SPSS CUSTOM TABLES - SAVING 'SETUP' and creating 'NETS'
Locked
SPSS CUSTOM TABLES - SAVING 'SETUP' and creating 'NETS'
 
|
Dear all,
I have a feeling that SPSS custom tables cannot be saved per se.
One can save the syntax of course but not the 'set' up?
For instance, I have around 40 survey questions, comprising 2,000 variables (many are multi choice with LOTS of response options) and around 20 breaks on these in the columns, with various NETS, suppressed categories and what not.
Having got this 'just so' I would like to keep it. It's possible, likely in fact, that I will nee to tweak it in the future (in ways that I can't predict - might be adding new variables or new breaks) but it seems that the only way I could do that would be via syntax (far form easy given the complexity of the tables) or by rebuilding the 'set up' manually in C Tables.
Can the 'set up' be saved? I think not but would love to be pleasantly surprised.
Relatedly seems that C tables cannot really do 'NETS'. e.g. I have some scale data (let's say 5 point strongly agree to strongly disagree'. I also want to have a 'NET' of agree + strongly agree. This is easily accomplished by adding a category in C tables but that is then treated literally as an additional category (i.e. as if there were 7 categories in the variable). What is needed of course is to calculate those two categories in the net seperateleey, off the base of all respondents. I achieved this in the end by a recode into a new variable but what a pain in the fundament!
|
Re: SPSS CUSTOM TABLES - SAVING 'SETUP' and creating 'NETS'
|
|
It sounds to me you need to learn how to approach jobs .. Now, SPSS Ctables aint the most detailed table program by far .. maybe you want Quantum or Wincross or something else - there are a variety of programs that literally build the tables cell by cell column by column and row by row .. will give you all the flexibility you want .. SPSS has limitations . it works with variables ... but for its cost and because it can be a pretty full functional statistical package at the same time .. TO ME .. it remains one of the best values in the market (and you tend to like what you know) .. SHEEESH I hate the days of column specs. I have jobs spanning 15 years - with tens of thousands of cases and variables (maybe into the 100s of thousands) producing near 500 - 1000 tables per run (I can run anywhere near 25 to 50 separate runs of these tables each year or more) .. The survey changes each year (in fact monthly) adds/subtracts, item bank changes, adds to coding list etc.. its all maintained in one piece. What it takes in PLANNING . And nothing does it for you - it is work .. but if you have a set of syntax to prepare your data and the tables syntax to run tables .. you should be 'reproducable' with the exception of changes to the survey .. I can go back to any job I have done in the past 20 years and reproduce it from start to finish - and I don't need to remember a thing (give me new data with the same structural characteristics - which can be determined with dictionaries - and with the exception of weighting) .. the syntax takes the raw data file I was given and runs to produce the tables .. if I need to know anything - its all in the syntax .. yes I have nets and yes its 'saved' . Ctables (SPSS) is more flexible than you may have discovered to this point. Nets (and particularly sorting) are not everything they could be - but they suffice. I have seen plenty of different 'setups' over the years - everybody does it differently (a testament to just how flexible Ctables is) and I am sure there are few things I could still learn to make it easier for me. BUT THERE IS NO SUBSTITUTE for detailing all your job in syntax including your prep . I don't use the GUI at all (with the exception of seeing the syntax for a command I may not have used in some time - but that's also what the Command Syntax Reference is for) EVERYTHING YOU DO TO A DATA FILE TO GET TO WHATEVER YOU PROCUDE SHOULD BE DOCUMENTED - AND IF YOU DO IT IN SYNTAX - ITS ALL THERE AND READY TO RUN - not only can you reproduce it but anyone else should be able to too as well. Date: Sat, 30 Jan 2016 05:39:32 -0700 From: [hidden email] Subject: SPSS CUSTOM TABLES - SAVING 'SETUP' and creating 'NETS' To: [hidden email] Dear all, I have a feeling that SPSS custom tables cannot be saved per se. One can save the syntax of course but not the 'set' up? For instance, I have around 40 survey questions, comprising 2,000 variables (many are multi choice with LOTS of response options) and around 20 breaks on these in the columns, with various NETS, suppressed categories and what not. Having got this 'just so' I would like to keep it. It's possible, likely in fact, that I will nee to tweak it in the future (in ways that I can't predict - might be adding new variables or new breaks) but it seems that the only way I could do that would be via syntax (far form easy given the complexity of the tables) or by rebuilding the 'set up' manually in C Tables. Can the 'set up' be saved? I think not but would love to be pleasantly surprised. Relatedly seems that C tables cannot really do 'NETS'. e.g. I have some scale data (let's say 5 point strongly agree to strongly disagree'. I also want to have a 'NET' of agree + strongly agree. This is easily accomplished by adding a category in C tables but that is then treated literally as an additional category (i.e. as if there were 7 categories in the variable). What is needed of course is to calculate those two categories in the net seperateleey, off the base of all respondents. I achieved this in the end by a recode into a new variable but what a pain in the fundament! View this message in context: SPSS CUSTOM TABLES - SAVING 'SETUP' and creating 'NETS' 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 |
Re: SPSS CUSTOM TABLES - SAVING 'SETUP' and creating 'NETS'
|
|
Timothy - thanks for your comments. I salute your systematic approach.
I also have 20+ years research experience in an academic environment so planning isn't something I am a stranger to. What I am lacking is not planning skills but knowledge of SPSS syntax re Ctables at a level which would enable me to quickly edit it to add in new variables or tweak any of the outputs. In the absence of that (and the time to learn it in the short term) I have to use menus to a great extent and certainly sourcing new software with an impending deadline software mid project is not an option. Thanks for the wide ranging err.. ''advice'' but you haven't really addressed any of my questions: i) can the Ctables 'set up' be saved ii) is there a way to do NETS that actually makes sense (you say that they suffice but no details on how) I suspect the answer to both of the above is ''no'' but just wanted to check with this community of experts in case I had missed something there. Regards Mike |
|
|
In reply to this post by Timothy Hennigar
Regarding NETS: if the categories you want to combine are adjacent, this is just a subtotal, which can optionally hide the subtotaled categories. More flexibly, are you aware of the postcompute functionality via the PCOMPUTE and PPROPERTIES subcommands? There is also an extension command, STATS CATEGORY ORDER that adds flexibility. This procedure creates macro definitions or multiple dichotomy sets for use
in Custom Tables in order to control the category order beyond what is built
in to that procedure. It can also create macros listing variables in order
of the frequency of occurrence of specified values. More generally, you can create template syntax via macro or, better, Python, that can have all the stock syntax with parameters that can be substituted for varying parts. On Sat, Jan 30, 2016 at 7:06 AM, Timothy Hennigar <[hidden email]> wrote:
|
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |