Ph.D. in Statistics from Harvard
may be completely incompetent as an APPLIED statistician.
I relunctantly gave said student a "gentleman's F", a "C", in
that course, having discussed the matter with Chernoff and
Rubin (who was chair of the Department at the time).

That student was of course an extreme outlier in ANY Ph.D.
program in statistics. But I can assure everyone that I have
seen all too many PhDs in Statistics who had never published
any paper NOR capable of doing a creditable statistican
application.

So, in that respect, the PhD degree in Statistics, from ANY
department, merely is a very broad and general way of
expressing some LEVEL of good research OR application.


So, I was referring to what you had expressed outside
of the context of the current thread. It does seem to conflict with your
current statement. Perhaps there was more to the context. I will accept
your clarification that preceded.

In that case, my further clarification above should kill all birds,
if you get the gist of what I was trying to convey -- something
that is not easily quantifiable, but the main POINT is that one
needs AS MUCH training to be a GOOD statistician (in whatever
sense of the word GOOD) as it takes one to get a M.D. degree,
or a degree in Law, or any of the other fields in which there
are CERTIFICATIONS before one is allowed to practice.

Alas, that's still another area that puts statistics in its current
state of affairs -- ANYBODY can call himself a statistician and
practice statistics -- even they are completely incompetent and
you can hear their "quacks" from miles away.


It was in reference to this that I was
making the MD, NP, PA analogy to statistical users.

The website I referenced that you could not access made a similar point,
using a craftsman, technician, technologist, engineer, scientist hierarchy
as an analogy to levels of statistical practice.

Again, whatever is discussed there is unlikely to be beyond the
spectrum of different LEVELS of training and competence I had
covered in my several posts in this thread (and others).


I hope you are enjoying your travels. Nha Trang looks beautiful. Are the
Thanh Long (Green Dragon fruit) orchards in bloom?

We are now "at sea" to my home away from home, Hong Kong,
where we'll have just enough time to have lunch with my
mother and brother who still live there.

My wife took a tour of the countryside of Nha Trang while I took the
low-brow tour of the Dam Market (an incredible Flea Market of
hundreds of stalls to find just about any kind of junk that exists in
Vietnam!). :-)

-- Bob.

.

Relevant Pages

  • What Medical Researchers Might Learn About Statistics From the CIA
    ... outside of statistics departments, that is. ... I was approached by a researcher who believed that high ... prospect of promotion, so I smiled and said fine, and we ran the analyses ... Run the analyses separately within each risk group, ...
    (alt.support.mult-sclerosis)
  • Re: Statistics in Psychology?
    ... That is why the Harvard joint medical-statistical Program, ... Harvard Ph.D. in the Statistics Department! ... RF> It can change ONLY when people in those disciplines take ... RF> and PRACTICE -- as Harvard being the extreme example. ...
    (sci.stat.edu)
  • Re: Statistics in Psychology?
    ... train a competent medical doctor either. ... statistics -- from the top notch researcher to the low level go-fers. ... mathematics as if you are doing APPLIED statistics. ...
    (sci.stat.edu)
  • Re: Statistics in Psychology?
    ... The same with a applied statistician or a data analyst, ... statistics -- from the top notch researcher to the low level go-fers. ... Its much easier to contribute in THOSE mathematics areas ...
    (sci.stat.edu)
  • Re: Statistics in Psychology?
    ... Is it really necessary to have a PhD in Statistics to apply statistics to ... of the requirement for a top-notch researcher in medicine using ... RF> and PRACTICE -- as Harvard being the extreme example. ... to leave little room for the "Garbage in; ...
    (sci.stat.edu)