URGENT help needed! : Object amount limit?

[David McFarlane]

dpschpak wrote:
>I had previously made separate experiments for each
>questionnaire, but given that the output would vary in number of rows
>and such, she decided that it would be easier to transfer into SAS if
>each question were it's own object

OK folks, here's the problem -- you are all (dpschpak, your 
professor, and liwenna) grossly misusing the term "object" in this 
context.  I understood this only because of Michiel's reply, he got 
it right.  E-Prime employs an *object*-oriented design.  In this 
context "object" has a very precise technical meaning, and you should 
all learn to use that term correctly just as you would any other 
scientific or technical term.  You cannot just throw that term around 
loosely and expect people to understand what you mean.

Your professor, reasonably enough, most likely means merely that she 
wants each question to appear in some loose sense as its own 
"item".  Now, if you do things in the ordinary EP way as Mich and liw 
advise then you will have one, say, TextDisplay "object" that you use 
repeatedly with differing text on each "trial", and the response for 
each trial will then appear as a separate row in the .edat 
file.  Then once you train yourself on E-Merge and E-DataAid (you 
*will* work faithfully through all the manuals and not overlook the 
wonderful data handling facilities of EP, won't you?) you can slice 
and dice and recombine and analyze the congregate six ways to Sunday, 
and if that is still not enough write some macros in Excel or Python 
or what have you, and from there bring it into SAS.

>, therefore one row of data in the output file.

Ah, but mayhaps your professor wants your output to look more like 
what you get from tools like MediaLab, where you end up with data 
from all the subject in one big table, one column for each "item" 
(>1000 columns in your case, I don't think even Excel can handle 
that!) and one row for each subject.  Given what you can do with 
E-Merge and E-DataAid that is a completely silly way to do things in 
EP, but I can see that if your prof wanted that outcome then it would 
be natural to ask for each "item" to be a separate "object" just so 
that each response went into a column with a unique name.  But even 
for such silliness you would not need >1000 objects.  You could 
accomplish the same thing with a single object and a little bit of 
inline code, e.g.,

g_itemCount = g_itemCount + 1
c.SetAttrib "Item" & g_itemCount, StimText.RESP

(you would of course have to declare g_itemCount as a global variable 
in the User Script area, see chapter 4 of the User's Guide that came 
with EP).  ...  oops, on further thought, even though this tactic 
would create a unique column for each "item", it would would still 
create more than a single row in your .edat file (and leave a bunch 
of "NULL" entries), but you get the idea.

But really, if you are going to use any tool, such as EP, then you 
should adapt to use the tool in its natural way or else get a tool 
already more to your liking.  Otherwise, you are just using a 
screwdriver for a hammer.

>  The response methods differ between and within these
>questionnaires. Some are Yes/No, some ask for the subject to select a
>number or letter corresponding to their answer, some questions ask for
>typed out answers.

OK, so you may need different "objects" for different *kinds* of 
question/response, but if you think through the design properly then 
that number should still remain small.  Once again, please work 
through *all* the tutorials in the manuals before you start any work 
in EP, you are not doing anyone any favors by jumping in with both 
feet before you can swim.

>On a side note, EPrime is a program that is supplied to us through our
>department for free, which may also be why she choose to use it for
>this reason.

Arrgh!  I HATE it when people choose a tool just because it is 
"free"!!!  What are we, Neandertals?!  Choose a tool because it is 
the proper tool for the job, for goodness sakes, and let cost act as 
merely a tie-breaker among otherwise equivalent tools.

>  What other programs would be better suited for this purpose?

AFAIK the preeminent tool for lab-based computerized social and 
psychology questionnaires is the aforementioned MediaLab, from 
Empirisoft (www.empirisoft.com ).  I don't think it's very expensive, 
and they have a very liberal policy toward site licenses to boot, we 
use it here.  Beyond that, I would try any modern database management 
system (e.g., MS Access), which is really meant for this sort of 
thing.  With a little more daring and cleverness you could do this 
all directly in JavaScript and run it from any web browser (and by 
mixing in a little ActiveX and runing on MS Internet Explorer even 
have it save data directly to disk), which would give you some 
practice in a widely used technology and also be, well, free.

-- David McFarlane, Professional Faultfinder

-- 
You received this message because you are subscribed to the Google Groups "E-Prime" group.
To post to this group, send email to e-prime at googlegroups.com.
To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/e-prime?hl=en.