[Heartlogic-dev] Re: surveys

[William L. Jarrold]

Dear OHL-ians,

Okay I've made a tiny step forward in finding correspondance btwn www
data and diss study 3 data.  But I'm thinking more and more beyond my
current obsessive bean counter stage, taking too much the worms eye
perspective.  I need to consult Kid Beyond or something and get on to
real work.  Stay tuned.

Meantime, enjoy the 2.398072829 beans counted below.

(NOTE: You will the fully anal evidence for the CONCLUSION farther
below.)

CONCLUSION: So, it looks like I am learning how to find
correspondances between items on the website and the dissertation
data.  And it looks like there is a pretty good correspondance between
study 3 data and www data.  All 10 respondants but 1 gave a rating of
1.  The other gave a rating of 2.  But this is not surprising.  The
item was a mindnumbly simple one.  Obviously unbelievable.  But at
least we are proving to ourselves that the data is sane and that we
can find correspondaneces.

I should also note that I discoverd that my backup of homebase does
not have the hugely important directories ..Dissertation/FinalOral nor
Dissertation/AfterFinalOral.  Presumably this is because my backup of
homebase in the fall of 2004 happend before these directories were
migrated to homebase???  Well, I have a new reason to do data
recovering on Homebase.  Presumably I moved it there successfully?  I
pray to God it is recoverable and on there.  By luck my laptop happens
to have AfterFinalOral.  Live and learn.

TO DO NEXT: Look over the web data and see if it makes sense from a
plain old believability point of view.  Then later go through the
pains of exactly comparing it with diss data.

On Tue, 26 Apr 2005, William L. Jarrold wrote:

> On Tue, 26 Apr 2005, Joshua N Pritikin wrote:
>
> > On Tue, 2005-04-26 at 00:34 -0500, William L. Jarrold wrote:
> > > Thanks.  Just letting you know I did rouglyl one hour more work
> > > tonight on this not conclusivei yet, more tomorrow i hope....
> > >
> > > So we are a tiny step closer to some sort of comparision...
> > >
> > > I took the file (which used to be yuour attachment) called
> > > stats-on-web-data.text in ...
> > >
> > > ~/OpenHeartLogic/AnalyzeDataObtainedUpTo2005-04-12/
> > >
> > > ...and found that what corresponds to scenario a1 used in group 6
> > > i.e. a computer generated, type i, strongly reversed item, corresponds
> > > to xref 6 in stats-on-web-data.text...See....
> > >
> > > |  6 |
> > > Goal: Tracy wants an apple.
> > > Situation: Mommy gives Tracy an apple for lunch.
> > > Question: How will Tracy feel when Mommy gives her an apple for lunch?
> > > | She got whats she wanted.
> > > | Sad |
> > >
> > > |    6 |        5 |        -2.0000 |            0.0000 |
> > >
> > > ...so we have only 5 people who filled this one out.  We really need
> > > 10 judgements to be sue of reliability. And everyone gave it a -2
> > > (highly unbelievable.)  Finish later too tired.
> >
> > I have some script which can do the matching semi-automatically.
> > However, I need your assurance of the accuracy of the groups you sent
> > me.  Verifying that the grp*.tex files match the hardcopy seems like a
> > top priority.
>
> It is pretty high but I have some other stuff that is higher.  Stay tuned,
> I'm percolating.
>
> Bill


Okay here is my diary entry for today except I snipped off started at
the CONCLUSION part and slapped that at the top of this email, just
after the first few lines.

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2005-04-27::00:09

Okay, Now I have a faint glimmer that the first item in any file, such
as grp6.tex or grp9.tex or whatever is actually the LAST item.
I.e. the order was reversed.  I should check this.  BUT, if this is
true, then when I yesterday said "Darn this does not match
prediction." I was wrong.

You see, if my faint glimmer is correct, then the last item in the
file is actually Item1.  So, using our piece of text from groups.text,
i.e....

item1   grp06   C       I       R2
item2   grp06   H       II      R1
item3   grp06   C       II      R2
item4   grp06   H       III     U
item5   grp06   C       III     U  (so it's 3U / 2R1 / 4R2 & 1U / 2R1 / 1R2)
item6   grp06   H       II      R1
item7   grp06   C       I       U
item8   grp06   H       III     R2
item9   grp06   C       II      R2

...we see that item1 is C- Computer Generated, I - type I and R2,
strongly reversed.  Bingo.  Prediction matches.

So, next I went to

bash-2.05b$ grep a1 *

A-XXX-XX-XXXX:a1 2 she will be happy because she got what she wanted.
B-XXX-XX-XXXX:a1 1 because tracy wanted an apply and that would make her happy.
C-XXX-XX-XXXX:a1 1 tracy got exactly what she wanted therefore she would not be 
sad.
D-XXX-XX-XXXX:a1 1 because she didn't get what she wanted.
E-XXX-XX-XXXX:a1 1 she got what she wanted so she should be happy.
F-XXX-XX-XXXX:a1 1 tracy will be happy she got what she wanted.
G-XXX-XX-XXXX:a1 1 she would be happy.
H-XXX-XX-XXXX:a1 1 she will be happy b/c she got what she wanted.
I-XXX-XX-XXXX:a1 1 the answer and reason don't match - she'd be happy.
J-XXX-XX-XXXX:a1 NOTHINGCIRCLED she can't feel sad because she got what she 
wanted.
K-XXX-XX-XXXX:a1 1 tracy got what she watned, so she should be happy.

(NOTE: where you see XXX-XX-XXXX in this diary file used to be SSN's.
I removed them for privacy reasons. (-:)

bash-2.05b$ pwd
/Users/billj/Desktop/Homebase-BillJ-2004/billj/ArchivedFiles/GradSchool/Dissertation/DataAnalysis/\
After2002/Study3/Group06
bash-2.05b$