Reduced functionality for UNIANOVA
Locked
Reduced functionality for UNIANOVA
 
|
I've just upgraded from v19 to v21 and am intrigued that syntax UNIANOVA no longer seems to accept multiple dependent variables. In v19 the code below happily executes 3 ANOVA, one with each of participation_rate_child, hours_enrolled_child and weeks_enrolled_child as dependent variable. In version 21 it informs me that "Multiple dependent variables are not allowed for univariate analysis. Execution of this command stops.". I can find no documentation of this change although I do note that the v19 functionality is not listed in the v19 command syntax reference. Presumably the parser has been re-written between the two releases to work as advertised rather than as useful! Fortunately oneway still takes a list :-)
UNIANOVA participation_rate_child hours_enrolled_child weeks_enrolled_child BY program_type_child /METHOD=SSTYPE(3) /INTERCEPT=INCLUDE /POSTHOC=program_type_child(FREGW) /PRINT=DESCRIPTIVE /CRITERIA=ALPHA(.05) /DESIGN=program_type_child. Cheers, Simon ===================== 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 |
Re: Reduced functionality for UNIANOVA
Administrator
|
ARGHHH Simon!
I feel your pain. I suspect any number of people are going to be PO at their broken code. Best I can suggest is wrap it in a macro (add whatever specs you need). You should be happy that they haven't defiled it further by porting the output into the Abomination known as the Model Viewer <Monstrosity> (MVM henceforth)! --- DEFINE MyUniANOVA (Deplist !CharEnd ("/") /Indep !Tokens(1)). !DO !V !IN (!DepList) UNIANOVA !V BY !Indep /METHOD=SSTYPE(3) /POSTHOC=!Indep(FREGW) /INTERCEPT=INCLUDE /CRITERIA=ALPHA(0.05) /DESIGN=!Indep. !DOEND. !ENDDEFINE . MyUniANOVA DepList x y z / Indep Random . --------------------------------------------------------------------------------------------------------------------- @IBM - PLEASE STOP DOING THIS!!! YOU FORCE EVERYONE WHO IS USING UNIANOVA TO CHANGE THEIR CODE AND THAT IS PRETTY SUCKY [Filtered] ON YOUR PART! PLEASE CHANGE IT BACK! ---------------------------------------------------------------------------------------------------------------------- Looking at ver 11.5 syntax diagram (my most current version prior to 21): UNIANOVA dependent varlist [BY factor list [WITH covariate list]] Looking at ver 21 syntax diagram: UNIANOVA dependent var [BY factor list [WITH covariate list]] Same with ver 20. What is the point of the change?
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?" |
Re: Reduced functionality for UNIANOVA
|
|
What can you say but Another IMPROVEMENT!
On the other hand, a curious person might wonder whether the people who come up with these and the people who decide implement them do or have done data analysis. I vote NO, they haven't. Gene Maguin -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of David Marso Sent: Friday, March 22, 2013 1:10 AM To: [hidden email] Subject: Re: Reduced functionality for UNIANOVA ARGHHH Simon! I feel your pain. I suspect any number of people are going to be PO at their broken code. Best I can suggest is wrap it in a macro (add whatever specs you need). You should be happy that they haven't defiled it further by porting the output into the Abomination known as the Model Viewer <Monstrosity> (MVM henceforth)! --- DEFINE MyUniANOVA (Deplist !CharEnd ("/") /Indep !Tokens(1)). !DO !V !IN (!DepList) UNIANOVA !V BY !Indep /METHOD=SSTYPE(3) /POSTHOC=!Indep(FREGW) /INTERCEPT=INCLUDE /CRITERIA=ALPHA(0.05) /DESIGN=!Indep. !DOEND. !ENDDEFINE . MyUniANOVA DepList x y z / Indep Random . --------------------------------------------------------------------------------------------------------------------- @IBM - PLEASE STOP DOING THIS!!! YOU FORCE EVERYONE WHO IS USING UNIANOVA TO CHANGE THEIR CODE AND THAT IS PRETTY SUCKY [Filtered] ON YOUR PART! PLEASE CHANGE IT BACK! ---------------------------------------------------------------------------------------------------------------------- Looking at ver 11.5 syntax diagram (my most current version prior to 21): UNIANOVA *dependent varlist* [BY factor list [WITH covariate list]] Looking at ver 21 syntax diagram: UNIANOVA *dependent var* [BY factor list [WITH covariate list]] Same with ver 20. What is the point of the change? Simon Palmer-4 wrote > I've just upgraded from v19 to v21 and am intrigued that syntax > UNIANOVA no longer seems to accept multiple dependent variables. In > v19 the code below happily executes 3 ANOVA, one with each of > participation_rate_child, hours_enrolled_child and > weeks_enrolled_child as dependent variable. In version 21 it informs > me that "Multiple dependent variables are not allowed for univariate > analysis. Execution of this command stops.". I can find no > documentation of this change although I do note that the v19 functionality is not listed in the v19 command syntax reference. > Presumably the parser has been re-written between the two releases to > work as advertised rather than as useful! Fortunately oneway still > takes a list > :-) > > UNIANOVA participation_rate_child hours_enrolled_child > weeks_enrolled_child BY program_type_child > /METHOD=SSTYPE(3) > /INTERCEPT=INCLUDE > /POSTHOC=program_type_child(FREGW) > /PRINT=DESCRIPTIVE > /CRITERIA=ALPHA(.05) > /DESIGN=program_type_child. > > Cheers, > Simon > > ===================== > 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" -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Reduced-functionality-for-UNIANOVA-tp5718957p5718963.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 |
Re: Reduced functionality for UNIANOVA
|
Administrator
|
My suspicion is that after 2 decades of 'real stats real easy' marketing tripe, someone with good intentions but absolutely no clue about the consequences of broken code suddenly stopped the speeding train and speculated in a design meeting: "Oh My...face palm action figure pose... This UNIANOVA permits specification of more than one DV. Oh Gosh. What if our poor users mistake it for MANOVA and start writing up their results like a multivariate analysis?. I think we have provided more than enough rope for them to hang themselves. Better fix it before the train crashes and explodes."
Result: General nods of assent by those attending, muffled shrieks by the silent minority who knew better but were overwhelmed by the sheer power of the cranial flatulence in the room. Does IBM eat their own dog food? ---
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?" |
Re: Reduced functionality for UNIANOVA
|
|
Hey David,
> My suspicion is that after 2 decades of 'real stats real easy' marketing > tripe, someone with good intentions but absolutely no clue about the > consequences of broken code suddenly stopped the speeding train and > speculated in a design meeting: "Oh My...face palm action figure pose... > This UNIANOVA permits specification of more than one DV. Oh Gosh. What if > our poor users mistake it for MANOVA and start writing up their results like > a multivariate analysis?. I think we have provided more than enough rope > for them to hang themselves. Better fix it before the train crashes and > explodes." I did wonder about that, but as I noted in my original post the V19 syntax manual does not allow for a variable list even though the software accepted it. So I speculated that the parser had been re-implemented by someone who followed the documented specs rather than the actual code. :-) Thanks for your macro suggestion, I considered writing something similar but realised I could use ONEWAY to achieve the same outcomes I needed with less typing. Cheers, Simon ===================== 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 |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |