[Top][All Lists]

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

Re: Possible to adjust output formatting for stat procedures?

From: John Darrington
Subject: Re: Possible to adjust output formatting for stat procedures?
Date: Sat, 14 Mar 2009 08:38:45 +0900
User-agent: Mutt/1.5.18 (2008-05-17)

I think this problem represents a serious usability issue.  According
to the spss manual, "print formats ... control the form in which
values are displayed by a procedure ...", but pspp doesn't do that.
We've not seriously addressed this issue before, because we've been
holding out for an improved output system.  It looks like that system
isn't going to happen soon. But now that users are complaining, I
think we need to do something as an intermediate measure. 

I suggest that we git rid of the last two arguments to the  tab_float
function, and replace them by a single (const struct fmt_spec *) argument.
This argument may be NULL, in which case the value of
settings_get_format() will be used. For values such as means, stddevs
and sums etc, it's appropriate to pass the print format of the
variable from which they are calculated.  Other values, particularly
dimensionless ones, should be passed NULL. 

Clearly this is not ideal, and may need refinement (counts should
normally displayed with zero decimal places, but not if there are
fractional weights involved).  However I think it'll be a worthwhile
stop-gap measure, until we have some intermediate output
representation which contains a lot more meta-data than we have


On Fri, Mar 13, 2009 at 03:03:07PM -0700, ErikF wrote:
     ErikF wrote:
     > When I perform a procedure in pspp the data come to the output window in
     > F8.2 format. Is there a way to get greater precision? (Tried "FORMATS"
     > command, but this only seems to affect how variables are displayed). Is
     > this unchangeable, or may I apply something like the FORMATS command to
     > results as well as variables? 
     I'm going to answer my own question with an assumption to close the issue,
     just in case anyone else was wondering. While you CAN use FORMATS to
     increase the number of decimals with which variables are displayed in
     output, it does not appear possible at this time to increase the precision
     of calculated results (like Pearson CORR or NPAR probability) displayed in
     the output window over what the given procedure uses (typically 2 points
     after the decimal/F8.2 standard) without tinkering with the sourcecode

PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See 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]