Batch Frequency Tables

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Batch Frequency Tables

Bob Dozier
Ok I decided after having read messages for about a year to at least get
my toe wet (if not take the plunge) in the SPSS programing world and
have run into a problem with which I could use some assistance.

I admit:
1) I do not know what I am doing,
2) It is most certainly a stupid novice error and
3) Am most certainly breaking more than one of the "10 commandments" of
statistical analysis and/or SPSS program.

That having been said please have mercy.


First the important basic stuff:

1) I am using SPSS 15 with extensions, one of which is tables.

2) I am working with an Emergency Medical Services WEB based
"satisfaction" type survey. It is made up of 4 point "Very Satisfied" to
"Very Dissatisfied"  (14 questions) and "Strongly Agree" to "Strongly
Disagree" (8 questions)  questions,  2 point "Important" or "Not
Important " (13 questions) questions,  classification/demographic
multiple choice questions(6 questions)  and open-ended commit text
fields (3 questions).

3) At the moment I have 572 completed survey which represent 11% of the
target population and my boss says "I don't care about this "sample
size" stuff, we use what we get."

I have a preliminary (can you say fishing) program that I use to
preview the data coming in (so I have some idea where I am going to go
with the analysis). It basically:
1) correlated everything that can be correlated
2) creates basic statistic and frequency tables with charts on every
response category and
3) creates crosstab tables and bar charts on everything by 4 categories
of certification types.

To do this I have created a template for the  bar charts and created a
batch file to run the commands. The code is basic. In abbreviate form it
says:
1) GET File=....
 2) NONPAR Corr /Variables=  ... /PRINT=KENDALL TWOTAIL
NOSIG/MISSING=PARWISE.
3) Frequency Variables=  ... /STATISTICS=MEAN MEDIAN MODE SUM SKEWNESS
SESKEW KURTOSIS SEKURT/BARCHART FREQ/ORDER=VARIABLE.
4) CROSSTABS /TABLES= ... BY Certification/FORMAT=AVALUE TABLES
/STATISTIC=BTAU /CELLS=COUNT ROW COLUMN TOTAL /COUNT ROUND CELL
/BARCHART.

Now the problem:

About half way through the last (the crosstab) portion of this program
SPSS appears to lose the template for the frequency bar charts and
reverts to something very very vanilla. The Warning message I get say
"an error occurred while applying a chart template."   It uses the
template for some variables and completes the program for all variables
so I do not think the problem is the code. Am I pushing for to much and
getting a memory problem or what?

Thanks
bob dozier

=====================
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
Reply | Threaded
Open this post in threaded view
|

Re:Batch Frequency Tables

Jerabek Jindrich
Hello Bob,

I must say I do not see a solution just now, anyway I would like to make some suggestion how to debug a longer SPSS syntax . And your program seems to be rather long and difficult.

1) Check that your program runs without errors and warnings. Yes, there is the warning about chart template, but there might be more errors:
Open your syntax file and then a draft output (ALT+ F/ N / R or with mouse MENU/FILE/NEW/Draft Output). Then run all your syntax. This prints output into a textfile, instead of the usual Viewer output. Then use find(CTRL F) in the draft output window, look for each ERROR and WARNING message and repair the syntax. It is much easier to look for errors in text than in normal Output viewer. There is a chance that there is a preceding error and debugging syntax this way could solve the problem. Anyway this checking should be done with every long syntax.

2) As soon as  only the chart template warning remains, look up which part of syntax caused it. Make SPSS log your syntax commands in output - MENU/EDIT/OPTIONS/VIEWER/and tick "Display commands in the log" - or insert command SET Printback=On  into your syntax. Run the syntax and you will see the warning together with the command that caused it. Post the piece of syntax together with small data sample to the list.

3) You wrote
Am I pushing for to much and
> getting a memory problem or what?
Yes it might be sometimes a problem. Just run only the small piece of syntax that resulted in warning. If it runs OK,  unsufficent HW sources are suspected.

good luck
Jindra





> ------------ Původní zpráva ------------
> Od: Bob Dozier <[hidden email]>
> Předmět: Batch Frequency Tables
> Datum: 29.4.2008 16:59:44
> ----------------------------------------
> Ok I decided after having read messages for about a year to at least get
> my toe wet (if not take the plunge) in the SPSS programing world and
> have run into a problem with which I could use some assistance.
>
> I admit:
> 1) I do not know what I am doing,
> 2) It is most certainly a stupid novice error and
> 3) Am most certainly breaking more than one of the "10 commandments" of
> statistical analysis and/or SPSS program.
>
> That having been said please have mercy.
>
>
> First the important basic stuff:
>
> 1) I am using SPSS 15 with extensions, one of which is tables.
>
> 2) I am working with an Emergency Medical Services WEB based
> "satisfaction" type survey. It is made up of 4 point "Very Satisfied" to
> "Very Dissatisfied"  (14 questions) and "Strongly Agree" to "Strongly
> Disagree" (8 questions)  questions,  2 point "Important" or "Not
> Important " (13 questions) questions,  classification/demographic
> multiple choice questions(6 questions)  and open-ended commit text
> fields (3 questions).
>
> 3) At the moment I have 572 completed survey which represent 11% of the
> target population and my boss says "I don't care about this "sample
> size" stuff, we use what we get."
>
> I have a preliminary (can you say fishing) program that I use to
> preview the data coming in (so I have some idea where I am going to go
> with the analysis). It basically:
> 1) correlated everything that can be correlated
> 2) creates basic statistic and frequency tables with charts on every
> response category and
> 3) creates crosstab tables and bar charts on everything by 4 categories
> of certification types.
>
> To do this I have created a template for the  bar charts and created a
> batch file to run the commands. The code is basic. In abbreviate form it
> says:
> 1) GET File=....
>  2) NONPAR Corr /Variables=  ... /PRINT=KENDALL TWOTAIL
> NOSIG/MISSING=PARWISE.
> 3) Frequency Variables=  ... /STATISTICS=MEAN MEDIAN MODE SUM SKEWNESS
> SESKEW KURTOSIS SEKURT/BARCHART FREQ/ORDER=VARIABLE.
> 4) CROSSTABS /TABLES= ... BY Certification/FORMAT=AVALUE TABLES
> /STATISTIC=BTAU /CELLS=COUNT ROW COLUMN TOTAL /COUNT ROUND CELL
> /BARCHART.
>
> Now the problem:
>
> About half way through the last (the crosstab) portion of this program
> SPSS appears to lose the template for the frequency bar charts and
> reverts to something very very vanilla. The Warning message I get say
> "an error occurred while applying a chart template."   It uses the
> template for some variables and completes the program for all variables
> so I do not think the problem is the code. Am I pushing for to much and
> getting a memory problem or what?
>
> Thanks
> bob dozier
>
> =====================
> 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