Re: 'Forest-plots' or Error bar-charts in SPSS

[Marta García-Granero]

Hi Laurie

LP> Thanks again for your very helpful suggestions. Please could you let me know
LP> which of the files in the META-SPSS.zip file I need to use

It depends. If you read the Readmefirst.txt file that comes along the
files

http://www.spsstools.net/Syntax/MetaAnalysis/ReadMeFirst.txt

You will see that each independent file has a precise purpose (OR with
Peto method, unstandardized means heterogeneous effects...). Each one
of them has a MATRIX part that computes the summary measures, Cochran
Q test and other statistics. The results are then exported to the
active dataset and a HILO graph is drawn using a chart template. You
need to create your own chart templates in C:\Temp\ folder (those
where created with SPSS 11), unless you modify the syntax to replace
C:\Temp by another folder. Basically, the templates do the following:

- Transpose chart coordinates
- Add a reference line at y=1 (if measure if log-normal, like OR or
  RR) or at y=0 (if the measure is means or risk difference)
- Set y scale to log (for OR & RR, not for meandiffs or RD)

Basically, all files have a syntax quite close to this (the one I
sent to you in my previous message):

GRAPH /HILO(SIMPLE)=VALUE( ciup cilow measure ) BY study.

Then edit the graph to transpose the coordinates system, change the
scale to log (if needed), add a reference line at 0 or 1 (depending on
the summary measure), save the chart template an apply it to new
graphs:

GRAPH /HILO(SIMPLE)=VALUE( ciup cilow measure ) BY study
 /TEMPLATE='Your template file.sgt'.

LP>  and whether they should be opened in the syntax window or in some
LP> other way? I am working with SPSS 12.0.1.

They are simple syntax files (.sps) and two chart templates. No .sbs
nor fancier files are used.

A newer version of these file (collapsed into a small set of macros
with a lot of options) has been posted at Code Center (by me), but, I
don't know what happens, it looks the either SPSS webpage has
undergone changes again (I found a fast and direct way to the
Algorithms page that wasn't there before, but I can't find how to
reach the Code Center) or either my age-related brain deterioration
rate is rocketing: I just can't find a "clicking road" from www.spss.com
homepage to the Code Center. The page saif that SPSS 14 was needed,
but don't worry, I haven't added anything that will make the code fail
in SPSS 12/13.

If you succeed reaching the page, tell me and I'll bookmark it.

Regards,
Marta

LP> -----Original Message----- From: Marta García-Granero
LP> [mailto:[hidden email]] Sent: 15 September 2006 01:44
LP> Subject: Re: 'Forest-plots' or Errror bar-charts in SPSS

LP>> Thanks to those who responded with very helpful messages in
LP>> response to my queries about using SPSS for meta-analysis. Now
LP>> I'm at the stage of reporting the results, please can anyone help
LP>> me with a way of drawing up a 'forest plot' like those produced
LP>> by the Cochrane RevMan software using SPSS? Perhaps someone knows
LP>> an ingenious way to import SPSS data to RevMan in order to do
LP>> this? Or, maybe someone has a macro to do the same thing using
LP>> SPSS? Alternatively, if someone could help me draw up a simple
LP>> error bar-chart in SPSS, showing the effect size with its 95%
LP>> confidence interval for each study, that would also help.

LP> * Sample dataset (processed data )*.
LP> DATA LIST LIST/ trial(F4) year(A5) study(A10) measure cilow ciup percwi(4
LP> F8.3).
LP> BEGIN DATA
LP>  1 "1989" "Hodnett  " .502 .262  .962   4.940
LP>  2 "1991" "Kennell  " .352 .216  .575   8.694
LP>  3 "1992" "Bréart-Fr" .785 .483 1.276   8.878
LP>  4 "1992" "Bréart-Bg" .811 .653 1.007  44.598
LP>  5 "1997" "Gagnon   " .867 .573 1.311  12.237
LP>  6 "1998" "Langer   " .280 .203  .384  20.654
LP>  7 "    " "Total    " .594 .514  .687 100.000
LP> END DATA.

LP> A) Assuming you have SPSS 14:

LP> SORT CASES BY trial(D).
LP> STRING YearAndStudy(A30).
LP> COMPUTE YearAndStudy=CONCAT(RTRIM(year)," ",study).
LP> COMPUTE RefLine=1.

LP> GGRAPH
LP>   /GRAPHDATASET NAME="graphdataset" VARIABLES=YearAndStudy ciup cilow
LP> measure percwi RefLine
LP>   MISSING=LISTWISE REPORTMISSING=NO
LP>  /GRAPHSPEC SOURCE=INLINE.

LP> * If the effect you are measuring is OR or RR, then this syntax can
LP>  do the task *.

LP> BEGIN GPL
LP>  SOURCE: s=userSource(id("graphdataset"))
LP>  DATA:   YearAndStudy=col(source(s),   name("YearAndStudy"),
LP> unit.category())
LP>  DATA:   ciup=col(source(s),    name("ciup"))
LP>  DATA:   cilow=col(source(s),   name("cilow"))
LP>  DATA:   measure=col(source(s), name("measure"))
LP>  DATA:   percwi=col(source(s),  name("percwi"))
LP>  DATA:   RefLine=col(source(s), name("RefLine"))
LP>  COORD:  transpose(rect(dim(1,2), transpose()))
LP>  GUIDE:  axis(dim(2), label(" Favours treatment               Favours
LP> Control"))
LP>  SCALE:  cat(dim(1))
LP>  SCALE:  log(dim(2))
LP>  ELEMENT:
LP> interval(position(region.spread.range(YearAndStudy*(cilow+ciup))),
LP>   shape(shape.line), color(color.black))
LP>  ELEMENT: point(position(YearAndStudy*measure), shape(shape.square),
LP>   size(percwi), color.interior(color.black))
LP>  ELEMENT: line(position(YearAndStudy*RefLine), shape(shape.line))
LP> END GPL.

LP> If your meta-analyzing means, then the reference line should be set at
LP> 0 [COMPUTE RefLine=0] and the scale should be linear instead of
LP> logarithmic [SCALE: linear(dim(2))].

LP> B) If you don't have SPSS 14, but older, then use the simpler syntax
LP> (uglier graph) you can find in the collection of meta-analytic files I
LP> wrote time ago:

LP> http://www.spsstools.net/Syntax/MetaAnalysis/META-SPSS.ZIP

LP> Basically:

LP> GRAPH /HILO(SIMPLE)=VALUE( ciup cilow measure ) BY study.

LP> (same sample dataset is used).

LP> Then edit the graph to transpose the coordinates system, change the
LP> scale to log (if needed), add a reference line at 0 or 1 (depending on
LP> the summary measure), save the chart template an apply it to new
LP> graphs:

LP> GRAPH /HILO(SIMPLE)=VALUE( ciup cilow measure ) BY study
LP>  /TEMPLATE='Your template file'.

LP> --
LP> Regards,
LP> Dr. Marta García-Granero,PhD           mailto:[hidden email]
LP> Statistician

LP> ---
LP> "It is unwise to use a statistical procedure whose use one does
LP> not understand. SPSS syntax guide cannot supply this knowledge, and it
LP> is certainly no substitute for the basic understanding of statistics
LP> and statistical thinking that is essential for the wise choice of
LP> methods and the correct interpretation of their results".

LP> (Adapted from WinPepi manual - I'm sure Joe Abrahmson will not mind)

[Marta García-Granero]

Hi again

MGG> A newer version of these file (collapsed into a small set of macros
MGG> with a lot of options) has been posted at Code Center (by me), but, I
MGG> don't know what happens, it looks the either SPSS webpage has
MGG> undergone changes again (I found a fast and direct way to the
MGG> Algorithms page that wasn't there before, but I can't find how to
MGG> reach the Code Center) or either my age-related brain deterioration
MGG> rate is rocketing: I just can't find a "clicking road" from www.spss.com
MGG> homepage to the Code Center. The page said that SPSS 14 was needed,
MGG> but don't worry, I haven't added anything that will make the code fail
MGG> in SPSS 12/13.

OK, it was my brain, after all (blushing):

http://www.spss.com/devcentral/index.cfm?pg=downloadDet&dId=35

(My excuses to SPSS webmaster).

Marta

[Sebastián Daza]

dear list,
I found the way to run without problems the syntax spss from ultraedit:
1) the application tdrunsyntax.exe works without problems in spss 14, one should only run spss before executing
2) so that it works the same as in textpad the following macro should be created that replaces the function of the parameter $clip, because there is not equivalent for $ clip in UltraEdit. But this is non problem. You can create a macro in a macro file which is set to be automatically loaded when UltraEdit is started. The macro contains following lines:

IfSel
Copy
Else
ToggleBookmark
SelectAll
Copy
EndSelect
GotoBookmark
ToggleBookmark
EndIf
RunTool "case-sensitive name of the user tool"

Assign the hot key you normally use to run your user tool to the macro.
It works me without problems. Another tip: A multiline tag can be created by using ^p in the tag line for CR/LF.

regards!!