Re: Using pspp

[John Darrington]

On Tue, Oct 27, 2009 at 10:18:18AM -0500, Leonard Evens wrote:
     I have been using pspp from the gui version psppire.   I am running
     version 0.6.1 under Fedora 11.  (The package is pspp-0.6.1-3.fc11.i586.)
     But I am  unclear about some things.
     
     1> Is it possible to delete variables?   How about entries?

Yes and yes.  Right click on the row/column concerned and choose "Clear" 
from the popup menu.  Alternatively select the row/column and from the Edit 
menu choose the Clear item.
     
     2> I've been running descriptive statistics for frequencies and
     crosstabs.  But I have to copy what is in the output window each time
     and paste it into a text file using gedit.  Is there some way using
     psppire to put the output in a file?  

The output is by default put into a file called "psppire.txt".


     Is there some way to control the
     format of the output?  It often divides a cross tabulation table into
     separate sub-tables, as might be required to continue on the next page,
     and I then have to merge by deleting extra lines in gedit. 

Not from the GUI.  It's possible from the command line interface.  See the 
section of the manual entitled "Output devices".
     
     Is there some way to show the details of how a recoded variable was
     recoded?

I'm not sure that I understand this question.  There is a record of all commands
run in pspp.jnl if that helps.
     
     3>  Supposedly pspp can produce graphs of various kinds, e.g.,
     histograms.  Is this only possible by running pspp in a terminal window
     at the command line?

Well you can also do it by running the commands in the GUI's syntax window.
However the problem is, that version 0.6.x doesn't have any way to actually
display the graphs.   This is currently being addressed - if you're prepared
to build the "output" branch from the git repository, then you can see what 
progress has been made so far.
     
     4>  I have split the data on the values of a certain variable and then
     run crosstabs on different pairs of variables.  Some of these variables
     are recodings into new variables of other variables. (The split is done
     on the basis of one of these recoded variables.) There appears to be
     some sort of bug.  Namely, in some cases, perhaps only in cases of
     recoded variables, the reported number of missing values is wrong in one
     of the two sets produced by the split. In the errant set, the number
     reported is the total number of missing values rather than just the
     number for that set.  As far as I can tell the cross tabulations are
     right for the valid data.
     
     I've reported this to the pgpp-bugs mailing list, but haven't gotten a
     detailed response yet.  I presume it is possible I did something wrong.
     e.g., in creating the recoded variables.  Can anyone suggest anything
     which I might have done wrong?

Probably nothing.  The crosstabs command was completely rewritten in version 
0.7.x so hopefully this problem won't arise there.
     
     5>  Is there some tutorial which goes through various examples,
     particularly  using psppire?  I've been studying the manual, but it
     isn't a whole lot of help in using the pspire.

Unfortunately not.  Pspp started as a free replacement for Spss.  However
it's starting to attract interest from people who have never used Spss, so
it maybe time to write such a tutorial.  Any volunteers?

J'

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

signature.asc
Description: Digital signature