Hi:

Just some code to calculate RSE for different type of estimates (means, correlations, RR and OR):

* MEANS * * Dataset*. DATA LIST FREE/group(f8.0) glucose(f8.0). BEGIN DATA 1 51 1 56 1 58 1 60 1 62 1 63 1 65 1 68 1 72 1 73 2 60 2 65 2 66 2 68 2 68 2 69 2 73 2 75 2 78 2 80 3 69 3 73 3 74 3 78 3 79 3 79 3 82 3 85 3 87 3 88 4 70 4 75 4 76 4 77 4 79 4 80 4 82 4 86 4 88 4 89 END DATA. VAR LABEL group 'AcType' /glucose 'Glucose levels'. VALUE LABELS group 1 'Control' 2 'Respiratory' 3 'Metabolic' 4'Mixed'.

* Computing RSE *. AGGREGATE /OUTFILE=* /BREAK=group /mean = MEAN(glucose) /sd = SD(glucose) /n=N. COMPUTE rse=100*sd/SQRT(n)/mean. VAR LABEL rse 'Relat. Std. Error. (%)'. EXEC.

* Report *. SUMMARIZE /TABLES=group mean sd n rse /FORMAT=LIST NOCASENUM TOTAL /TITLE='Case Summaries' /MISSING=VARIABLE /CELLS=NONE.

* CORRELATIONS *

* A summary dataset must be constructed *. DATA LIST LIST /n(F8.0) cor(F8.2). BEGIN DATA 131 0.51 129 0.48 155 0.30 121 0.21 111 0.60 119 0.46 112 0.22 145 0.25 END DATA. VAR LABEL cor 'Pearson`s correlation coefficient' /n 'Sample size'.

* Computing RSE *. COMPUTE rse=100*SQRT((1-cor**2)/(n-2))/cor. VAR LABEL rse 'Relat. Std. Error. (%)'. EXEC.

* Report *. SUMMARIZE /TABLES=cor n rse /FORMAT=LIST NOCASENUM TOTAL /TITLE='Case Summaries' /MISSING=VARIABLE /CELLS=NONE.

* RSE from Z transformed data *. COMPUTE zcor=0.5*LN((1+cor)/(1-cor)). COMPUTE #zse=1/SQRT(n-3). COMPUTE zrse=100*#zse/zcor. VAR LABEL zcor 'Z transformed cor.'. VAR LABEL zrse 'Relat. Std. Error. (%)'. EXEC.

* Report *. SUMMARIZE /TABLES=zcor n zrse /FORMAT=LIST NOCASENUM TOTAL /TITLE='Case Summaries' /MISSING=VARIABLE /CELLS=NONE.

* RR *

* Dataset *. DATA LIST LIST /study(A10) ai(F8.0) n1i(F8.0) ci(F8.0) n0i(F8.0). BEGIN DATA Kirkman 14 40 24 40 VanGelder 3 27 8 30 Nashan-A 51 171 73 166 Kahan 57 173 79 173 Lawen 9 59 17 64 Ponticelli 31 168 50 172 Vincenti 28 126 47 134 Nashan-B 39 140 63 133 END DATA. VAR LABEL study 'Trial ID name' /ai 'Events (acute rejections) in treated group' /n1i 'Total N treated group' /ci 'Events (acute rejections) in control group' /n0i 'Total N control group'.

* Computing statistics *. COMPUTE rr=(ai/n1i)/(ci/n0i). COMPUTE invrr=1/rr. COMPUTE selog=SQRT((1/ai)-(1/n1i)+(1/ci)-(1/n0i)). COMPUTE rse=100*selog/ABS(LN(rr)). VAR LABEL rse 'Relat. Std. Error. (%)'. VAR LABEL rr 'Relative Risk'. VAR LABEL invrr 'Inverse Relative Risk (1/RR)'. VAR LABEL selog 'SE log(RR)'.

* Report *. SUMMARIZE /TABLES=rr invrr selog rse /FORMAT=LIST NOCASENUM TOTAL /TITLE='Case Summaries' /MISSING=VARIABLE /CELLS=NONE.

* OR *

* Dataset *. DATA LIST LIST /study(A10) ai(F8.0) n1i(F8.0) ci(F8.0) n0i(F8.0). BEGIN DATA Kirkman 14 40 24 40 VanGelder 3 27 8 30 Nashan-A 51 171 73 166 Kahan 57 173 79 173 Lawen 9 59 17 64 Ponticelli 31 168 50 172 Vincenti 28 126 47 134 Nashan-B 39 140 63 133 END DATA. VAR LABEL study 'Trial ID name' /ai 'Events (acute rejections) in treated group' /n1i 'Total N treated group' /ci 'Events (acute rejections) in control group' /n0i 'Total N control group'.

* Computing statistics *. COMPUTE bi=n1i-ai. COMPUTE di=n0i-ci. COMPUTE orr=(ai*di)/(ci*bi). COMPUTE invorr=1/orr. COMPUTE selog=SQRT((1/ai)+(1/bi)+(1/ci)+(1/di)). COMPUTE rse=100*selog/ABS(LN(orr)). VAR LABEL rse 'Relat. Std. Error. (%)'. VAR LABEL orr 'Odds Ratio'. VAR LABEL invorr 'Inverse Odds Ratio (1/OR)'. VAR LABEL selog 'SE log(OR)'.

* Report *. SUMMARIZE /TABLES=orr invorr selog rse /FORMAT=LIST NOCASENUM TOTAL /TITLE='Case Summaries' /MISSING=VARIABLE /CELLS=NONE.

Regards

Marta García-Granero

"Arthur J. Kendall" ha escrito:

> I know of no place where it is directly calculated. > Using the definition at > http://www.cdc.gov/nchs/datawh/nchsdefs/relativestandarderror.htm > > You would simply divide the standard error of whatever estimate you are > interested in by the estimate itself. If you have only a few, you would > do by hand. Depending on the kind of estimate you are interested in you > might be able to save estimates and their sd's or se' from some > procedures into a new system file. > > Hope this helps. > > Art > Art@DrKendall.org > Social Research Consultants > University Park, MD USA > (301) 864-5570 > > Jason Payne wrote: > > Dear All, > > > > Does anyone know if this is possible to calculate the RSE in SPSS... and if so, where i might find reference to the syntax that will assist me. > > > > Regards > > > > Jason Payne > > AUSTRALIA > > > > ____________________________________ > > Jason Payne > > Violence, Property Crime and Drugs > > Australian Institute of Criminology > > GPO Box 2944 Canberra ACT 2601 > > > > E-mail: jason.payne@aic.gov.au > > Tel +61 2 6260 9283 Fax +61 2 6260 9201 > > > > Visit the AIC homepage www.aic.gov.au > > > > > > -----Original Message----- > > From: Vincent Toh [mailto:vincent04@pacific.net.sg] > > Sent: Thursday, 11 December 2003 12:43 PM > > To: SPSSX-L@LISTSERV.UGA.EDU > > Subject: Re: Manipulation of Pivot Table by Scripting > > > > > > Dear All, > > > > I am re-posting this mail, as I think the previous one that I sent didn't > > turn out well. > > > > I have a scripting question which I hope someone would be able to assist. > > > > I would like to make changes to the values of each row such that: > > > > 1) if n < 3, suppress result by replacing all the values by "s" . > > 2) if n = 3, show only median values, suppress lower and upper quartile > > values by replacing with "s" . > > > > An example is as follows: > > > > The pivot table that I have created is: > > > > CATEGORY n Lower Quartile Median Upper Quartile > > E&E 34 2300 2349 2461 > > T&E 2 2337 3013 3689 > > P&P 3 3690 3740 3790 > > A&A 0 . . > > . > > > > and my required format is as follows: > > > > CATEGORY n Lower Quartile Median Upper Quartile > > E&E 34 2300 2349 > > 2461 > > T&E 2 s s > > s > > P&P 3 s 3740 > > s > > A&A 0 s s > > s > > > > > > > > As I am unable to attach the documents and post it to the list, should > > anyone require, I could also send it to your emails directly. > > > > Would greatly appreciate if anyone could enlighten! > > > > Thanks and Best Regards, > > Vincent Toh > > Singapore > > > > > > ********************************************************************** > > This email and any files transmitted with it are confidential and > > intended solely for the use of the individual or entity to whom they > > are addressed. If you have received this email in error please notify > > the system manager. > > > > This footnote also confirms that this email message has been swept by > > MIMEsweeper for the presence of computer viruses. > > > > www.mimesweeper.com > > ********************************************************************** > >