bootstrapping with logistic regression on SPSS 17.0

[Miriam Haverkamp]

Hello,
I am new to this list (and SPSS) and have the following problem:
I have a database out of which we are trying to predict death of children
with pneumonia in the developing world. As we luckily do nto ahve that
many deaths 43 out of 958 we would like to do the bootstrapping procedure
to develop a repdcition model. We tried to split the database but the
model gets too unstable...
So< here is my coding (adjusted from what the help function in sPSS
provides and it just does not work, any suggestions? Alternatively if this
is easier for you you might send me your coding so I can adjust mine.

PRESERVE.
SET TVARS NAMES.
DATASET DECLARE bootstrap_example.
OMS /DESTINATION VIEWER=NO /TAG='suppressall'.
OMS
  /SELECT TABLES
  /IF COMMANDS=['Regression'] SUBTYPES=['Coefficients']
  /DESTINATION FORMAT=SAV OUTFILE='bootstrap_example'
  /COLUMNS DIMNAMES=['Variables'  'Statistics']
  /TAG='reg_coeff'.

DATASET ACTIVATE DataSet1.

DEFINE regression_bootstrap (samples=!TOKENS(1)
                             /depvar=!TOKENS(1)
                             /indvars=!CMDEND)

COMPUTE dummyvar=1.
AGGREGATE
  /OUTFILE=* MODE=ADDVARIABLES
  /BREAK=dummyvar
  /filesize=N.
DO REPEAT !other=1 !TO !samples.
SET SEED RANDOM.
WEIGHT OFF.
FILTER OFF.
DO IF $casenum=1.
COMPUTE #samplesize=filesize.
COMPUTE #filesize=filesize.
END IF.
DO IF (#samplesize>0 and #filesize>0).
- COMPUTE sampleWeight=rv.binom(#samplesize, 1/#filesize).
- COMPUTE #samplesize=#samplesize-sampleWeight.
- COMPUTE #filesize=#filesize-1.
ELSE.
- COMPUTE sampleWeight=0.
END IF.
WEIGHT BY sampleWeight.
FILTER BY sampleWeight.
REGRESSION
  /STATISTICS COEFF
  /DEPENDENT !depvar
  /METHOD=ENTER !indvars.
!DOEND
!ENDDEFINE.

DATASET ACTIVATE DataSet1.

Regression_bootstrap
   samples=100
   depvar=Earlydeath
   indvars=Leucocytosis(1)Pulse(1) Treatmentarm(1)hypotensive (1).

OMSEND.
DATASET ACTIVATE bootstrap_example.
FREQUENCIES
  VARIABLES=Leucocytosis_B Leucocytosis_BetaPulse_B Pulse_Beta
Treatmentarm_B Treatmentarm_Beta hypotensive_B hypotensive_Beta
meanO2_10percentdecrements_B meanO2_10percentdecrements_Beta
  /FORMAT NOTABLE
  /PERCENTILES= 2.5 97.5
  /HISTOGRAM  NORMAL.
RESTORE.

I have also added the coding for the logistic regression from the full
dataset, just so you can see what the regression is supposed to look like:


DATASET ACTIVATE DataSet1.
LOGISTIC REGRESSION VARIABLES Earlydeath
  /METHOD=ENTER Leucocytosis Treatmentarm meanO2_10percentdecrements
hypotensive Pulse
  /CONTRAST (Leucocytosis)=Indicator(1)
  /CONTRAST (Pulse)=Indicator(1)
  /CONTRAST (Treatmentarm)=Indicator(1)
  /CONTRAST (hypotensive)=Indicator(1)
  /PRINT=GOODFIT CI(95)
  /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

Thanks!

Miriam

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Michael Kruger]

Miriam,

I don't know if this helps but I have attached the syntax for a
bootstrapped logistic regression that I used on a research project
earlier this year. I found the basic code
for it on  Raynald  SPSS Tools website that I have placed the link for
below:

http://pages.infinit.net/rlevesqu/SampleSyntax.htm#Bootstrap

The last portion of the syntax that begins with the AGGREGATE command
was used to create a data file with specific portions of the output that
I wanted to save to a new SPSS data file to obtain specific parameter
estimates. Good luck.

Michael Kruger

>
> PRESERVE.
> SET TVARS NAMES.
>
> ***first OMS command just suppresses Viewer output***.
> OMS /DESTINATION VIEWER=YES.
>
> DATASET DECLARE bootstrap_example.
>
> ***select regression coefficients tables and write to data file***.
> ***Note that DIMNAMES values vary based on output language***.
> ***/COLUMNS SEQUENCE=[R2 C1] will achieve the same result in all languages***.
>
> OMS /SELECT TABLES
>     /IF COMMANDS=['Logistic Regression'] SUBTYPES = ['Variables in the Equation']
>    /DESTINATION FORMAT=SAV OUTFILE='bootstrap_example'
>    /COLUMNS DIMNAMES=['Variables'  'Statistics']
>    /TAG='logeg_coeff'.
>
> ***define a macro to draw samples with replacement and run Regression commands***.
> DEFINE  regression_bootstrap (samples=!TOKENS(1)
>                                            /depvar=!TOKENS(1)
>                                            /indvars=!CMDEND)
>
> COMPUTE dummyvar=1.
> AGGREGATE
>   /OUTFILE = * MODE = ADDVARIABLES
>   /BREAK=dummyvar
>   /filesize=N.
> !DO !other=1 !TO !samples
> SET SEED RANDOM.
> WEIGHT OFF.
> FILTER OFF.
> DO IF $casenum=1.
> - COMPUTE #samplesize=filesize.
> - COMPUTE #filesize=filesize.
> END IF.
> DO IF (#samplesize>0 and #filesize>0).
> - COMPUTE sampleWeight=rv.binom(#samplesize, 1/#filesize).
> - COMPUTE #samplesize=#samplesize-sampleWeight.
> - COMPUTE #filesize=#filesize-1.
> ELSE.
> - COMPUTE sampleWeight=0.
> END IF.
> WEIGHT BY sampleWeight.
> FILTER BY sampleWeight.
>
> LOGISTIC REGRESSION VARIABLES  survival
>   /METHOD = ENTER Race_Cat Best_GA Steroid AC PPROM FL_Cat
>   /CONTRAST (Race_Cat)=Deviation(1)  /CONTRAST (Steroid)=Indicator(1)  /CONTRAST (PPROM)=Indicator(1)
>   /CONTRAST (FL_Cat)=Indicator(1)
>   /PRINT = CI(95)
>   /CRITERIA = PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .
>
> !DOEND
> !ENDDEFINE.
>
> ***insert any valid path\data file name***.
> GET FILE='e:\miscellaneous\ms_rbs_2009.sav'.
> ***Call the macro, and specify number of samples, dependent variable, and independent variables***.
> regression_bootstrap
>    samples=1000
>
> OMSEND.
>
> DATASET ACTIVATE bootstrap_example.
>
> RESTORE.
>
> AGGREGATE
>   /OUTFILE='J:\Documents\FILES\RBS\MS_RBS_2009_AGGR.sav'
>   /BREAK=Var2
>   /B_mean=MEAN(B)
>   /B_mi=Min(B)
>   /B_max=Max(B)
>   /ExpB_mean=MEAN(ExpB)
>   /Exp_B_min=Min(ExpB)
>   /Exp_B_max=Max(ExpB)
>   /Sig_mean=mean(Sig)
>   /S.E_mean=MEAN(S.E)
>   /Upper_mean=MEAN(Upper)
>   /Lower_mean=MEAN(Lower)
>   /N_BREAK=N.
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Bruce Weaver]

Michael Kruger wrote

Miriam,

I don't know if this helps but I have attached the syntax for a
bootstrapped logistic regression that I used on a research project
earlier this year. I found the basic code
for it on  Raynald  SPSS Tools website that I have placed the link for
below:

http://pages.infinit.net/rlevesqu/SampleSyntax.htm#Bootstrap

The last portion of the syntax that begins with the AGGREGATE command
was used to create a data file with specific portions of the output that
I wanted to save to a new SPSS data file to obtain specific parameter
estimates. Good luck.

Michael Kruger

That is the old version of Raynald's site, and I don't think it is maintained any longer.  For the new version, go to:

   http://www.spsstools.net/
   http://www.spsstools.net/SampleSyntax.htm#Bootstrap