http://www.nnseek.com/e/comp.softsys.stat.spss/ Posts for comp.softsys.stat.spss Wed, 27 Aug 2008 12:12:56 PDT http://www.nnseek.com/ http://www.nnseek.com/img/64.png 64 64 NNSeek http://www.nnseek.com/e/comp.softsys.stat.spss/what_s_wrong_with_this_syntax_69095918t.html http://www.nnseek.com/e/comp.softsys.stat.spss/what_s_wrong_with_this_syntax_69095918t.html
DO IF sex=1 .
EXECUTE.

IF (age >=10 & age <11 & HeightPercentileforage>=5 &
HeightPercentileforage<10 & SystolicBP <97) HSBP=1.
EXECUTE .
IF (age >=10 & age <11 & HeightPercentileforage>=5 &
HeightPercentileforage<10 & SystolicBP >= 97 & SystolicBP <111)
HSBP=2.
EXECUTE .

Posted In: comp.softsys.stat.spss

7 Comments

Reply

]]> Wed, 27 Aug 2008 12:12:56 PDT http://www.nnseek.com/e/comp.softsys.stat.spss/violating_the_homogeneity_of_variance_assumption_69065454t.html http://www.nnseek.com/e/comp.softsys.stat.spss/violating_the_homogeneity_of_variance_assumption_69065454t.html
I examined scores (below, average, high) as my DV for three age
groups. I did a one-way ANOVA. Results of the Test for Homogeneity of
Variances are:

Levene Statistic=24.258
df1 = 2
df2=4102
sig=.000

My ANOVA results are F(2,4102)=15.789, p=.000. Eta squared (SS between
groups/SS total)=.008.

I believe I may have violated the homogeneity of variance assumption
since its sig is less than .05, but I'm not sure I know what this
means. Help please? Thanks.

Posted In: comp.softsys.stat.spss

10 Comments

Reply

]]> Wed, 27 Aug 2008 07:47:15 PDT http://www.nnseek.com/e/comp.softsys.stat.spss/version_16_gui_problem_69059822t.html http://www.nnseek.com/e/comp.softsys.stat.spss/version_16_gui_problem_69059822t.html I couldn't find anything on it in the archives just now. Just thought
I'd mention it in case anyone else is running into the same problem.

In version 16, I often find that the GUI is only accessible from the
data window. If I try to access it from a syntax window, for example,
quite often, nothing happens. I have to bring the data window to the
front, and access the GUI from there. And before anyone asks, YES, I
have applied all of the relevant patches & hotfixes. (Windoze version
of SPSS 16.0.2, running under XP Professional, SP2.)

--
Bruce Weaver
bweaver@lakeheadu.ca
www.angelfire.com/wv/bwhomedir
"When all else fails, RTFM."

Posted In: comp.softsys.stat.spss

2 Comments

Reply

]]> Wed, 27 Aug 2008 06:59:39 PDT http://www.nnseek.com/e/comp.softsys.stat.spss/compute_percentiles_from_z_scores_69026286t.html http://www.nnseek.com/e/comp.softsys.stat.spss/compute_percentiles_from_z_scores_69026286t.html
Has anyone a macro/syntax or a formula I could use to calculate
percentiles & percentile ranks from Z-scores?

Thanks

Posted In: comp.softsys.stat.spss

2 Comments

Reply

]]> Tue, 26 Aug 2008 22:44:42 PDT http://www.nnseek.com/e/comp.softsys.stat.spss/median_across_variables_in_new_column_68996846t.html http://www.nnseek.com/e/comp.softsys.stat.spss/median_across_variables_in_new_column_68996846t.html gmail.com
wrote:

>Thanks, Rich.
>
>So, one remaining question: it seems as if it's always necessary to
>FLIP the data in order to calcualate the median in SPSS.

Huh? Read my response. Getting the median in another
form, one line at a time, is what the (untested) code was about...

>
>In my case, this means I'm not able to perform this since I have data
>sheets with 5 million rows...

... and it works the same for 5 million rows as for 300,000 rows.



--
Rich Ulrich

Posted In: comp.softsys.stat.spss

no comments

Reply

]]> Tue, 26 Aug 2008 14:30:52 PDT http://www.nnseek.com/e/comp.softsys.stat.spss/median_across_variables_in_new_column_68933870t.html http://www.nnseek.com/e/comp.softsys.stat.spss/median_across_variables_in_new_column_68933870t.html > On Mon, 25 Aug 2008 11:46:41 -0700 (PDT), graefe.andreas@gmail.com
> wrote:
>
>> Hi,
>>
>> I open a new post since I'm afraid the other one (http://
>> groups.google.com/group/comp.soft-sys.stat.spss/browse_thread/thread/
>> 8726eacae27ea0a7#) is too old to get any answers.
>>
>> I have a huge data file (about 300,000 rows) with ten variables, v1 to
>> v10. For these ten variables, I want to calculate the median in a new
>> column / variable.
>>
>> Experimenting with copy and paste from other posts, I came to this:
>
> [snip, what follows]
>
> Perhaps I miss the question, but I think you want the
> median of ten variables that exist in one line. For that,
> I do not imagine how VARSTOCASES could help; it seems
> like that proper first step should be FLIP, to transpose
> the data. (Is that what that VARSTOCASES achieves?)

Yes. If you have 10 variables per subject, VARSTOCASES will
change it to one variable per subject with 10 rows of data, and an
index variable with numbers 1-10. Then you can use AGGREGATE
(with ID as the break variable) to compute the median for each
subject, which is what the OP wanted.


> However, I would not do that with 300,000 lines.

I don't understand your point here, Rich.

--
Bruce Weaver
bweaver@lakeheadu.ca
www.angelfire.com/wv/bwhomedir
"When all else fails, RTFM."

Posted In: comp.softsys.stat.spss

1 Comment

Reply

]]> Mon, 25 Aug 2008 19:37:35 PDT http://www.nnseek.com/e/comp.softsys.stat.spss/can_anyone_help_with_optimal_scaling_in_spss_68915438t.html http://www.nnseek.com/e/comp.softsys.stat.spss/can_anyone_help_with_optimal_scaling_in_spss_68915438t.html > Hi,
>
> can anybody give me some advice on optimal scaling in SPSS?
> Specifically, on the non-linear principal component analysis and the
> generalised canonical correlation analysis (overals).
>
>
> I find the optimal scaling procedures extremely useful. My problem
> ist
> that I really need to rotate the axis in order to understand and
> interpret the results. I know that a different kind of inteprretation
> is suggested, but I am not the visual type and I have problems
> interpretating unrotated dimensions. If I do employ subsequent
> varimax
> rotation, I get great results with a very clear structure.
>
>
> So far so good. The problem is, however, that the optimal scaling
> module in SPSS does not offer an intergrated rotation procedure so
> that right now I have to compute the varimax rotation by manually
> inserting the values in the matrix. The disadvantage is that I can
> never save the individual scores of my participants on the rotated
> dimensions. But that is what I want because I want to employ rotated
> scores for further validity analysis.
>
>
> Does anybody know a syntax in SPSS which makes it possible to
> directly
> rotate and then also to get the individual scores on the rotated
> dimensions? It is very easy to save the scores on the unrotated
> dimensions, but I need the rotated ones....
>
>
> I really would appreciate any advice, thank´s a lot!
>
>
> Peter

Peter, when posting the same message to multiple groups, please
cross-post (i.e., one post to the multiple groups) rather than
multi-posting (separate posts).

--
Bruce Weaver
bweaver@lakeheadu.ca
www.angelfire.com/wv/bwhomedir
"When all else fails, RTFM."

Posted In: comp.softsys.stat.spss

no comments

Reply

]]> Sun, 24 Aug 2008 17:51:07 PDT http://www.nnseek.com/e/comp.softsys.stat.spss/crosstabs_multiple_response_sets_help_68903918t.html http://www.nnseek.com/e/comp.softsys.stat.spss/crosstabs_multiple_response_sets_help_68903918t.html I have this table (questions) as a part of a survey (university
project)
http://img398.imageshack.us/my.php?image=capturetj5.jpg

It is used to know which methods the employers used to recruit their
workers and employees and what method is used most of each type of
employee (skilled, unskilled etc..)

I need help of how to design the SPSS (variables etc..) So in the end
I can have as an output:
1. The most 3 important methods for recruitment in general
2. The most 3 important methods for recruitment for each type of
workers
3. Frequencies for each method of recruitment

and Also what SPSS tests i need to run?
btw, the others in question 1 can have different answers obviously do
i have to create a variable in SPSS for each answer given? ( it may be
over 20)
Using SPSS 16.0 btw
Thanks for your help

Posted In: comp.softsys.stat.spss

no comments

Reply

]]> Sun, 24 Aug 2008 14:08:11 PDT http://www.nnseek.com/e/comp.softsys.stat.spss/can_anyone_help_with_optimal_scaling_68861166t.html http://www.nnseek.com/e/comp.softsys.stat.spss/can_anyone_help_with_optimal_scaling_68861166t.html
can anybody give me some advice on optimal scaling in SPSS?
Specifically, on the non-linear principal component analysis and the
generalised canonical correlation analysis (overals).

I find the optimal scaling procedures extremely useful. My problem ist
that I really need to rotate the axis in order to understand and
interpret the results. I know that a different kind of inteprretation
is suggested, but I am not the visual type and I have problems
interpretating unrotated dimensions. If I do employ subsequent varimax
rotation, I get great results with a very clear structure.

So far so good. The problem is, however, that the optimal scaling
module in SPSS does not offer an intergrated rotation procedure so
that right now I have to compute the varimax rotation by manually
inserting the values in the matrix. The disadvantage is that I can
never save the individual scores of my participants on the rotated
dimensions. But that is what I want because I want to employ rotated
scores for further validity analysis.

Does anybody know a syntax in SPSS which makes it possible to directly
rotate and then also to get the individual scores on the rotated
dimensions? It is very easy to save the scores on the unrotated
dimensions, but I need the rotated ones....

I really would appreciate any advice, thank´s a lot!

Peter

Posted In: comp.softsys.stat.spss

no comments

Reply

]]> Sun, 24 Aug 2008 04:08:35 PDT http://www.nnseek.com/e/comp.softsys.stat.spss/how_to_calculate_a_3_way_interaction_68836078t.html http://www.nnseek.com/e/comp.softsys.stat.spss/how_to_calculate_a_3_way_interaction_68836078t.html I could really use someone's help. I am new to SPSS and I need someone
to dumb it down for me. I want to calculate a three way interaction
between three categorical variables: A, B and C with a dependent
variable X. How do I do this on SPSS? I have very limited stats
knowledge and, like I said, I am very new to SPSS. So, please provide
me with a "SPSS for dummies" explanation. Thank you!!

Posted In: comp.softsys.stat.spss

1 Comment

Reply

]]> Sat, 23 Aug 2008 19:09:51 PDT