2 Jan
2015
2 Jan
'15
2:13 p.m.
I'm working on transitioning our undergrad introductory statistics course to one that
involves simulation-based ideas alongside parametric, theory-based models. At the risk of
sounding naïve, I would like your opinion on the following: With the advent of computers
that make simulation-based p-value calculations fast and easy, and for those of you that
teach this idea in parallel with theory-based p-values, how do you explain the motivation
for even using theory-based p-values to your students?
In other words, parametric models require assumptions that may not be met. If we can use
simulation to obtain a reliable p-value without worrying about these assumptions, then why
bother with a t-test (e.g.) for a p-value at all? It seems like a historical argument
("It's been done and probably will continue to be done for quite some time")
is one, but... I'm left feeling like I have to tell my students to conduct these
theory-based tests "just because."
Thanks for your thoughts.
Megan
2 Jan
2 Jan
2:22 p.m.
Dear Megan,
My simple answer is the one you gave. That is, the course I teach is
almost completely a service course. I actually teach at an institution
where the only intro stat class is the one that I (as the statistician on
campus) control (I can't teach all sections - there are too many - but I
pick the book, design the syllabus, etc.)
I switched our course to teaching both simulation and traditional methods
together about two years ago, after having many talks with the client
disciplines about the idea that students should come out of such a course
having a better understanding of the ideas of inference, and yet still know
the traditional methods that these departments are still using.
Unfortunately one department has gotten so upset with our teaching these
"new-fangled" ideas that they have seriously discussed dropping the major
requirement for our course and teaching their own intro stat class. Trust
me when I say that this would be a disaster in all kinds of ways for their
students.
So I'm walking a very tight line. I tell my students that I think it is
very important for them to see both kinds of methods. I tell them that I'm
convinced that in about 10 years, all or most client disciplines will be
using simulation based methods. So the students should be exposed to them
now. But, I go on, there are many people still out there using the
traditional methods, and there are many research papers previously
published (and continuing to be published) that use the traditional
methods. For that reason, the students still need to be exposed to the
traditional methods.
Will this work in the long run? I don't know. This is the first time in
over 30 years that this particular department has complained loudly about
the intro stat course in the math/stat department.
Does that help?
I suppose it really depends on who your audience is and how progressive the
client departments are at your institution.
Ann Cannon
Professor of Statistics
Cornell College
Mt. Vernon, Iowa
On Fri, Jan 2, 2015 at 1:13 PM, Olson Hunt, Megan <olsonhum(a)uwgb.edu> wrote:
I’m working on transitioning our undergrad
introductory statistics
course to one that involves simulation-based ideas alongside parametric,
theory-based models. At the risk of sounding naïve, I would like your
opinion on the following: With the advent of computers that make
simulation-based *p*-value calculations fast and easy, and for those of
you that teach this idea in parallel with theory-based *p*-values, how do
you explain the motivation for even using theory-based * p*-values to
your students?
In other words, parametric models require assumptions that may not be met.
If we can use simulation to obtain a reliable *p*-value without worrying
about these assumptions, then why bother with a * t*-test (e.g.) for a *p*-value
at all? It seems like a historical argument (“It’s been done and probably
will continue to be done for quite some time”) is one, but… I’m left
feeling like I have to tell my students to conduct these theory-based tests
“just because.”
Thanks for your thoughts.
Megan
_______________________________________________
SBI mailing list
SBI(a)causeweb.org
https://www.causeweb.org/mailman/listinfo/sbi
2:26 p.m.
Thats a great question Megan and I know I often give the before we had
computers argument and its not that satisfying. And I agree that we
shouldnt just see them as an alternative when theory-based methods dont
work, but as a reasonable approach on their own with large and small sample
sizes. But you can say they are still what is predominantly done in practice
and you can still talk about test statistics as giving additional
information beyond the p-value. Because I still tend to do fairly
traditional confidence intervals, then the theory-based p-value and test
statistic give me a nice transition there (though I often talk more in terms
of 2SEs in both cases and dont worry too much about the exact
multiplier/critical value). I think I would be pretty happy teaching a
course (like the CATALST folks do) that does very little with the
theory-based methods, but I dont think my client departments would be very
happy with me yet. Mostly I want students to understand what the p-value
and confidence interval mean and then turn things over to a computer (and to
be honest maybe not always know the exact details of what the computer is
doing) to get the output and focus on the conclusions we can draw
It is nice to show them two different reasonable approaches to help contrast
them. In fact, maybe we should be doing more with exact methods in place of
both of these?
Beth
From: sbi-bounces(a)causeweb.org [mailto:sbi-bounces@causeweb.org] On Behalf
Of Olson Hunt, Megan
Sent: Friday, January 02, 2015 11:13 AM
To: 'sbi(a)causeweb.org'
Subject: [SBI] Explaining motivation for theory-based models
Im working on transitioning our undergrad introductory statistics course to
one that involves simulation-based ideas alongside parametric, theory-based
models. At the risk of sounding naïve, I would like your opinion on the
following: With the advent of computers that make simulation-based p-value
calculations fast and easy, and for those of you that teach this idea in
parallel with theory-based p-values, how do you explain the motivation for
even using theory-based p-values to your students?
In other words, parametric models require assumptions that may not be met.
If we can use simulation to obtain a reliable p-value without worrying about
these assumptions, then why bother with a t-test (e.g.) for a p-value at
all? It seems like a historical argument (Its been done and probably will
continue to be done for quite some time) is one, but
Im left feeling like
I have to tell my students to conduct these theory-based tests just
because.
Thanks for your thoughts.
Megan
3414
days inactive
3414
days old
2 comments
3 participants
participants (3)
-
Ann Cannon -
Beth Chance -
Olson Hunt, Megan