Hey everybody,
does anybody happen to know where you can configure the number of decimals SPSS should use when exporting to a tab separated TXT-file using OMS? I have a syntax file that generates a lot of analyses. The output is saved into a txt-file which is imported by one of the sheets in an excel file, which then formats the output and calculates some extra stuff (i.e. Cohen's d etc). However, because I do so many tests, I have to correct my alpha, requiring me to have much more precise p-values than the ones SPSS currently includes in the OMS-ed TXT-file (which are rounded to three decimals). I assume this setting ('round to three decimals') is somewhere, but I'm not sure it's accessible to 'mere users', so I was hoping perhaps somebody else encountered the same problem? Thanks in advance, kind regards, Gjalt-Jorn Peters Open University Netherlands ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
What version of SPSS are you using, including the update version?
Matthew J Poes Research Data Specialist Center for Prevention Research and Development University of Illinois 510 Devonshire Dr. Champaign, IL 61820 Phone: 217-265-4576 email: [hidden email] -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Gjalt-Jorn Peters Sent: Friday, May 25, 2012 4:10 AM To: [hidden email] Subject: Decimals in OMS Hey everybody, does anybody happen to know where you can configure the number of decimals SPSS should use when exporting to a tab separated TXT-file using OMS? I have a syntax file that generates a lot of analyses. The output is saved into a txt-file which is imported by one of the sheets in an excel file, which then formats the output and calculates some extra stuff (i.e. Cohen's d etc). However, because I do so many tests, I have to correct my alpha, requiring me to have much more precise p-values than the ones SPSS currently includes in the OMS-ed TXT-file (which are rounded to three decimals). I assume this setting ('round to three decimals') is somewhere, but I'm not sure it's accessible to 'mere users', so I was hoping perhaps somebody else encountered the same problem? Thanks in advance, kind regards, Gjalt-Jorn Peters Open University Netherlands ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
In reply to this post by Gjalt-Jorn Peters
You can't control the OMS formatting behavior,
but why not create an Excel file directly from OMS? That will have
full precision (as will the XML format.) Or you can have OMS create
a new Statistics dataset and then set the variable formats explicitly and
save it in csv or other formats.
HTH Jon Peck (no "h") aka Kim Senior Software Engineer, IBM [hidden email] new phone: 720-342-5621 From: Gjalt-Jorn Peters <[hidden email]> To: [hidden email] Date: 05/25/2012 09:22 AM Subject: [SPSSX-L] Decimals in OMS Sent by: "SPSSX(r) Discussion" <[hidden email]> Hey everybody, does anybody happen to know where you can configure the number of decimals SPSS should use when exporting to a tab separated TXT-file using OMS? I have a syntax file that generates a lot of analyses. The output is saved into a txt-file which is imported by one of the sheets in an excel file, which then formats the output and calculates some extra stuff (i.e. Cohen's d etc). However, because I do so many tests, I have to correct my alpha, requiring me to have much more precise p-values than the ones SPSS currently includes in the OMS-ed TXT-file (which are rounded to three decimals). I assume this setting ('round to three decimals') is somewhere, but I'm not sure it's accessible to 'mere users', so I was hoping perhaps somebody else encountered the same problem? Thanks in advance, kind regards, Gjalt-Jorn Peters Open University Netherlands ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Hey Jon,
The reason I asked what I asked was because you may recall an email from me in which I indicated that certain versions of SPSS were only exporting
to 3 decimal places, but that when I ran the update patch, it went back to full precision, and if you knew anything about that. I think you said you weren’t sure, but that it seemed plausible. If he is running that version (I think it was 19 with no patch
and 18 with a certain patch), exporting directly will not solve it (and that may be what he is already doing).
Matthew J Poes Research Data Specialist Center for Prevention Research and Development University of Illinois 510 Devonshire Dr. Champaign, IL 61820 Phone: 217-265-4576 email:
[hidden email] From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of Jon K Peck You can't control the OMS formatting behavior, but why not create an Excel file directly from OMS? That will have full precision (as will the XML format.) Or you can have
OMS create a new Statistics dataset and then set the variable formats explicitly and save it in csv or other formats.
|
Dear all,
that you so much for your reactions! Matthew: I use version 19 (release 19.0.0 - do those last two numbers constitute the update version? The 'about' dialog does not provide any additional version numbers as far as I can see . . .) ViAnn and Jon: I could switch to Excel, but that would decrease the usability of the Excel spreadsheet quite a lot (an argument which I'm sure you can sympathise with :-)). Currently, this Excel spreadsheet has some worksheets that look for 'keywords' in a "PasteHere" worksheet, then extracting the necessary information, calculating new stuff, and presenting it all in a format that's convenient for me to compile when writing papers. This system uses so-called indirect cell references, where it creates the 'coordinates' of the cell it needs to refer to from information from other cells. The "PasteHere" worksheet contains a 'text import' which you can update; i.e. you can point it to a different textfile. If I would use an excel file, I would have to update the indirect cell references. Or, I realise now, I could change them to retrieve the location of the relevant Excel file from a cell somewhere, then I'd only have to update the indirect cell references one, and can then update the contents of that "excel file" cell. This seems a viable solution, thank you! Oh, and the reason I don't work through new datasets if because I try to automate the process as much as possible. I guess with a new dataset I could use some of the macro's at http://spsstools.net to automate changing the format of all variables automatically, and then export. But this would be more work than if it would've been possible to just tell SPSS to stop rounding the output :-) Again, everybody, thank you very much, it's great that I can move on already now! :-) Have a good weekend, kind regards, Gjalt-Jorn --- Gjalt-Jorn Peters Open University Netherlands On 25 May 2012 20:12, Poes, Matthew Joseph <[hidden email]> wrote: > > Hey Jon, > > The reason I asked what I asked was because you may recall an email from me in which I indicated that certain versions of SPSS were only exporting to 3 decimal places, but that when I ran the update patch, it went back to full precision, and if you knew anything about that. I think you said you weren’t sure, but that it seemed plausible. If he is running that version (I think it was 19 with no patch and 18 with a certain patch), exporting directly will not solve it (and that may be what he is already doing). > > > > Matthew J Poes > > Research Data Specialist > > Center for Prevention Research and Development > > University of Illinois > > 510 Devonshire Dr. > > Champaign, IL 61820 > > Phone: 217-265-4576 > > email: [hidden email] > > > > > > From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Jon K Peck > Sent: Friday, May 25, 2012 11:54 AM > To: [hidden email] > Subject: Re: Decimals in OMS > > > > You can't control the OMS formatting behavior, but why not create an Excel file directly from OMS? That will have full precision (as will the XML format.) Or you can have OMS create a new Statistics dataset and then set the variable formats explicitly and save it in csv or other formats. > > HTH > > Jon Peck (no "h") aka Kim > Senior Software Engineer, IBM > [hidden email] > new phone: 720-342-5621 > > > > > From: Gjalt-Jorn Peters <[hidden email]> > To: [hidden email] > Date: 05/25/2012 09:22 AM > Subject: [SPSSX-L] Decimals in OMS > Sent by: "SPSSX(r) Discussion" <[hidden email]> > > ________________________________ > > > > > Hey everybody, > > does anybody happen to know where you can configure the number of > decimals SPSS should use when exporting to a tab separated TXT-file > using OMS? > > I have a syntax file that generates a lot of analyses. The output is > saved into a txt-file which is imported by one of the sheets in an excel > file, which then formats the output and calculates some extra stuff > (i.e. Cohen's d etc). > > However, because I do so many tests, I have to correct my alpha, > requiring me to have much more precise p-values than the ones SPSS > currently includes in the OMS-ed TXT-file (which are rounded to three > decimals). I assume this setting ('round to three decimals') is > somewhere, but I'm not sure it's accessible to 'mere users', so I was > hoping perhaps somebody else encountered the same problem? > > Thanks in advance, kind regards, > > Gjalt-Jorn Peters > Open University Netherlands > > ===================== > To manage your subscription to SPSSX-L, send a message to > [hidden email] (not to SPSSX-L), with no body text except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD > ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Free forum by Nabble | Edit this page |