Re: Sarah Palin- creationist VP candidate?

On Sep 20, 8:41�am, tgdenn...@xxxxxxxxxxxxx wrote:
On Sep 19, 8:30�pm, Tim Norfolk <timsn...@xxxxxxx> wrote:





On Sep 19, 1:09 pm, tgdenn...@xxxxxxxxxxxxx wrote:

On Sep 18, 9:43 pm, Tim Norfolk <timsn...@xxxxxxx> wrote:

On Sep 18, 9:06 am, tgdenn...@xxxxxxxxxxxxx wrote:

On Sep 17, 2:04 pm, Tim Norfolk <timsn...@xxxxxxx> wrote:

On Sep 17, 10:17 am, tgdenn...@xxxxxxxxxxxxx wrote:
<yet another massive snip>

*

As for my counter-experiment, I am not sure how far you intend to go.
Obviously, we can eliminate all arithmetic, since that can be done on
a dime-store calculator. The rules of arithmetic then become those in
algebra, such as the distributive law : a*(b+c) = a*b+a*c, which
immediately implies for the average student that 1/(a+b) = 1/a+1/b and
sqrt(a+b) = sqrt(a)+sqrt(b) - after all, these are just symbols.

So, we give all students access to Maple and spreadsheets, or whatever
you want them to have. We also have "taught" them algorithm-free
arithmetic, algebra, etc.

Let's start with your group 1, the 'average' person, who we would wish
to make inferences. Assume that they have only had the 'algebra'
level.

If your gedankenexperiment holds, 'everyone' should be able to answer
this using a laptop and internet connection:

A test for a disease gives false-positive results p% of the time, and
false-negative results n% of the time. The incidence of the disease in
the general population is i%.

a) Given a positive test, what is the probability that a person has
the disease?
b) Given a negative test, what is the probability that a person has
the disease?
c) What is the sample size required to have a c% probability that at
least one person in the sample has the disease, yet tests negative?

We have, of course, a control group, who have been trained at least
through algebra in the traditional manner, then taught to use the
spread*** software.

Give both groups a set of 200 data points (p,n,i,c) in a spread***,
with a time limit of say 1 hour. Count the number in each group who
actually did the problems correctly in the time allotted, and their
mean time to completion (or just look at the percentage of correct
answers). My prediction is that the score for the control group will
be higher, as they are more likely have phrased the questions in
algebraic form, and then coded them.

This can certainly be illustrative, but I'm just going to restate
slightly to be sure we are talking about the same thing. You are
actually presenting 200 identical problems where there are different
values for pnic, correct?

Ok, the problem is that I really don't understand what you mean by
'phrasing in algebraic form and then coding them'. This seems a fairly
direct matter of plugging numbers into a formula. So if the students
know the formula, they will enter it into the spread*** and fill the
column.

I would say again that you have created a strawman (and I don't even
mean intentionally,) in which 'my' students somehow don't know what a
formula is or that formulas can be combined and modified for
particular applications.

In fact, my students will know the formula or how to look it up
because they have been doing such problems for an entire semester,
using real-world examples, while yours were learning things like
'grouping like terms', and '16 ways to 'solve' a quadratic equation',
and so on. Why do you think that your group will do any better (or
worse) if we give them the formula and instruct them in spread***
use?

-tg- Hide quoted text -

- Show quoted text -

I forgot to add the following to my post:

How about an actual experiment which will not require you to even get
up?

Get me the 'formulae' to solve the problems described. See if you can
do so within 30 minutes, using your internet connection.

Tim, you've started snipping and dodging already. If you want to snip
a long string of previous posts, that's fine, but you should leave the
post to which you are responding intact. (Refers to other reply,)

As to the question in this post:

http://faculty.vassar.edu/lowry/clin2.html

And it may have taken all of 2 minutes to find since I've slept in and
am only on the second cup of coffee.

And you get a grade of D, since you didn't answer all of the questions
- neither does that code.


Not even a formula, Tim, but your worst nightmare---a bit of code that
does what you claim you can't teach your students.


My worst nightmare is better not described here. You seem to forget
that I not only use a lot of software in my work, but have spent
several years in software design. I didn't say I couldn't teach my
students, I said that they don't learn, and the easier you make it,
the less they seem to do so.

You might read Mr. McKendry's comment that using a wrench doesn't make
you a mechanic.

Now if you wish to have a serious discussion, please try to focus on
the following.

To properly design an experiment, you have to be far more specific
than you have been here.

As Mr. McKendry suggested that you do, and you ignored?

If we start arbitrarily with HS freshmen,
the two groups---yours and mine---should be compared after some
elapsed time. After say one year, your students will have sat in an
algebra class doing what I not-that-facetiously described as '16 ways
to solve quadratic equations' (that's a metaphor ok?) �They will *not*
have been introduced to Bayes' Theorem, since they have been training
their little minds to rearrange symbols. My students, on the other
hand, will have spent a fraction of that time learning to use a
computer to manipulate those same symbols, and then have spent the
rest of the time working on more complex problems, thereby reinforcing
their skills in those manipulations.

I still make the claim, based on experience, that they won't reinforce
any skill.


Since we are talking about Group 1, �I may very well have spent time
having them work on Citizen-type statistics, of which yours is an
example. �Perhaps we would have used both the spread*** with Bayes'
Equation and that little free calculator from Vassar, and discussed
when the results are more or less meaningful. Perhaps as an exercise I
would have had them do a simple Google search so that they would see
that such tools are readily available.

Try it with poker players, a fairly broad cross-section of society.
While success strongly depends on simple probability calculations, not
only can the average player not do them, they don't 'believe' the
results even when given them.


If you wish to cite some of those 'educational literature' cases that
you claim exist, we can see what exactly they are comparing. I would
be willing to bet even more than a dollar that it is nothing like what
I am talking about.

-tg- Hide quoted text -

- Show quoted text -

.

Loading