Problems with 24
Locked
 
|
I’m having problems with SPSS24 not performing transformations with EXECUTE or after FREQUENCIES. The file as originally supplied has a range of positive values to be treated as missing: they have value labels, but none of the values have been declared as missing. The basic process is to recode these positive values to negative and to set missing values as (lo thru -1) where 0 is a valid value or as (lo thru 0) where 0 denotes “Skipped question” etc. I have a perfectly good set of syntax for: VARIABLE LEVEL RECODE MISSING VALUES ADD VALUE LABELS . . and am running jobs with well under 100 variables at a time. The commands are repeated in the spv without errors or warnings. After running the job with EXECUTE or running FREQUENCIES to check, the new measurement levels appear in the data editor, but the negative value labels do not. The recoded values appear as missing in the frequency table, but without their value labels. The file I’m working on very large 10,765 variables, 95,630 cases, but to date all my syntax files have worked. I know there is a bug in RECODE which has been reported, but the syntax is correct. Is anyone else having the same or similar problems? John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop |
|
So the specific issue is value labels assigned
to negative values via the ADD VALUE LABELS command?
Rick Oliver Senior Information Developer IBM Business Analytics (SPSS) E-mail: [hidden email] From: John F Hall <[hidden email]> To: [hidden email] Date: 05/17/2016 01:32 PM Subject: Problems with 24 Sent by: "SPSSX(r) Discussion" <[hidden email]> I’m having problems with SPSS24 not performing transformations with EXECUTE or after FREQUENCIES. The file as originally supplied has a range of positive values to be treated as missing: they have value labels, but none of the values have been declared as missing. The basic process is to recode these positive values to negative and to set missing values as (lo thru -1) where 0 is a valid value or as (lo thru 0) where 0 denotes “Skipped question” etc. I have a perfectly good set of syntax for:
VARIABLE LEVEL RECODE MISSING VALUES ADD VALUE LABELS
. . and am running jobs with well under 100 variables at a time. The commands are repeated in the spv without errors or warnings. After running the job with EXECUTE or running FREQUENCIES to check, the new measurement levels appear in the data editor, but the negative value labels do not. The recoded values appear as missing in the frequency table, but without their value labels.
The file I’m working on very large 10,765 variables, 95,630 cases, but to date all my syntax files have worked. I know there is a bug in RECODE which has been reported, but the syntax is correct. Is anyone else having the same or similar problems?
John F Hall (Mr) [Retired academic survey researcher]
Email: johnfhall@... Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop
===================== To manage your subscription to SPSSX-L, send a message to LISTSERV@...(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
|
|
|
Rick Here is an example of what I start with: Version 34 (rows 3307 to 3318)
Run a set of syntax, one command at a time (no errors or warnings) variable level ~ ~ ~ recode ~ ~ ~ missing values ~ ~ ~ add value labels ~ ~ ~ exec. All the variables above are included in all sections of the (shortened) syntax below, which runs without error or warning. missing values ssocclas smanual shortint prefjob empliv (7=-7)(8=-8) (9=-9). add value labels /ssocclas -7 "SK:LOOK AFT HOME" -8 "ARMED FORCES" -9 "INSUFFIC INFO" /smanual -7 "SK:LOOK AFT HOME" -8 "ARMED FORCES" -9 "UNABLE TO CLASSIFY" /shortint prefjob -7 "OTHER" -8 "DK" -9 "NA" /empliv -7 "SK/MORE T LIVNG" -8 "DK" -9 "NA". Version 35 (rows 3307 to 3318)
As you can see, some variables display {0 SKIPPED} in the values column even after values 8, 9, 98, 99 etc are recoded to -8, -9, -98, -99 etc and missing values are declared in my syntax as (lo thru 0). Ignore None as missing for rschtype to soccode. prefjob to srgcomp have the correct range in the Missing column, but not in the Values column even though add value labels was correctly specified. Some of the missing values contain six or seven digits: one has 5 digits plus two decimal places! It’s not my data set and whoever wrote the original setup files had clearly not been to any of my classes. John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Rick Oliver So the specific issue is value labels assigned to negative values via the ADD VALUE LABELS command?
The file as originally supplied has a range of positive values to be treated as missing: they have value labels, but none of the values have been declared as missing. The basic process is to recode these positive values to negative and to set missing values as (lo thru -1) where 0 is a valid value or as (lo thru 0) where 0 denotes “Skipped question” etc. I have a perfectly good set of syntax for: VARIABLE LEVEL RECODE MISSING VALUES ADD VALUE LABELS . . and am running jobs with well under 100 variables at a time. The commands are repeated in the spv without errors or warnings. After running the job with EXECUTE or running FREQUENCIES to check, the new measurement levels appear in the data editor, but the negative value labels do not. The recoded values appear as missing in the frequency table, but without their value labels. The file I’m working on very large 10,765 variables, 95,630 cases, but to date all my syntax files have worked. I know there is a bug in RECODE which has been reported, but the syntax is correct. Is anyone else having the same or similar problems? John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop ===================== 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 John F Hall
Hello John,
I get a daily digest of the messages sent to the SPSS list. This morning, I read your recent email to the SPSS list regarding problems with SPSS 24, and I became greatly alarmed when I saw the following statement from you as a veteran user of SPSS: I know there is a bug in RECODE which has been reported . Although I routinely get IBM notifications about these kinds of problems, I do not recall seeing anything about a bug in RECODE in SPSS 24. However, because I am currently involved in extensive recoding operations using SPSS 24, I would certainly welcome any insights which you and/or other members of the list might have regarding this reported bug in RECODE in SPSS 24. Thank you. David N. ===================== 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 John F Hall
Jon,
Thank you for your helpful and reassuring reply. David N. From: Jon Peck [mailto:[hidden email]] Sent: Wednesday, May 18, 2016 9:37 AM To: David Nolle <[hidden email]> Cc: John F Hall <[hidden email]> Subject: Re: [SPSSX-L] Problems with 24 David, The RECODE bug has to do with very long multiple variable lists. It will not give you wrong results, though. It is a parsing error, so the RECODE command issues an incorrect error message and stops instead of doing the recodes. If you don't get such messages, everything is fine. This problem also exists in V23 but had not been discovered (by John) until recently. On Wed, May 18, 2016 at 7:33 AM, David Nolle <[hidden email]> wrote: Hello John, I get a daily digest of the messages sent to the SPSS list. This morning, I read your recent email to the SPSS list regarding problems with SPSS 24, and I became greatly alarmed when I saw the following statement from you as a veteran user of SPSS: “I know there is a bug in RECODE which has been reported….” Although I routinely get IBM notifications about these kinds of problems, I do not recall seeing anything about a bug in RECODE in SPSS 24. However, because I am currently involved in extensive recoding operations using SPSS 24, I would certainly welcome any insights which you and/or other members of the list might have regarding this reported “bug in RECODE” in SPSS 24. Thank you. David N. ===================== 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 -- Jon K Peck [hidden email] ===================== 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 John F Hall
John,
Thank you for adding further documentation on your previously reported bug in RECODE. In addition, thank you for providing details on your current problem with value labels. David N. From: John F Hall [mailto:[hidden email]] Sent: Wednesday, May 18, 2016 1:35 PM To: 'David Nolle' <[hidden email]> Cc: [hidden email] Subject: RE: Problems with 24 David The bug was reported by Jon Peck two or three weeks ago after I managed to cause several catastrophic errors (something to do with the back engine). When this happens, the working *.sav file is lost: luckily SPSS saves the *.sps and *.spv files when the processor is restarted, but it gave me quite a fright when it first happened. Because I was taking screenshots with MS-Snip I sent Jon my error reports off-line. He is now retired and doesn't see the schedules anymore, but expects that this will get fixed in the next fixpack. I managed to recreate the problem twice and it seems to be caused by a missing forward slash separating sets of variables within recode commands (easy to do with long lists as, if they are at the beginning of a line, they are not that easy to see by eyeball check). The thing is that SPSS did not pick up this error in the syntax editor and highlight it in red, but it did report the error in the spv. The current problem is that changes are not made to the value labels even when they are correctly specified. I've just run a test job on a few variables which still displayed {0 SKIPPED} in the Labels column even though they had recoded negative values and (lo thru 0) was displayed in the Missing column recode prefjob (8=-8) (9=-9) / rindsect sseggrp srgclass srgcomp (9=-9) / empliv (7=-7)(8=-8) (9=-9). missing values rindsect prefjob sseggrp srgclass srgcomp empliv (lo thru 0). add value labels prefjob empliv -8 "DK" -9 "NA" /rindsect sseggrp srgclass srgcomp -9 "CANT CLASSIFY" /srgclass srgcomp -9 "FORCES/ NO INFO". . . and it worked: Name Measure Values Missing rindsect Nominal {-9, CANT CLASSIFY}... LO - 0 prefjob Nominal {-9, NA}... LO - 0 sseggrp Nominal {-9, CANT CLASSIFY}... LO - 0 srgclass Ordinal {-9, FORCES/ NO INFO}... LO - 0 srgcomp Ordinal {-9, FORCES/ NO INFO}... LO - 0 It may be something to do with the number of variables processed, the number of unique value combinations to recode or the number of different labels to add. With almost 11,000 variables it's bad enough doing 25 - 30 at a time: if I have to cut it down to 5 or 10 the job will never get finished. Luckily there some quite long contiguous blocks which can be picked with <varx> to <vary>. John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop ===================== 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 John F Hall
Rick Associated thread [SPSSX-L] FW: Rejected posting to [hidden email]] appended below. As part of his investigations, Jon asked me to send him a file with a single case with all 10,756 variables. Whilst preparing this for him I caused a couple more crashes. Fortunately SPSS prompts to saves the sps and spv files before restarting the processor. He is travelling for a few days and will be away from his computer, so to bring you up to date . . It looks as if the problem occurs in long lists, but it doesn’t help that simple errors are not always flagged in the syntax editor. For basic commands, I always use syntax in preference to the GUI: for me, if SPSS syntax displayed periods and forward slashes as bold (and in a different colour) these errors would be less likely. They are very hard to spot in dense, small text and I can’t use CTRL+scroll to increase the size. I finally got what I wanted after splitting the recode command into four separate commands, running each one separately followed by a freq command for a single variable in the preceding recode list (to force a data pass). When all four recodes had been processed, I ran variable level, missing values and add value labels with all recoded variables in the same list. Tedious, but effective. Only another 7,000 variables to go! Regards John From: John F Hall [mailto:[hidden email]] Jon It just happened again. This time it was during a shortened recode command for single digit values recode dutyresp rindsect sseggrp srgclass srgcomp (9=-9) / elseduty (7=-7)(9=-9) /biginf1 biginf2 dutyresp to oldresp semploye sempwork spartful phoneacc altadd advance francliv to indiainf girladv1 to boyadv8 presssat to phonesat damage techeff1 techeff2 govtdopr govtdoun govtdosl govtdoss fruitpr to ratespr jobmf1 to jobmf11 mwextras to femhome empfeel empearn (8=-8) (9=-9) /rsector selse ssector /ssocclas smanual shortint prefjob empliv (7=-7)(8=-8) (9=-9). SPSS closed down and restarted the processor, but when I re-opened cum24.sav I got this message: GET FILE='C:\Users\John\Desktop\peckvar\cum34.sav'. Warning # 67. Command name: GET FILE The document is already in use by another user or process. If you make changes to the document they may overwrite changes made by others or your changes may be overwritten by others. File opened C:\Users\John\Desktop\peckvar\cum34.sav DATASET NAME DataSet7 WINDOW=FRONT. I’ll try it again, but if it still causes a catastrophic error, I’ll chop it into smaller slices until it works [Catastrophic error message here] Split the job into 50/50 recode dutyresp rindsect sseggrp srgclass srgcomp (9=-9) / elseduty (7=-7)(9=-9) /biginf1 biginf2 dutyresp to oldresp semploye sempwork spartful phoneacc altadd advance francliv to indiainf girladv1 to boyadv8 (7=-7)(8=-8) (9=-9). freq boyadv8.
recode dutyresp rindsect sseggrp srgclass srgcomp (9=-9) / elseduty (7=-7)(9=-9) /biginf1 biginf2 dutyresp to oldresp semploye sempwork spartful phoneacc altadd advance francliv to indiainf girladv1 to boyadv8 presssat to phonesat damage techeff1 techeff2 govtdopr govtdoun govtdosl govtdoss fruitpr to ratespr jobmf1 to jobmf11 mwextras to femhome empfeel empearn (8=-8) (9=-9) /rsector selse ssector /ssocclas smanual shortint prefjob empliv (7=-7)(8=-8) (9=-9). [Catastrophic error message here] Split job again Didn’t find this until later, but it may have caused the error: /rsector selse ssector /ssocclas smanual shortint prefjob empliv plus there was a period at the end of another line before the end of add value labels I ran the job in smaller sections with a frequency count after each section and everything has worked perfectly. All recoded values and labels now correct, but there some pre-existing negative values which don’t show up in the ellipse and have no labels. The ellipse info doesn’t match the frequency table [Ellipse screenshot here] (Values -3, -2 and -1 not shown)
This could be a bit of a headache later, as it will mean diving into the documentation to find out what such codes indicate. At some point I shall need to delete the previous positive values and labels: any chance of a new command? delete value labels Time to stop: I’ve been at this all day. John |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Administrator
|
In reply to this post by John F Hall
"At some point I shall need to delete the previous positive values and labels: any chance of a new command?
delete value labels" Hogwash! Simply use VALUE LABELS command and reassign all relevant labels as you see fit. I think you make your life very complicated by NOT bothering to acquire a little bit of programming skills. Carry on!
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
Using Value Labels requires recreating the whole set, which should not be necessary to make incremental changes. Good subject for an extension command.
On Friday, May 20, 2016, David Marso <[hidden email]> wrote: "At some point I shall need to delete the previous positive values and -- ===================== 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 |
|
|
Incremental appending already works with ADD VALUE LABELS ... A logical supplement would be REMOVE VALUE LABELS ... Mario Jon Peck <[hidden email]> schrieb am 16:02 Freitag, 20.Mai 2016: Using Value Labels requires recreating the whole set, which should not be necessary to make incremental changes. Good subject for an extension command. On Friday, May 20, 2016, David Marso <[hidden email]> wrote: "At some point I shall need to delete the previous positive values and -- ===================== 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 David Marso
For 10,765 variables with dozens of different positive value combinations to
be treated as missing, and hundreds of different labels for the valid values, you're welcome to try. I've told you before, I'm a (sort of sociologist) survey researcher, not an IT guru. It's only my 50-year old Algol experience (and 42 years of using SPSS on hundreds of surveys) that enables me to find workarounds that work for me and are quicker. I simply don't have that much time to spare, and neither do the teachers, students and fellow researchers for whom I'm doing this. -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of David Marso Sent: 20 May 2016 15:27 To: [hidden email] Subject: Re: Problems with 24 "At some point I shall need to delete the previous positive values and labels: any chance of a new command? delete value labels" Hogwash! Simply use VALUE LABELS command and reassign all relevant labels as you see fit. I think you make your life very complicated by NOT bothering to acquire a little bit of programming skills. Carry on! ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Problems-with-24-tp5732161p573 2187.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== 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 spss.giesel@yahoo.de
Exactly what I suggested. Marso, go sit on the naughty step! From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Mario Giesel Incremental appending already works with ADD VALUE LABELS ... A logical supplement would be REMOVE VALUE LABELS ...
Jon Peck <[hidden email]> schrieb am 16:02 Freitag, 20.Mai 2016: Using Value Labels requires recreating the whole set, which should not be necessary to make incremental changes. Good subject for an extension command.
Jon K Peck
===================== 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 |
|
Administrator
|
"Marso, go sit on the naughty step!"
I will not! What would be the point of REMOVE VALUE LABELS when you can just use ADD VALUE LABELS with a blank for the label you want to clobber? Corner time and pointy hat for you John ;-) DATA LIST FREE / x. BEGIN DATA 1 2 3 9 -1 END DATA. MISSING VALUES x (-1,9). ADD VALUE LABELS x 1 'one' 2 'two' 3 'three' 9 'nine MISS' -1 'neg1 miss'. FREQUENCIES x. ADD VALUE LABELS x 3 ''. FREQUENCIES x.
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
It may or may not be worth noting that
you can remove all value labels for a variable with the Value Labels command:
value labels varname. It also may or may not be worth noting that commands that only modify the dictionary (like Value Labels and Add Value Labels) take effect immediately and do not require a data pass. data list free /varname. begin data 1 2 3 end data. value labels varname 1 'one' 2 'two' 3 'three'. display dictionary. value labels varname. display dictionary. Rick Oliver Senior Information Developer IBM Business Analytics (SPSS) E-mail: [hidden email] From: David Marso <[hidden email]> To: [hidden email] Date: 05/20/2016 10:29 AM Subject: Re: Problems with 24 Sent by: "SPSSX(r) Discussion" <[hidden email]> "Marso, go sit on the naughty step!" I will not! What would be the point of REMOVE VALUE LABELS when you can just use ADD VALUE LABELS with a blank for the label you want to clobber? Corner time and pointy hat for you John ;-) DATA LIST FREE / x. BEGIN DATA 1 2 3 9 -1 END DATA. MISSING VALUES x (-1,9). ADD VALUE LABELS x 1 'one' 2 'two' 3 'three' 9 'nine MISS' -1 'neg1 miss'. FREQUENCIES x. ADD VALUE LABELS x 3 ''. FREQUENCIES x. John F Hall wrote > Exactly what I suggested. Marso, go sit on the naughty step! > > From: SPSSX(r) Discussion [mailto: > SPSSX-L@.UGA > ] On Behalf Of Mario Giesel > Sent: 20 May 2016 16:46 > To: > SPSSX-L@.UGA > Subject: Re: Problems with 24 > > Incremental appending already works with > > ADD VALUE LABELS ... > > A logical supplement would be > > REMOVE VALUE LABELS ... > > Mario > Jon Peck < > jkpeck@ > <mailto: > jkpeck@ > > > schrieb am 16:02 Freitag, 20.Mai 2016: > > > Using Value Labels requires recreating the whole set, which should not be > necessary to make incremental changes. Good subject for an extension > command. > > On Friday, May 20, 2016, David Marso < > david.marso@ > <mailto: > david.marso@ > > > wrote: > "At some point I shall need to delete the previous positive values and > labels: any chance of a new command? > delete value labels" > > Hogwash! Simply use VALUE LABELS command and reassign all relevant labels > as you see fit. > I think you make your life very complicated by NOT bothering to acquire a > little bit of programming skills. > Carry on! > > > > > ----- > Please reply to the list and not to my personal email. > Those desiring my consulting or training services please feel free to > email me. > --- > "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos > ne forte conculcent eas pedibus suis." > Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in > abyssum?" > -- > View this message in context: > http://spssx-discussion.1045642.n5.nabble.com/Problems-with-24-tp5732161p5732187.html > Sent from the SPSSX Discussion mailing list archive at Nabble.com. > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA > (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 > > > -- > Jon K Peck > jkpeck@ > <mailto: > jkpeck@ > > > > ===================== To manage your subscription to SPSSX-L, send a > message to > LISTSERV@.UGA > <mailto: > LISTSERV@.UGA > > (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 > LISTSERV@.UGA > <mailto: > LISTSERV@.UGA > > (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 > LISTSERV@.UGA > (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 ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Problems-with-24-tp5732161p5732195.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== 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 David Marso
Think I'll wait for Jon to get back from his travels. He offered to modify his Python code to remove the superfluous positive (missing) values. It may not even be necessary: once all the positive values to be treated as missing have been recoded to negative, and Missing set to (lo thru -1) or (lo thru 0) they won't appear in tabulations. They still appear in the ellipse in the Values column, but this is an aesthetic rather than a technical problem.
Since most survey data are categorical, why do we need cumulative %% Isn’t it about time that the frequencies command was modified to have options. eg: frequencies ASApplic /sta cou val . For quick checks I always use Data >> Define Variable Properties: it gives all the info needed for checking, including counts. I never use it to define properties: it’s quicker to use syntax. That’s enough for today. John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop -----Original Message----- "Marso, go sit on the naughty step!" I will not! What would be the point of REMOVE VALUE LABELS when you can just use ADD VALUE LABELS with a blank for the label you want to clobber? Corner time and pointy hat for you John ;-) DATA LIST FREE / x. BEGIN DATA 1 2 3 9 -1 END DATA. MISSING VALUES x (-1,9). ADD VALUE LABELS x 1 'one' 2 'two' 3 'three' 9 'nine MISS' -1 'neg1 miss'. FREQUENCIES x. ADD VALUE LABELS x 3 ''. FREQUENCIES x. John F Hall wrote > Exactly what I suggested. Marso, go sit on the naughty step! > > From: SPSSX(r) Discussion [mailto: > ] On Behalf Of Mario Giesel > Sent: 20 May 2016 16:46 > To: > Subject: Re: Problems with 24 > > Incremental appending already works with > > ADD VALUE LABELS ... > > A logical supplement would be > > REMOVE VALUE LABELS ... > > Mario > Jon Peck < > jkpeck@ > <mailto: > jkpeck@ > > > schrieb am 16:02 Freitag, 20.Mai 2016: > > > Using Value Labels requires recreating the whole set, which should not > be necessary to make incremental changes. Good subject for an > extension command. > > On Friday, May 20, 2016, David Marso < > david.marso@ > <mailto: > david.marso@ > > > wrote: > "At some point I shall need to delete the previous positive values and > labels: any chance of a new command? > delete value labels" > > Hogwash! Simply use VALUE LABELS command and reassign all relevant > labels as you see fit. > I think you make your life very complicated by NOT bothering to > acquire a little bit of programming skills. > Carry on! > > > > > ----- > Please reply to the list and not to my personal email. > Those desiring my consulting or training services please feel free to > email me. > --- > "Nolite dare sanctum canibus neque mittatis margaritas vestras ante > porcos ne forte conculcent eas pedibus suis." > Cum es damnatorum possederunt porcos iens ut salire off sanguinum > cliff in abyssum?" > -- > View this message in context: > http://spssx-discussion.1045642.n5.nabble.com/Problems-with-24-tp57321 > 61p5732187.html Sent from the SPSSX Discussion mailing list archive at > Nabble.com. > > ===================== > To manage your subscription to SPSSX-L, send a message to > (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 > > > -- > Jon K Peck > jkpeck@ > <mailto: > jkpeck@ > > > > ===================== To manage your subscription to SPSSX-L, send a > message to > <mailto: > > (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 > <mailto: > > (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 > (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 ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Problems-with-24-tp5732161p5732195.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== 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 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Administrator
|
John,
Please trim your posts!!! Management
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
In reply to this post by Rick Oliver-3
I know that. That’s why I use add value labels. I’ve seen a few tears in the past when someone has inadvertently deleted all their value labels without keeping a copy of the original syntax or of the file they were working on. Ditto for missing values x (). I daren’t delete any values as I don’t have the original setup files and neither, it seems, does anyone else. This whole exchange started with a query from me whether there was a way to reverse engineer syntax from a *.sav file. Apparently there isn’t. (See thread [Reverse engineering from *.sav to *.sps] John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Rick Oliver It may or may not be worth noting that you can remove all value labels for a variable with the Value Labels command:
|
|
|
Well, without reading that thread, I can
think of a couple of way to reverse engineer dictionary definition syntax
from a .sav file, although they might be a bit convoluted, and the least
convoluted would involve Python.
Rick Oliver Senior Information Developer IBM Business Analytics (SPSS) E-mail: [hidden email] From: John F Hall <[hidden email]> To: [hidden email] Date: 05/21/2016 12:23 AM Subject: Re: Problems with 24 Sent by: "SPSSX(r) Discussion" <[hidden email]> I know that. That’s why I use add value labels. I’ve seen a few tears in the past when someone has inadvertently deleted all their value labels without keeping a copy of the original syntax or of the file they were working on. Ditto for missing values x (). I daren’t delete any values as I don’t have the original setup files and neither, it seems, does anyone else. This whole exchange started with a query from me whether there was a way to reverse engineer syntax from a *.savfile. Apparently there isn’t. (See thread [Reverse engineering from *.sav to *.sps]
John F Hall (Mr) [Retired academic survey researcher]
Email: johnfhall@... Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop
From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of Rick Oliver
It may or may not be worth noting that you
can remove all value labels for a variable with the Value Labels command:
===================== To manage your subscription to SPSSX-L, send a message to LISTSERV@...(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 John F Hall
I can’t even think about how to do it, but I suspect the product would yield thousands of lines of syntax, one per variable for each of at least: value labels missing values An example is the original setup file for the 1989 British Social Attitudes from Strathclyde University. It runs to 4155 lines for 1070 variables, each of which is separately specified: variable labels Married "Marital Status(Summary)derived" value labels Married 1"Married/living as marr" 2"Separated/divorced" 3"Widowed" 4"Never married" 9"No information" variable labels ReligSum "Religion (Summary) derived" value labels ReligSum 1"C of E,C of I,Anglican" 2"Roman Catholic" 3"Other Christian" 4"Non-Christian" 5"No religion" 8"DK/NA" Variable levels are not specified, so are default Numeric from data list. Missing values are occasionally declared, but do not appear in the *.sav file: missing values jobsatis (0,-1) .. appears in the ellipse with 0 as the value for each label. Later files using CAPI and BLAISE do not have the equivalent setup files. The problem would still be to identify lines like: 9"No information" 8"DK/NA" . . and change them to: - 9"No information" - 8"DK/NA" I need to check this against the setup for the 1987 wave, but the FA Cup Final, Crystal Palace vs Manchester United is about to start. John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop From: Rick Oliver [mailto:[hidden email]] Well, without reading that thread, I can think of a couple of way to reverse engineer dictionary definition syntax from a .sav file, although they might be a bit convoluted, and the least convoluted would involve Python.
John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Rick Oliver It may or may not be worth noting that you can remove all value labels for a variable with the Value Labels command:
===================== 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 |
|
|
The 3rd party product Stat/Transfer from Circle Systems will take an IBM SPSS Statistics SAV file and “convert” it to “SPSS Program + Ascii data.” As an example, I took “census_small.sav” from the training materials and converted it. It is a push-button operation. Here is the generated syntax, which reads the ascii file and builds the SAV file. /* This SPSS program was automatically generated by Stat/Transfer ver.13.1.487.0426 */ /* ENCODING CP1252 */ /* The following line should contain the complete path and name of your raw data file */ /* The last line of this file contains the path to your output '.sav' file */ FILE HANDLE DATA / NAME="C:/Users/Tony/Documents/temp/try.dat" LRECL=103 . DATA LIST FILE=DATA/ id 1-4 sex 5 race 6 region 7 born 8 hhsize 9-10 adults 11-12 age 13-15 agekdbrn 16-18 sibs 19-20 dwelling 21-40 (A) wrkstat 41-42 degree 43 educ 44-45 paeduc 46-47 maeduc 48-49 speduc 50-51 happy 52 marital 53 hapmar 54 agemar 55-57 satjob 58 postlife 59 health 60 class 61 sei 62-65 pasei 66-69 masei 70-73 spsei 74-77 rincome 78-79 income 80-81 polviews 82 natroad 83 natmass 84 natsci 85 natchld 86 sexornt 87 newsfrom 88-89 intecon 90 inteduc 91 intspace 92 intintl 93 intmil 94 intenvir 95 helpful 96 cantrust 97 conmedic 98 coneduc 99 confinan 100 intrhome 101 tvhours 102-103 . FORMATS sei (F5.1) / pasei (F5.1) / masei (F5.1) / spsei (F5.1) / . VARIABLE LABELS id 'R ID NUMBER' sex 'RESPONDENTS SEX' race 'RESPONDENTS RACE' region 'REGION OF INTERVIEW' born 'WAS R BORN IN THIS COUNTRY' hhsize 'NUMBER OF PERSONS IN HOUSEHOLD' adults 'HOUSEHOLD MEMBERS 18 YRS AND OLDER' age 'AGE' agekdbrn "R'S AGE WHEN 1ST CHILD BORN" sibs 'NUMBER OF BROTHERS AND SISTERS' dwelling 'TYPE OF STRUCTURE' wrkstat 'LABOR FORCE STATUS' degree 'RS HIGHEST DEGREE' educ 'HIGHEST YEAR OF SCHOOL COMPLETED' paeduc 'HIGHEST YEAR SCHOOL COMPLETED, FATHER' maeduc 'HIGHEST YEAR SCHOOL COMPLETED, MOTHER' speduc 'HIGHEST YEAR SCHOOL COMPLETED, SPOUSE' happy 'GENERAL HAPPINESS' marital 'MARITAL STATUS' hapmar 'HAPPINESS OF MARRIAGE' agemar 'AGE WHEN FIRST MARRIED' satjob 'SATISFACTION WITH JOB OR HOUSEWORK' postlife 'BELIEF IN LIFE AFTER DEATH' health 'CONDITION OF HEALTH' class 'SUBJECTIVE CLASS IDENTIFICATION' sei 'RESPONDENT SOCIOECONOMIC INDEX' pasei "R'S FATHER'S SOCIOECONOMIC INDEX " masei "R'S MOTHER'S SOCIOECONOMIC INDEX" spsei "R'S SPOUSE'S SOCIOECONOMIC INDEX" rincome 'RESPONDENTS INCOME (CATEGORY)' income 'TOTAL FAMILY INCOME (CATEGORY)' polviews 'THINK OF SELF AS LIBERAL OR CONSERVATIVE' natroad 'HIGHWAYS AND BRIDGES' natmass 'MASS TRANSPORTATION' natsci 'SUPPORTING SCIENTIFIC RESEARCH ' natchld 'ASSISTANCE FOR CHILDCARE' sexornt 'SEXUAL ORIENTATION' newsfrom 'MAIN SOURCE OF INFORMATION ABOUT EVENTS IN THE NEWS' intecon 'INTERESTED IN ECONOMIC ISSUES' inteduc 'INTERESTED IN EDUCATION ISSUES' intspace 'INTERESTED IN SPACE EXPLORATION' intintl 'INTERESTED IN INTERNATIONAL ISSUES' intmil 'INTERESTED IN MILITARY POLICY' intenvir 'INTERESTED IN ENVIRONMENTAL ISSUES' helpful 'PEOPLE HELPFUL OR LOOKING OUT FOR SELVES' cantrust 'PEOPLE CAN BE TRUSTED OR CANT BE TOO CAREFUL' conmedic 'CONFIDENCE IN MEDICINE' coneduc 'CONFIDENCE IN EDUCATION' confinan 'CONFIDENCE IN BANKS & FINANCIAL INSTITUTIONS' intrhome "INTERNET ACCESS IN R'S HOME" tvhours 'HOURS PER DAY WATCHING TV' . VALUE LABELS sex 1 'MALE' 2 'FEMALE' / race 1 'WHITE' 2 'BLACK' 3 'OTHER' / region 1 'NORTH' 2 'SOUTH' 3 'EAST' 4 'WEST' 5 'NORTH EAST' 6 'NORTH WEST' 7 'SOUTH EAST' 8 'SOUTH WEST' / born 1 'YES' 2 'NO' / hhsize 99 'REFUSED TO ANSWER' / adults 99 'REFUSED TO ANSWER' / age 999 'REFUSED TO ANSWER' / agekdbrn 998 "DON'T KNOW" 999 'REFUSED TO ANSWER' / sibs 99 'REFUSED TO ANSWER' / wrkstat 1 'WORKING FULLTIME' 2 'WORKING PARTTIME' 3 'TEMP NOT WORKING' 4 'UNEMPL, LAID OFF' 5 'RETIRED' 6 'SCHOOL' 7 'KEEPING HOUSE' 8 'OTHER' 99 'REFUSED TO ANSWER' / degree 0 'LESS THAN PRIMARY/HIGH SCHOOL' 1 'PRIMARY/HIGH SCHOOL' 2 'POST PRIMARY/HIGH SCHOOL' 3 'BACHELOR' 4 'GRADUATE' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / educ 98 "DON'T KNOW" 99 'REFUSED TO ANSWER' / paeduc 97 'IAP' 98 "DON'T KNOW" 99 'REFUSED TO ANSWER' / maeduc 97 'IAP' 98 "DON'T KNOW" 99 'REFUSED TO ANSWER' / speduc 97 'IAP' 98 "DON'T KNOW" 99 'REFUSED TO ANSWER' / happy 1 'VERY HAPPY' 2 'PRETTY HAPPY' 3 'NOT TOO HAPPY' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / marital 1 'MARRIED' 2 'WIDOWED' 3 'DIVORCED' 4 'SEPARATED' 5 'NEVER MARRIED' 9 'REFUSED TO ANSWER' / hapmar 1 'VERY HAPPY' 2 'PRETTY HAPPY' 3 'NOT TOO HAPPY' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / agemar 999 'REFUSED TO ANSWER' / satjob 0 'IAP' 1 'VERY SATISFIED' 2 'MOD. SATISFIED' 3 'A LITTLE DISSAT' 4 'VERY DISSATISFIED' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / postlife 1 'YES' 2 'NO' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / health 0 'IAP' 1 'EXCELLENT' 2 'GOOD' 3 'FAIR' 4 'POOR' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / class 1 'LOWER CLASS' 2 'WORKING CLASS' 3 'MIDDLE CLASS' 4 'UPPER CLASS' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / sei -1 'IAP' / pasei -1 'IAP' / masei -1 'IAP' / spsei -1 'IAP' / rincome 0 'IAP' 1 'LT $1000' 2 '$1000 TO 2999' 3 '$3000 TO 3999' 4 '$4000 TO 4999' 5 '$5000 TO 5999' 6 '$6000 TO 6999' 7 '$7000 TO 7999' 8 '$8000 TO 9999' 9 '$10000 - 14999' 10 '$15000 - 19999' 11 '$20000 - 24999' 12 '$25000 OR MORE' 98 "DON'T KNOW" 99 'REFUSED TO ANSWER' / income 0 'IAP' 1 'LT $1000' 2 '$1000 TO 2999' 3 '$3000 TO 3999' 4 '$4000 TO 4999' 5 '$5000 TO 5999' 6 '$6000 TO 6999' 7 '$7000 TO 7999' 8 '$8000 TO 9999' 9 '$10000 - 14999' 10 '$15000 - 19999' 11 '$20000 - 24999' 12 '$25000 OR MORE' 98 "DON'T KNOW" 99 'REFUSED TO ANSWER' / polviews 1 'EXTREMELY LIBERAL' 2 'LIBERAL' 3 'SLIGHTLY LIBERAL' 4 'MODERATE' 5 'SLGHTLY CONSERVATIVE' 6 'CONSERVATIVE' 7 'EXTREMELY CONSERVATIVE' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / natroad 1 'TOO LITTLE' 2 'ABOUT RIGHT' 3 'TOO MUCH' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / natmass 1 'TOO LITTLE' 2 'ABOUT RIGHT' 3 'TOO MUCH' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / natsci 1 'TOO LITTLE' 2 'ABOUT RIGHT' 3 'TOO MUCH' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / natchld 1 'TOO LITTLE' 2 'ABOUT RIGHT' 3 'TOO MUCH' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / sexornt 0 'IAP' 1 'GAY, LESBIAN OR HOMOSEXUAL' 2 'BISEXUAL' 3 'HETEROSEXUAL OR STRAIGHT' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / newsfrom 0 'IAP' 1 'NEWSPAPERS' 2 'MAGAZINES' 3 'INTERNET' 4 'BOOKS OTHER PRINTED MATERIALS' 5 'TV' 6 'RADIO' 7 'GOVERNMENT AGENCIES' 8 'FAMILY' 9 'FRIENDS COLLEAGUES' 10 'OTHER' 98 'DONT KNOW' 99 'REFUSED TO ANSWER' / intecon 0 'IAP' 1 'VERY INTERESTED' 2 'MODERATELY INTERESTED' 3 'NOT AT ALL INTERESTED' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / inteduc 0 'IAP' 1 'VERY INTERESTED' 2 'MODERATELY INTERESTED' 3 'NOT AT ALL INTERESTED' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / intspace 0 'IAP' 1 'VERY INTERESTED' 2 'MODERATELY INTERESTED' 3 'NOT AT ALL INTERESTED' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / intintl 0 'IAP' 1 'VERY INTERESTED' 2 'MODERATELY INTERESTED' 3 'NOT AT ALL INTERESTED' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / intmil 0 'IAP' 1 'VERY INTERESTED' 2 'MODERATELY INTERESTED' 3 'NOT AT ALL INTERESTED' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / intenvir 0 'IAP' 1 'VERY INTERESTED' 2 'MODERATELY INTERESTED' 3 'NOT AT ALL INTERESTED' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / helpful 0 'IAP' 1 'HELPFUL' 2 'LOOKOUT FOR SELF' 3 'DEPENDS' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / cantrust 0 'IAP' 1 'ALWAYS TRUSTED' 2 'USUALLY TRUSTED' 3 'USUAL NOT TRUSTED' 4 'ALWYS NOT TRUSTED' 8 "CAN'T CHOOSE" 9 'REFUSED TO ANSWER' / conmedic 0 'IAP' 1 'A GREAT DEAL' 2 'ONLY SOME' 3 'HARDLY ANY' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / coneduc 0 'IAP' 1 'A GREAT DEAL' 2 'ONLY SOME' 3 'HARDLY ANY' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / confinan 0 'IAP' 1 'A GREAT DEAL' 2 'ONLY SOME' 3 'HARDLY ANY' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / intrhome 0 'IAP' 1 'YES' 2 'NO' 8 "DON'T KNOW" 9 'REFUSED TO ANSWER' / tvhours -1 'IAP' 98 "DON'T KNOW" 99 'REFUSED TO ANSWER' / . /* User Missing Value Specifications */ MISSING VALUES hhsize (99) adults (99) age (999) agekdbrn (998, 999) sibs (99) wrkstat (99) degree (8, 9) educ (98, 99) paeduc (97, 98, 99) maeduc (97, 98, 99) speduc (97, 98, 99) happy (8, 9) marital (9) hapmar (8, 9) agemar (999) satjob (0, 8, 9) postlife (8, 9) health (8, 9, 0) class (8, 9) sei (-1) pasei (-1) masei (-1) spsei (-1) rincome (0, 98, 99) income (0, 98, 99) polviews (8, 9) natroad (8, 9) natmass (8, 9) natsci (8, 9) natchld (8, 9) sexornt (0, 8, 9) newsfrom (0, 98, 99) intecon (0, 8, 9) inteduc (0, 8, 9) intspace (0, 8, 9) intintl (0, 8, 9) intmil (0, 8, 9) intenvir (0, 8, 9) helpful (0, 8, 9) cantrust (0, 8, 9) conmedic (0, 8, 9) coneduc (0, 8, 9) confinan (0, 8, 9) intrhome (0, 8, 9) tvhours (-1, 98, 99) . SAVE OUTFILE='C:/Users/Tony/Documents/temp/try.sav'. Tony Babinec From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of John F Hall I can’t even think about how to do it, but I suspect the product would yield thousands of lines of syntax, one per variable for each of at least: value labels missing values An example is the original setup file for the 1989 British Social Attitudes from Strathclyde University. It runs to 4155 lines for 1070 variables, each of which is separately specified: variable labels Married "Marital Status(Summary)derived" value labels Married 1"Married/living as marr" 2"Separated/divorced" 3"Widowed" 4"Never married" 9"No information" variable labels ReligSum "Religion (Summary) derived" value labels ReligSum 1"C of E,C of I,Anglican" 2"Roman Catholic" 3"Other Christian" 4"Non-Christian" 5"No religion" 8"DK/NA" Variable levels are not specified, so are default Numeric from data list. Missing values are occasionally declared, but do not appear in the *.sav file: missing values jobsatis (0,-1) .. appears in the ellipse with 0 as the value for each label. Later files using CAPI and BLAISE do not have the equivalent setup files. The problem would still be to identify lines like: 9"No information" 8"DK/NA" . . and change them to: - 9"No information" - 8"DK/NA" I need to check this against the setup for the 1987 wave, but the FA Cup Final, Crystal Palace vs Manchester United is about to start. John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop From: Rick Oliver [[hidden email]] Well, without reading that thread, I can think of a couple of way to reverse engineer dictionary definition syntax from a .sav file, although they might be a bit convoluted, and the least convoluted would involve Python.
John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Rick Oliver It may or may not be worth noting that you can remove all value labels for a variable with the Value Labels command:
===================== 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 |