Re: [Heartlogic-dev] the plan

[William L. Jarrold]

On Tue, 12 Apr 2005, Joshua N Pritikin wrote:

> On Tue, 2005-04-12 at 00:27 -0500, William L. Jarrold wrote:
> > On Sun, 10 Apr 2005, Joshua N Pritikin wrote:
> > > OHL v2 will have three activities:
> > >
> > > 1. Rating the believability of goal status (Goal, Anti-goal, NoGoal) of
> > > various participants.  This corresponds to level 'g'.
> >
> > Can you expand #1 ever so slightly.  E.g. a one or two paragraph version.
> > Shoot I know we talked about this, but we talked about many things.
> > Is there a hypothesis?  What is it?
>
> Without going too deep, the hypothesis is that a pair of goal statuses
> (same topic, same mindreader, different appraisers) characterize the
> situation in an affectively meaningful way.  We can take this hypothesis
> in lots different directions.

I confused.  I thought the goal statuses were Goal, AntiGoal and NoGoal.

> > But basically, I think I am game for trying out a mixture of different
> > studies.  We'll need to get one or two or maybe three pretty well nailed
> > down.
>
> Are you hinting at a tree or flow-chart style extension of WLJ 2003?  If
> so, sounds good.

I'm not sure.  Reimplimenting WLJ 2003 would not involve a flowchart. 
Doing your study might.

Here's what I'm thinking.  JoeUser logins in.  A computer (or a 
ROBOT!!!!) flips a coin.  If it's heads, JoeUser gets stuck with WLJ 
2003.  JoeUser will be required to complete about 10 items for his data 
to be used.  If it's tails, JoeUser gets to do your study.  Hopefully, 
the amount of effort required by JoeUser would be roughly the same 
regardless of whether he gets to do yours versus gets stuck with mine.

> >   Then we can publicize and get many respondants.
>
> Part of my purpose with this email is to try to narrow down on the
> minimum set of things we need to get ready.

Yep.  Especially important given that the distance between my ears is 
shrinking.

> Can you propose a specific list of things you want ready before we go
> live?  It seems like your list is longer than mine.

EXCELLENT question.  Good way to make sure I am not forgetting something 
important (such as my parachute before jumping out of the airplane).

Roughly this:

Okay, just thining about WLJ 2003 replication here is what needs to be 
done...

Compare currently collected data to WLJ 2003.  If data appears sane 
then continuing marching forward with replication of WLJ 2003.  Maybe talk 
through the project with UT statistical services to make sure everything 
is sane.  Make sure I can run the data and test the hypotheses on a 
moments notice using SAS.  Make sure I know precisely what constitutes 
replication.  Give Joshua study 3 items from 2 or 3 more groups.  Make 
sure website can randomly decide which group each new subject goes into 
yet keep the groups balanced.  Review ethical issues associated 
with doing WWW study and implement/modify informed consent/etc as 
necessary.  Publicisize the website a little to see what kinds of 
response we get for what kind of publicity.   If response is 
suprisingly  big, then we will want to be careful to not blow all our 
e.g. 1000's of subjects unnecessarily.  Instead, have other studies lined 
up.  If response is tepid, then this is less of a worry.  Once completed, 
distill all remaining study 3 items for the of the 12 groups and send off 
to Joshua to be put up on website.  Publicize as necessary to get 
sufficient subjects.

As for your deal...

Get clear on the terms/constructs, research questions and hypotheses. 
Develop the experimental method.  Probably write it up to prove we know 
what we are doing.  Plan the statistical analysis -- e.g. seek input from 
UT Statistical Services as necessary.  Pilot he items on 1 - 20 friends 
and tweak based on their feedback.  Run study.  Analyze results.  Write 
up.  Publish.

Once both of the above are done... Become famous and watch the $$$ roll 
in as we sip margaritas in Goa or somewhere!!!!! (-;

Bill

> > There may be a way of replicating only a portion of my dissertation so as
> > not to require so many subjects.  After I analyze the data collected so
> > far, I can ask statistical services consulting at UT about this.
>
> OK, good.
>
> --
> If you are an American then support http://fairtax.org
> (Permanently replace 50,000+ pages of tax law with about 200 pages.)