Math Forum / Math Software / SPSS / April 2006

I have a question about SPSS. The univariate ANOVA
allows one to specify the variables as fixed or
random. The repeated-measures ANOVA however does not.
What is the default in a repeated-measures design?

I suppose that most of you will find this question very simple but I am
a novice in statistics so please be patient with me.
I did a survey on 45 people. Each participant had to choose between 5
different answers. This question was repeated 10 times per user and

How do I write up results for tests like Mann Whitney and Wilcoxon?
There seems to be so much variation in published reserach reports that
it'd be great if someone could tell me how exactly to write these (what
to include and omit- mean rank or sum of ranks?  Number of ties?).  If

A colleague is being questioned on her sample size.  I can't find my old
paper file with the simple formula written out.  Help within SPSS didn't
provide that either as far as I could find.  I'd rather not have a program
calculate it for me - I'd rather use that simple formula I ...

I am trying to assign VAR LABS based on a variable value.
First I created some MACROS
Define !M1()
VAR LABS Q2 "Lance".

I'm building ARIMA model and I came accross this warning:
Our tests have determined that the estimated model lies close to the
boundary of the invertibility region. Although the moving average
parameters are probably correctly estimated, their standard errors and

Hi.  SAS to SPSS convert here...
In SAS I could place date ranges into cohorts with a 'proc format':
proc format;
    value finyrfmt

Some advice please...
I have a (normally-distributed ratio) measure of muscle strength
measured 4 times over 6 years (irregular intervals) and I have
testosterone (normally-distributed ratio) measured at baseline and at 3

in conducting multiple regression, is it important that the data be
normally distributed (or close to, via statistical transformations). If
so, does this apply BOTH to the predictors and criterion variables?
Thanks.

Hi, I'm developing syntax to be run from the Production Facility
however along with the way I would like to save out to Excel and PPT.
Does anyone have a syntax that does that? Thaks, LP

Hi, I've run a straight forward logistic regression using
interval/ratio scores to predict a dichotomous outcome. However, the
result I have is a bit confusing in that there is one significant
predictor, but it's associated odds ratio in predicting the outcome is

I would be very grateful if someone could direct me to how to perform a
median polish in SPSS 14. I can't find anything in the help contents.
Thanks,
E

Dear colleagues,
San Pablo - CEU University in collaboration with other five
universities (Málaga, Politécnica de Madrid, País Vasco, Rey Juan
Carlos, and Castilla La Mancha), nine companies, CSIC and IEEE

i just need confirmation on this as I 'think' I know the answer.
What is the difference between a partial correlation, and a linear
multiple regression (using 2 predictors)? I'm thinking that although
you get additional stats in the output with a multiple regression, that

  I am trying perform several iterations for various values of the
variables i, j and k. Also I want the loop to ignore the case when d ge
1 and go again to the beginning of the loop with the next possible
values for i, j and k etc.