Hi, I've been using SQL for years but am still relatively new to
SPSS. Is there any way to efficiently manipulate SPSS data sets with
SQL commands instead of SPSS' native language? - Doing this while
running Python would be even better. I did some searches, and every
result that had SPSS and SQL on it was related to Import/Export
procedures, not straight data manipulation. I realize that up until
one or two versions ago, SPSS couldn't even handle multiple data sets
running in the same instance, so I doubt I'll have much luck with
this.

Thanks

Hello everyone,

I am new to spss, so this may be a simple question, but I appreciate
if somenone answers it.

I want to compute confidence interval for the mean of a variable, for
which only mean, standard deviation and sample size are given, but the
sample data is not available. I know I can use a Transform / Compute
variable menu and use the formula to compute standard error and
confidence interval, but I wonder whether there is another way of
computing it. If the data were available, Descriptives/Explore menu
calculate the desired confidence interval, but I don't have the data
and only have summary statistics mean,sd and n. I have seen a rowtype_
variable in spss and tried to solve the problem using it but I
couldn't. Thanks for any help.

Is there anyone know how to do trend analysis using chi-square?
Currently, I am using SAS 9.0. Could you give some SAS code? Thanks.

Hi, I'm fairly new to SPSS so please bear with me...

I have this code:

DO IF ((Gender=1) AND (EstAge <18)).
RECODE varana (0 thru 99=1) (99.0001 thru hi=2) INTO varana_n.
END IF.
DO IF ((Gender=2) AND (EstAge <18)).
RECODE varana (0 thru 99=1) (99.0001 thru hi=2) INTO varana_n.
END IF.

I know I can look at the counts for Hi and Lo by gender, etc...

I'd like to do a proportion of the HIGH COUNTS for each GENDER ...
(summarize..tables...cells...count..mean...)

Any thoughts? Any help greatly appreciated.

Thanks very much.

Does anyone know how to create an error bar chart that is not based on
raw data, but *specific* values for the point and both ends of the
confidence intervals? The distance from the upper CL and point and
lower CL and point are *not* equal.

e.g.

. -------
. |
. |
. * ____
. | |
. | |
. -------- *
. |
. ____
.
.............................................

I'm thinking I may have to use a program like Excel for this, but
thought I'd ask before moving on. Does anyone know of a user friendly
program they'd recommend?

Hi, I'm relatively new to SPSS. I've done a lot of work in FoxPro,
where you call variables by reference simply by placing a & in front
of the name. I've seen a lot of SPSS code with #'s and $'s in front
of variables, but I can't for the life of me find the page or two of
documentation that explains what these symbols mean.

1. Can anyone point me toward the appropriate syntax page in the
reference manual?

2. As a simple example, if I wanted to create twenty variables named
test93X through test113X (each equal to the square of their number),
how would I do it in SPSS?

Here's the FoxPro code:
for i = 93 to 113
VarName = 'test' + str(i) + 'X'
&VarName = i * i
next

Hi all-

Quick question. Here's my dataset...

subj var1 var2
1 1 0
1 2 1
1 3 0
2 1 1
2 2 0
2 3 0
.
.
.

I would like to collapse this dataset, such that cases 2 and 3 for
var1 are collpased to one case, with the max value on var 2 being
assigned as the new value for var2 for that collpased case.

Here's what I want...

Hi,

I'm getting an error that I can not explain, since I ran the same
procedure (MATCH FILES, adding a variable) with two other, identically
structured data sets and it worked just fine.

Basically, I'm getting the 5130 Error that accuses me of not sorting
the files properly before trying to match them. Below is the output:

here is the syntax that produces this:

Hi all, can you help?

I'm hoping that SPSS has an equivalent to the SAS procedure VARCLUS.
Varclus groups variables with similar variance and is essentially a
variable reduction technique but one, while not as complete as PCA,
that maintains the original data in a few key variables that explain
the majority of the variance in the population.

Many thanks,
Dave

Is there any way to use this command without altering the active
dataset i.e. writing out a new one instead?