[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: PSPP-BUG: Suggestions to make PSPP much easier to use

From: John Darrington
Subject: Re: PSPP-BUG: Suggestions to make PSPP much easier to use
Date: Mon, 18 Jun 2012 21:42:15 +0000
User-agent: Mutt/1.5.18 (2008-05-17)

Thank you for your feedback on PSPP.  I am glad that you have found the
program useful.

PSPP started as a program intended to be used by those already familiar
with SPSS.  But I think you are right when you suggest that it might 
well be taken further than that.  Your idea of combining it with a form
of expert system is an interesting one, albeit not trivial to do - I think
it would require a lot of input from experts in that field.

Reading your text, I think it is appropriate to remind you that p values
are NOT error probabilities.  Rather it is the probability of obtaining a 
particular result, when the null hypothesis is true.  In other words, if a 
t-test comparing X with Y, gives a p value of 0.05, this does NOT mean 
you can be 95% certain that X < Y.  Rather it means there is a 5% chance
that the t-test would give that result when X>=Y.

So I tend to agree with your statement that PSPP is a domain-specific
calculator, but I don think that is how it should be.  A really valuable
program would provide the numbers AND enlighten the user as to their
meaning, and their pre-conditions.  Unfortunately too many programs have
concentrated on ??making it easy for the user to get the rsults?? that
they also make it easy for him to misunderstand the results. Avoiding that
pitfall is hard!

Thanks once again for your feedback.

On Sun, Jun 17, 2012 at 11:06:18AM +0000, Rick van Rein wrote:
     Thanks for making PSPP.  We used it successfully in an MSc assignment,
     and never had to resort to its closed relative.
     Looking back, I am stunned how much statistical and mathematical
     knowledge we had to put into the whole endeavour, and it got me thinking
     that this is in fact a shortcoming in the mindset of SPSS/PSPP: IMHO,
     these tools should encapsulate as much as possible the knowledge of the
     statistical domain of expertise.
     With this in mind, I've written a small essay that describes an extended
     logic that could be used for these programs.  I know that PSPP is a
     sort-of clone of SPSS, but it might be useful if it set up its own
     trajectory and let SPSS clone what it has done ;-) so I am sending this
     to you guys first of all.  I hope this is useful to you!
     Once more, I was happy with the calculations it could automate for us,
     and this essay should merely be seen as an attempt by me to give
     omething back with a bit of "conceptual analysis" of the tool.  If I
     used it a lot, I would start coding it too, but that is not the case.
     Rick van Rein
     Bug-gnu-pspp mailing list

PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://keys.gnupg.net or any PGP keyserver for public key.

Attachment: signature.asc
Description: Digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]