Date:         Tue, 12 Apr 2005 12:04:29 +0200
Reply-To:     Marta García-Granero
              <biostatistics@terra.es>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Marta García-Granero
              <biostatistics@terra.es>
Organization: Asesoría Bioestadística
Subject:      Re: ridiculous error with log scaling!
In-Reply-To:  <425B9A1E.4973.53D5EA@uvw.uni-bayreuth.de>
Content-Type: text/plain; charset=ISO-8859-1

Hi Thomas,

I checked it, and you are absolutely right! BTW, I also tried an IGRAPH,
but I could not find how to change the Y scale to log...

DATA LIST LIST/x y.
BEGIN DATA
1 0.001
2 0.01
3 0.1
4 1
5 10
6 100
7 1000
END DATA.

GRAPH
  /SCATTERPLOT(BIVAR)=x WITH y
  /MISSING=LISTWISE .
* Edit graph & change Y scale to Log (wrong results) *.

COMPUTE logy = LG10(y) .
EXECUTE .
GRAPH
  /SCATTERPLOT(BIVAR)=x WITH logy
  /MISSING=LISTWISE .
* Graph is OK (straight line) *.

(The same code works OK in SPSS 11).

In other recent message to the list I complained about SPSS 12/13
graphing system. I had a collection of templates I used for
meta-analysis (others for survival analysis, regression,
interactions...). Most of them required a change to log scale. When I
tried to update them to SPSS 12, some features could not be applied
anymore, some of them behaved strangely (now I understand), and, I
NEVER NEVER NEVER succeeded trying to apply a chart template using:
/TEMPLATE='C:\Templatename.sgt'.

I like some SPSS 12 features, like OMS, but I HATE the graphs. I use
SPSS 12 for statistical analyses, but I keep SPSS 11.0 in another
computer for graphing purposes (that's annoying sometimes, because of
incompatibility of output files).

--
Regards,

Marta García-Granero              mailto:biostatistics@terra.es

TS> I found an absolutely ridiculous error in SPSS 12.0.1, 12.0.2 and
TS> 13.0.1!

TS> Type in the values:
TS> x y
TS> 1 0.001
TS> 2 0.01
TS> 3 0.1
TS> 4 1
TS> 5 10
TS> 6 100
TS> 7 1000

TS> Now make a plot of it. Line, Bar or Scatter. Change the scale of the
TS> y-axis to logarithmic.

TS> Remembering basic maths, the dots should be on a straight line.
Indeed, the dots >>= 1 are lying on a straight line. But the rest is a
TS> mess!! The system seams to try linear scale for values smaller than
TS> one!

TS> German Support told me there will be no quick solution for that! Am I
TS> the only having that problem?