Confidence Interval

[Rubier Perez]

Hello all,
I am a relatively new user in SPSS (I have SPSS Statistics 17.0), and
I'd like to know how to carry out estimate Confidence Interval for the
Variance, and for the proportion.
Thanks


----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.



--

Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

[Marta Garcia-Granero]

El 16/11/2012 22:14, Rubier Perez escribió:
> Hello all,
> I am a relatively new user in SPSS (I have SPSS Statistics 17.0), and
> I'd like to know how to carry out estimate Confidence Interval for the
> Variance, and for the proportion.
Hola Rubier (greetings from Spain).

Use these macros:

DEFINE IC1PROP(!POSITIONAL !TOKENS(1) /!POSITIONAL !TOKENS(1)).
DATASET NAME Datos.
DATASET DECLARE Resultados WINDOW=HIDDEN.
PRESERVE.
SET ERRORS=NONE RESULTS=NONE.
MATRIX.
COMPUTE num= !1.
COMPUTE den= !2.
COMPUTE p=num/den .
COMPUTE z = 1.959964.
COMPUTE zsq = z*z.
COMPUTE x1 = 2*num+zsq .
COMPUTE x2 = z*(zsq+4*num&*(1-p))&**0.5 .
COMPUTE x3 = 2*(den+zsq) .
COMPUTE x4 = (x1 - x2) / x3 .
COMPUTE x5 = (x1 + x2) / x3 .
COMPUTE vnames={'X','N','Porcenta','Inferior','Superior'}.
SAVE {num,den,100*p,100*x4,100*x5} /OUTFILE=Resultados /NAMES=vnames.
END MATRIX.
RESTORE.
DATASET ACTIVATE Resultados.
FORMAT X N(F8) Porcenta TO Superior (PCT4.2).
VAR LABEL Porcenta 'Porcentaje'.
OMS /SELECT TABLES
  /IF COMMANDS='Summarize' SUBTYPES='Case Processing Summary'
  /DESTINATION VIEWER=NO.
SUMMARIZE
  /TABLES=ALL
  /FORMAT=LIST NOCASENUM NOTOTAL
  /TITLE='IC95% para una proporción (*)'
  /CELLS=NONE.
OMSEND.
DATASET ACTIVATE Datos.
DATASET CLOSE Resultados.
ECHO '(*) Método de Wilson'.
!ENDDEFINE.


DEFINE TESTCHI2 (!POS=!TOKENS(1) /Sigma=!TOKENS(1)).
DATASET COPY WorkingData WINDOW=HIDDEN.
DATASET ACTIVATE WorkingData .
RANK VARIABLES=!1(A) /N INTO N /PRINT=NO.
DO IF $casenum EQ 1.
. COMPUTE ChiLow = IDF.CHISQ(0.025,N-1) .
. COMPUTE ChiUpp = IDF.CHISQ(0.975,N-1) .
END IF.
MATRIX.
PRINT /TITLE='TEST DE CHI-CUADRADO E IC 95% PARA UNA VARIANZA'.
GET Data /VAR=!1.
GET ChiLow /VAR=ChiLow /MISSING=OMIT.
GET ChiUpp /VAR=ChiUpp /MISSING=OMIT.
COMPUTE Varianza=!Sigma**2.
COMPUTE N=NROW(Data).
COMPUTE S2=(CSSQ(Data)-(CSUM(Data)**2)/N)/(N-1).
COMPUTE Low95=(n-1)*S2/ChiUpp.
COMPUTE Upp95=(n-1)*S2/ChiLow.
PRINT {Low95,Upp95;SQRT(Low95),SQRT(Upp95)}
  /FORMAT='F8.3'
  /CLABEL='Inferior','Superior'
  /RLABEL='Varianza','Desv tip'
  /TITLE='Intervalo de confianza (95%) para la varianza y la desviación
típica'.
COMPUTE Chi2=(n-1)*S2/Varianza.
DO IF S2 GT varianza.
. COMPUTE Chi2Sig=1-CHICDF(Chi2,(N-1)).
. PRINT /TITLE='S>Sigma: test de una cola a la derecha'.
ELSE.
. COMPUTE Chi2Sig=CHICDF(Chi2,(N-1)).
. PRINT /TITLE='S<Sigma: test de una cola a la izquierda'.
END IF.
PRINT {S2,Varianza,chi2,Chi2Sig}
  /FORMAT='F8.3'
  /CLABEL='S²','Sigma²','Chi²','Sig.'
  /TITLE='Estadísticos'.
END MATRIX.
DATASET CLOSE WorkingData.
!ENDDEFINE.

Best regards,
Prof. Marta Gª-Granero
Universidad de Navarra
SPAIN

=====================
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

What do others use when SPSS graphics and Excel don't seem adequate or
elegant - displaying multiple dimensions, for example. Open source
recommendations would be best but don't want to limit options to that.

Thanks,
        JLPH

=====================
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]

Hi Marta.  Nice macros--and great to see Wilson's method for the CI of a proportion.  Of course, you could fairly easily add another argument for users to provide the desired confidence level (if not 95, which could be the default).  (I used 95 instead of .95 for consistency with other procedures in SPSS -- e.g., the /STATISTICS sub-command for REGRESSION has CI(95).)

Cheers,
Bruce

Marta García-Granero-2 wrote

El 16/11/2012 22:14, Rubier Perez escribió:
> Hello all,
> I am a relatively new user in SPSS (I have SPSS Statistics 17.0), and
> I'd like to know how to carry out estimate Confidence Interval for the
> Variance, and for the proportion.
Hola Rubier (greetings from Spain).

Use these macros:

DEFINE IC1PROP(!POSITIONAL !TOKENS(1) /!POSITIONAL !TOKENS(1)).
DATASET NAME Datos.
DATASET DECLARE Resultados WINDOW=HIDDEN.
PRESERVE.
SET ERRORS=NONE RESULTS=NONE.
MATRIX.
COMPUTE num= !1.
COMPUTE den= !2.
COMPUTE p=num/den .
COMPUTE z = 1.959964.
COMPUTE zsq = z*z.
COMPUTE x1 = 2*num+zsq .
COMPUTE x2 = z*(zsq+4*num&*(1-p))&**0.5 .
COMPUTE x3 = 2*(den+zsq) .
COMPUTE x4 = (x1 - x2) / x3 .
COMPUTE x5 = (x1 + x2) / x3 .
COMPUTE vnames={'X','N','Porcenta','Inferior','Superior'}.
SAVE {num,den,100*p,100*x4,100*x5} /OUTFILE=Resultados /NAMES=vnames.
END MATRIX.
RESTORE.
DATASET ACTIVATE Resultados.
FORMAT X N(F8) Porcenta TO Superior (PCT4.2).
VAR LABEL Porcenta 'Porcentaje'.
OMS /SELECT TABLES
  /IF COMMANDS='Summarize' SUBTYPES='Case Processing Summary'
  /DESTINATION VIEWER=NO.
SUMMARIZE
  /TABLES=ALL
  /FORMAT=LIST NOCASENUM NOTOTAL
  /TITLE='IC95% para una proporción (*)'
  /CELLS=NONE.
OMSEND.
DATASET ACTIVATE Datos.
DATASET CLOSE Resultados.
ECHO '(*) Método de Wilson'.
!ENDDEFINE.


DEFINE TESTCHI2 (!POS=!TOKENS(1) /Sigma=!TOKENS(1)).
DATASET COPY WorkingData WINDOW=HIDDEN.
DATASET ACTIVATE WorkingData .
RANK VARIABLES=!1(A) /N INTO N /PRINT=NO.
DO IF $casenum EQ 1.
. COMPUTE ChiLow = IDF.CHISQ(0.025,N-1) .
. COMPUTE ChiUpp = IDF.CHISQ(0.975,N-1) .
END IF.
MATRIX.
PRINT /TITLE='TEST DE CHI-CUADRADO E IC 95% PARA UNA VARIANZA'.
GET Data /VAR=!1.
GET ChiLow /VAR=ChiLow /MISSING=OMIT.
GET ChiUpp /VAR=ChiUpp /MISSING=OMIT.
COMPUTE Varianza=!Sigma**2.
COMPUTE N=NROW(Data).
COMPUTE S2=(CSSQ(Data)-(CSUM(Data)**2)/N)/(N-1).
COMPUTE Low95=(n-1)*S2/ChiUpp.
COMPUTE Upp95=(n-1)*S2/ChiLow.
PRINT {Low95,Upp95;SQRT(Low95),SQRT(Upp95)}
  /FORMAT='F8.3'
  /CLABEL='Inferior','Superior'
  /RLABEL='Varianza','Desv tip'
  /TITLE='Intervalo de confianza (95%) para la varianza y la desviación
típica'.
COMPUTE Chi2=(n-1)*S2/Varianza.
DO IF S2 GT varianza.
. COMPUTE Chi2Sig=1-CHICDF(Chi2,(N-1)).
. PRINT /TITLE='S>Sigma: test de una cola a la derecha'.
ELSE.
. COMPUTE Chi2Sig=CHICDF(Chi2,(N-1)).
. PRINT /TITLE='S<Sigma: test de una cola a la izquierda'.
END IF.
PRINT {S2,Varianza,chi2,Chi2Sig}
  /FORMAT='F8.3'
  /CLABEL='S²','Sigma²','Chi²','Sig.'
  /TITLE='Estadísticos'.
END MATRIX.
DATASET CLOSE WorkingData.
!ENDDEFINE.

Best regards,
Prof. Marta Gª-Granero
Universidad de Navarra
SPAIN

=====================
To manage your subscription to SPSSX-L, send a message to
<email>LISTSERV@.UGA (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

[Art Kendall]

I use healthlandscape.org to map geocoded data. It is free if you do not need to upload very large sets of data. It has a vast array of geocoded data.

Otherwise, I do not know of specific software to recommend.

However,  what do you mean by multiple multiple dimensions?

There are parallel coordinate plots i you do not need a fixed scale for all the variables.

I have been after SPSS to include profile graphs, and cartograms for a couple of decades but no luck.
Profile plots are similar to parallel coordinate plots except all of the variables are one the same scale. The are used in reporting standardized test scores, profiles for clusters, repeated measures, MANOVA if the DV are rescaled, etc.

What I do is  post what I hope is a clear description of what I want and usually someone has a way using GPL to do it. For example, I posted my efforts at doing profile plots and ViAnn was kind enough to improve on my efforts although there are a couple of spacing issues around the frame that we were not able to tweak.

GPL is a very powerful way to create graphs.  Sometimes what works is to use the GUI to paste syntax  for something close to what I want.  Then I use the graph editor to tweak it.  I can then save a template that can be used in the syntax for the original graph.

If you describe what you want to do, perhaps others know of software that does what you want.  On the other hand someone may have already written the GPL to do it.

HTH






On 11/17/2012 8:30 AM, John Lee Holmes wrote:

What do others use when SPSS graphics and Excel don't seem adequate or
elegant - displaying multiple dimensions, for example. Open source
recommendations would be best but don't want to limit options to that.

Thanks,
        JLPH

=====================
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

===================== 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

Thanks very much for taking time on this, Dr. Kendall. To answer your question and clarify mine I meant the kind of data visualization items as discussed in this somewhat dated article (pasted abstract below in case you don’t want to visit the link) from a decade ago:

http://digital.cs.usu.edu/~erbacher/publications/BSC303.pdf

Pastizzo MJ, Erbacher RF and Feldman LB, Multidimensional data visualization. Behav Res Methods Instrum Comput 34: 158–162(2002).

 

Abstract

Historically, data visualization has been limited primarily to 2 dimensions (e.g.,

histograms, scatter plots). Available software packages (e.g., Data Desk® 6.1, MatLab®

6.1, SAS©-JMPTM 4.04, SPSS© 10.0) are capable of producing 3-D scatter plots with

(varying degrees of) user interactivity. We constructed our own data visualization

application with The Visualization Toolkit (Schroeder, Martin, & Lorensen, 1998) and

Tcl/Tk to display multivariate data through the application of glyphs (Ware, 2000). A

glyph is a visual object onto which many data parameters may be mapped, each with a

different visual attribute (e.g., size, color). We used our Multi-Dimensional Data

Viewer to explore data from several psycholinguistic experiments. The graphical

interface provides flexibility when users dynamically explore the multi-dimensional

image rendered from raw experimental data. We highlight advantages of multidimensional

data visualization and consider some potential limitations.

 

In the body of the article the authors mention that “The capability to display simultaneously 4 (or more) continuous variables or to dynamically select variable ranges is less common” and I was hoping for some recommendations that update this because I assume it is a less true statement now. And some software are better designed for flexible presentation than others. Thanks again, If others have input on this as well that would be much appreciated.

-          JLPH

 

      

 

 

From: Art Kendall [mailto:[hidden email]]
Sent: Saturday, November 17, 2012 5:32 PM
To: [hidden email]
Cc: [hidden email]
Subject: Re: [SPSSX-L] Any recommendations for Graphics software?

 

I use healthlandscape.org to map geocoded data. It is free if you do not need to upload very large sets of data. It has a vast array of geocoded data.

Otherwise, I do not know of specific software to recommend.

However,  what do you mean by multiple multiple dimensions?

There are parallel coordinate plots i you do not need a fixed scale for all the variables.

I have been after SPSS to include profile graphs, and cartograms for a couple of decades but no luck.
Profile plots are similar to parallel coordinate plots except all of the variables are one the same scale. The are used in reporting standardized test scores, profiles for clusters, repeated measures, MANOVA if the DV are rescaled, etc.

What I do is  post what I hope is a clear description of what I want and usually someone has a way using GPL to do it. For example, I posted my efforts at doing profile plots and ViAnn was kind enough to improve on my efforts although there are a couple of spacing issues around the frame that we were not able to tweak.

GPL is a very powerful way to create graphs.  Sometimes what works is to use the GUI to paste syntax  for something close to what I want.  Then I use the graph editor to tweak it.  I can then save a template that can be used in the syntax for the original graph.

If you describe what you want to do, perhaps others know of software that does what you want.  On the other hand someone may have already written the GPL to do it.

HTH

Art Kendall

Social Research Consultants

On 11/17/2012 8:30 AM, John Lee Holmes wrote:

What do others use when SPSS graphics and Excel don't seem adequate or

elegant - displaying multiple dimensions, for example. Open source

recommendations would be best but don't want to limit options to that.

Thanks,

        JLPH

=====================

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

 

[Art Kendall]

Visualization is indeed a great help in exploring and analyzing data. That article refers to SPSS10.  SPSS21 is the current version.

 If you click <help> and type "exploring visualizations" in the edit box you come to info on

Exploring Visualizations

There are several interactive tools for exploring visualizations. Exploring is done through Explore mode. Explore mode allows you to:

• Animate visualizations.

• Rotate 3-D visualizations.

• View tooltips to determine exact data values.

 I have not used the animation so can not speak about it. The only dynamic graphic I have used is the 3D scatter plot. Which can differentiate data point according to shape, color, size. You can get a few additional dimensions depending on how clear distinctions are among the combinations of data point characteristic are practical for the particular application. Something close to the experimental data visualization in the article can be done.

For static visualization,  parallel coordinate plots, panels of graphs, and  matrices of graphs can also help. Different juxtapositions for comparison can be done by tweaking the syntax. On today's machines it is possible to tweak the syntax and rerun it is a few seconds, as a workaround for the lack of dynamics.

After using algorithmic explorations, e.g.,  clustering, multidimensional scaling, factor analysis, correspondence analysis, many of the variables that can be SAVEd can be explored.

It goes without saying that the residuals from application of the general linear model should be visualized.


perhaps other list members have used the animation feature and can speak about it.



On 11/17/2012 9:52 AM, John Lee Holmes wrote:

Thanks very much for taking time on this, Dr. Kendall. To answer your question and clarify mine I meant the kind of data visualization items as discussed in this somewhat dated article (pasted abstract below in case you don’t want to visit the link) from a decade ago:

http://digital.cs.usu.edu/~erbacher/publications/BSC303.pdf

Pastizzo MJ, Erbacher RF and Feldman LB, Multidimensional data visualization. Behav Res Methods Instrum Comput 34: 158–162(2002).

 

Abstract

Historically, data visualization has been limited primarily to 2 dimensions (e.g.,

histograms, scatter plots). Available software packages (e.g., Data Desk® 6.1, MatLab®

6.1, SAS©-JMPTM 4.04, SPSS© 10.0) are capable of producing 3-D scatter plots with

(varying degrees of) user interactivity. We constructed our own data visualization

application with The Visualization Toolkit (Schroeder, Martin, & Lorensen, 1998) and

Tcl/Tk to display multivariate data through the application of glyphs (Ware, 2000). A

glyph is a visual object onto which many data parameters may be mapped, each with a

different visual attribute (e.g., size, color). We used our Multi-Dimensional Data

Viewer to explore data from several psycholinguistic experiments. The graphical

interface provides flexibility when users dynamically explore the multi-dimensional

image rendered from raw experimental data. We highlight advantages of multidimensional

data visualization and consider some potential limitations.

 

In the body of the article the authors mention that “The capability to display simultaneously 4 (or more) continuous variables or to dynamically select variable ranges is less common” and I was hoping for some recommendations that update this because I assume it is a less true statement now. And some software are better designed for flexible presentation than others. Thanks again, If others have input on this as well that would be much appreciated.

-          JLPH

 

      

 

 

From: Art Kendall [[hidden email]]
Sent: Saturday, November 17, 2012 5:32 PM
To: [hidden email]
Cc: [hidden email]
Subject: Re: [SPSSX-L] Any recommendations for Graphics software?

 

I use healthlandscape.org to map geocoded data. It is free if you do not need to upload very large sets of data. It has a vast array of geocoded data.

Otherwise, I do not know of specific software to recommend.

However,  what do you mean by multiple multiple dimensions?

There are parallel coordinate plots i you do not need a fixed scale for all the variables.

I have been after SPSS to include profile graphs, and cartograms for a couple of decades but no luck.
Profile plots are similar to parallel coordinate plots except all of the variables are one the same scale. The are used in reporting standardized test scores, profiles for clusters, repeated measures, MANOVA if the DV are rescaled, etc.

What I do is  post what I hope is a clear description of what I want and usually someone has a way using GPL to do it. For example, I posted my efforts at doing profile plots and ViAnn was kind enough to improve on my efforts although there are a couple of spacing issues around the frame that we were not able to tweak.

GPL is a very powerful way to create graphs.  Sometimes what works is to use the GUI to paste syntax  for something close to what I want.  Then I use the graph editor to tweak it.  I can then save a template that can be used in the syntax for the original graph.

If you describe what you want to do, perhaps others know of software that does what you want.  On the other hand someone may have already written the GPL to do it.

HTH

Art Kendall

Social Research Consultants

On 11/17/2012 8:30 AM, John Lee Holmes wrote:

What do others use when SPSS graphics and Excel don't seem adequate or

elegant - displaying multiple dimensions, for example. Open source

recommendations would be best but don't want to limit options to that.

Thanks,

        JLPH

=====================

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

 

===================== 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

[Vik Rubenfeld]

Have a look at gnuplot.

On Nov 17, 2012, at 5:30 AM, John Lee Holmes wrote:

> What do others use when SPSS graphics and Excel don't seem adequate or
> elegant - displaying multiple dimensions, for example. Open source
> recommendations would be best but don't want to limit options to that.
>
> Thanks,
>        JLPH
>
> =====================
> 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

=====================
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]

Hello JLPH.  This is not an answer to your question, but please see this old thread.

http://spssx-discussion.1045642.n5.nabble.com/A-request-Please-compose-a-new-message-to-start-a-new-thread-td4864781.html

To facilitate this, members who post via their e-mail program need to add the address of the SPSSX-L list to their address-books (if they have not done so), and then simply compose a new message.

Thanks.  ;-)

[e-mail] wrote

What do others use when SPSS graphics and Excel don't seem adequate or
elegant - displaying multiple dimensions, for example. Open source
recommendations would be best but don't want to limit options to that.

Thanks,
        JLPH

=====================
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

[Marta Garcia-Granero]

Hi Bruce:

Long time no see. Knee deep in Stata (all my class notes need BIG
modifications to switch from SPSS to Stata). I thought I might indulge
in a relapse into old adictions ;)

See below the adaptation of Wilson's method to include CL (95 or 99).
Just for fun... big sight. I turned it to DE (*) some time ago, just in
case anyone wants it too.

(*) Deprecated Excel ;)

------------------------------------------------------------
BTW, if you like confidence intervals as measures of effect size, I
recommend you Robert Newcombe's last book (a bit heavy on maths, I kind
of like that, but I agree it could be too much for non-teachers/non-math
addicts): "Confidence intervals for proportions and related measures of
effect size", CRC Press. ISBN 978-1-4398-1278-5. I convinced the
University library to buy one, an it is all mine (until January 4th,
when I'm supposed to return it and let other teachers get a glimpse at
it). His effect size measures for Mann-Whitney and Wilcoxon tests are great!

Best regards,
Marta GG
Stata-rookie

DEFINE IC1PROP(!POSITIONAL !TOKENS(1) /!POSITIONAL
!TOKENS(1)/!POSITIONAL !DEFAULT(95) !TOKENS(1)).
DATASET NAME Data.
DATASET DECLARE Results WINDOW=HIDDEN.
PRESERVE.
SET ERRORS=NONE RESULTS=NONE.
MATRIX.
DO IF (!3 EQ 95) OR (!3 EQ .95). *Just in case people use the fraction *.
- COMPUTE z = 1.959964.
ELSE IF (!3 EQ 99) OR (!3 EQ .99).
- COMPUTE z = 2.575829.
END IF.
COMPUTE num= !1.
COMPUTE den= !2.
COMPUTE p=num/den .
COMPUTE zsq = z*z.
COMPUTE x1 = 2*num+zsq .
COMPUTE x2 = z*(zsq+4*num&*(1-p))&**0.5 .
COMPUTE x3 = 2*(den+zsq) .
COMPUTE x4 = (x1 - x2) / x3 .
COMPUTE x5 = (x1 + x2) / x3 .
COMPUTE vnames={'X','N','Percent','Lower','Upper'}.
SAVE {num,den,100*p,100*x4,100*x5} /OUTFILE=Results /NAMES=vnames.
END MATRIX.
RESTORE.
DATASET ACTIVATE Results.
FORMAT X N(F8) Percent TO Upper (PCT4.2).
OMS /SELECT TABLES
  /IF COMMANDS='Summarize' SUBTYPES='Case Processing Summary'
  /DESTINATION VIEWER=NO.
SUMMARIZE
  /TABLES=ALL
  /FORMAT=LIST NOCASENUM NOTOTAL
  /TITLE="Confidence interval for one proportion (Wilson's method) (*)"
  /CELLS=NONE.
OMSEND.
DATASET ACTIVATE DatA.
DATASET CLOSE Results.
ECHO !QUOTE(!CONCAT('(*) Confidence level: ',!3,'%')).
!ENDDEFINE.

IC1PROP 16 48 95.
IC1PROP 16 48 99.

=====================
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]

Hi Marta.  Always good to see you make an appearance on this list.  Re your modification to accommodate both 95% and 99% CIs, I was forgetting that MATRIX doesn't have inverse density functions, which makes things more difficult than they should be.   Here is some straight syntax I wrote a while ago, and cleaned up a bit just now.  Because it uses straight syntax, users can have whatever confidence level they want.

Cheers,
Bruce

*  ============================================================= .
*  File:   CI_for_proportion.SPS .
*  Date:   19-Nov-2012 .
*  Author:  Bruce Weaver, [e-mail] .
*  ============================================================= .

* Get confidence interval for a binomial proportion using:
   - Wald method
   - Adjusted Wald method (Agresti & Coull, 1998)
   - Wilson score method (identical to Ghosh's 1979 method)
   - Jeffreys method
.
* The data used here are from Table I in Newcombe (1998), Statistics
   in Medicine, Vol 17, 857-872.

DATA LIST LIST /x(f8.0) n(f8.0) confid(f5.3) .
BEGIN DATA.
81 263 .95
15 148 .95
0   20 .95
1   29 .95
81 263 .90
15 148 .90
0   20 .90
1   29 .90
81 263 .99
15 148 .99
0   20 .99
1   29 .99
16  48 .95
16  48 .99
END DATA.

compute alpha = 1 - confid.
compute p = x/n.
compute q = 1-p.
compute z = probit(1-alpha/2).

*  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .

* Wald method (i.e., the usual normal approximation).

compute #se = SQRT(p*q/n).
compute Lower1 = p - z*#se.
if Lower1 LT 0 Lower1 = 0.
compute Upper1 = p + z*#se.
if Upper1 GT 1 Upper1 = 1.

*  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .

* Adjusted Wald method due to Agresti & Coull (1998).

compute #p = (x + z**2/2) / (n + z**2).
compute #q = 1 - #p.
compute #se = SQRT(#p*#q/(n+z**2)).
compute Lower2 = #p - z*#se.
if Lower2 LT 0 Lower2 = 0.
compute Upper2 = #p + z*#se.
if Upper2 GT 1 Upper2 = 1.

*  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .

* Wilson score method (Method 3 in Newcombe, 1998) .
* Code adapted from Robert Newcombe's code posted here:
    http://archive.uwcm.ac.uk/uwcm/ms/Robert2.html .

* The method of Ghosh (1979), as described in Glass & Hopkins
* (1996, p 326) is identical to Wilson's method.
* Glass & Hopkins describe it as the "method of choice for all values
   of p and n" .

COMPUTE #x1 = 2*n*p+z**2 .
COMPUTE #x2 = z*(z**2+4*n*p*(1-p))**0.5 .
COMPUTE #x3 = 2*(n+z**2) .
COMPUTE Lower3 = (#x1 - #x2) / #x3 .
COMPUTE Upper3 = (#x1 + #x2) / #x3 .

*  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .

* Jeffreys method shown on the IBM-SPSS website at
* http://www-01.ibm.com/support/docview.wss?uid=swg21474963 .

compute Lower4 = idf.beta(alpha/2,x+.5,n-x+.5).
compute Upper4 = idf.beta(1-alpha/2,x+.5,n-x+.5).

*  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .

* Format variables and list the results of all methods .

formats p q Lower1 to Upper4 (f5.4).
sort cases by p confid.

list var x n confid p Lower1 to Upper4 .

* Method 1:  Wald method (i.e., the usual normal approximation) .
* Method 2:  Adjusted Wald method (using z**2/2 and z**2 rather than 2 and 4).
* Method 3:  Wilson score method (from Newcombe paper), identical to Ghosh (1979).
* Method 4:  Jeffreys method (http://www-01.ibm.com/support/docview.wss?uid=swg21474963).

* Data from Newcombe (1998), Table I.

variable labels
 x "Successes"
 n "Trials"
 p "p(Success)"
 confid "Confidence Level"
 Lower1 "Wald: Lower"
 Upper1 "Wald: Upper"
 Lower2 "Adj Wald: Lower"
 Upper2 "Adj Wald: Upper"
 Lower3 "Wilson score/Ghosh: Lower"
 Upper3 "Wilson score/Ghosh: Upper"
 Lower4 "Jeffreys: Lower"
 Upper4 "Jeffreys: Upper"
.

SUMMARIZE
  /TABLES=x n p confid Lower1 Upper1 Lower2 Upper2 Lower3 Upper3 Lower4 Upper4
  /FORMAT=VALIDLIST NOCASENUM TOTAL
  /TITLE='Confidence Intervals for Binomial Proportions'
  /MISSING=VARIABLE
  /CELLS=NONE.

*  ============================================================= .


OUTPUT from the LIST command above:
 
       x        n confid     p Lower1 Upper1 Lower2 Upper2 Lower3 Upper3 Lower4 Upper4
 
       0       20   .900 .0000  .0000  .0000  .0000  .1413  .0000  .1192  .0001  .0905
       0       20   .950 .0000  .0000  .0000  .0000  .1898  .0000  .1611  .0000  .1166
       0       20   .990 .0000  .0000  .0000  .0000  .2894  .0000  .2491  .0000  .1768
       1       29   .900 .0345  .0000  .0902  .0000  .1508  .0077  .1407  .0061  .1271
       1       29   .950 .0345  .0000  .1009  .0000  .1863  .0061  .1718  .0037  .1501
       1       29   .990 .0345  .0000  .1218  .0000  .2620  .0041  .2383  .0012  .2001
      15      148   .900 .1014  .0605  .1422  .0668  .1502  .0674  .1496  .0663  .1479
      15      148   .950 .1014  .0527  .1500  .0614  .1615  .0624  .1605  .0604  .1576
      15      148   .990 .1014  .0375  .1653  .0515  .1854  .0536  .1833  .0501  .1777
      81      263   .900 .3080  .2612  .3548  .2633  .3566  .2633  .3566  .2629  .3562
      81      263   .950 .3080  .2522  .3638  .2552  .3663  .2553  .3662  .2545  .3656
      81      263   .990 .3080  .2347  .3813  .2400  .3854  .2401  .3853  .2386  .3843
      16       48   .950 .3333  .2000  .4667  .2162  .4751  .2168  .4746  .2127  .4733
      16       48   .990 .3333  .1581  .5086  .1870  .5202  .1881  .5191  .1800  .5180

Number of cases read:  14    Number of cases listed:  14

* Method 1:  Wald method (i.e., the usual normal approximation) .
* Method 2:  Adjusted Wald method (using z**2/2 and z**2 rather than 2 and 4).
* Method 3:  Wilson score method (from Newcombe paper), identical to Ghosh (1979).
* Method 4:  Jeffreys method (http://www-01.ibm.com/support/docview.wss?uid=swg21474963).
 
* Data from Newcombe (1998), Table I.

But note that the last two rows are from Marta's two macro calls below.

Marta García-Granero-2 wrote

Hi Bruce:

Long time no see. Knee deep in Stata (all my class notes need BIG
modifications to switch from SPSS to Stata). I thought I might indulge
in a relapse into old adictions ;)

See below the adaptation of Wilson's method to include CL (95 or 99).
Just for fun... big sight. I turned it to DE (*) some time ago, just in
case anyone wants it too.

(*) Deprecated Excel ;)

------------------------------------------------------------
BTW, if you like confidence intervals as measures of effect size, I
recommend you Robert Newcombe's last book (a bit heavy on maths, I kind
of like that, but I agree it could be too much for non-teachers/non-math
addicts): "Confidence intervals for proportions and related measures of
effect size", CRC Press. ISBN 978-1-4398-1278-5. I convinced the
University library to buy one, an it is all mine (until January 4th,
when I'm supposed to return it and let other teachers get a glimpse at
it). His effect size measures for Mann-Whitney and Wilcoxon tests are great!

Best regards,
Marta GG
Stata-rookie

DEFINE IC1PROP(!POSITIONAL !TOKENS(1) /!POSITIONAL
!TOKENS(1)/!POSITIONAL !DEFAULT(95) !TOKENS(1)).
DATASET NAME Data.
DATASET DECLARE Results WINDOW=HIDDEN.
PRESERVE.
SET ERRORS=NONE RESULTS=NONE.
MATRIX.
DO IF (!3 EQ 95) OR (!3 EQ .95). *Just in case people use the fraction *.
- COMPUTE z = 1.959964.
ELSE IF (!3 EQ 99) OR (!3 EQ .99).
- COMPUTE z = 2.575829.
END IF.
COMPUTE num= !1.
COMPUTE den= !2.
COMPUTE p=num/den .
COMPUTE zsq = z*z.
COMPUTE x1 = 2*num+zsq .
COMPUTE x2 = z*(zsq+4*num&*(1-p))&**0.5 .
COMPUTE x3 = 2*(den+zsq) .
COMPUTE x4 = (x1 - x2) / x3 .
COMPUTE x5 = (x1 + x2) / x3 .
COMPUTE vnames={'X','N','Percent','Lower','Upper'}.
SAVE {num,den,100*p,100*x4,100*x5} /OUTFILE=Results /NAMES=vnames.
END MATRIX.
RESTORE.
DATASET ACTIVATE Results.
FORMAT X N(F8) Percent TO Upper (PCT4.2).
OMS /SELECT TABLES
  /IF COMMANDS='Summarize' SUBTYPES='Case Processing Summary'
  /DESTINATION VIEWER=NO.
SUMMARIZE
  /TABLES=ALL
  /FORMAT=LIST NOCASENUM NOTOTAL
  /TITLE="Confidence interval for one proportion (Wilson's method) (*)"
  /CELLS=NONE.
OMSEND.
DATASET ACTIVATE DatA.
DATASET CLOSE Results.
ECHO !QUOTE(!CONCAT('(*) Confidence level: ',!3,'%')).
!ENDDEFINE.

IC1PROP 16 48 95.
IC1PROP 16 48 99.

=====================
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

[Jon K Peck]

There is also an extension command named PROPOR available from the SPSS Community site (www.ibm.com/developerworks/spssdevcentral).  It provides Jeffreys Binomial and Poisson CIs.

The simplest usage would be
PROPOR NUM=16 DENOM=48.

It supports any alpha.  On this example, Wilson gives (.1881, .5191) using Marta's code (scaled by 100) while Jeffreys gives (.213, .473) for alpha=.05


The summary paragraph from
CONFIDENCE INTERVALS FOR A BINOMIAL PROPORTION AND ASYMPTOTIC EXPANSIONS BY LAWRENCE D. BROWN, T. TONY CAI AND ANIRBAN DASGUPTA, Annals of Statistics, 2002, says this.

To summarize, the conclusion is that the Agresti–Coull interval dominates
the other intervals in coverage, but is also longer on an average and is quite
conservative for p near 0 or 1. The Wilson, the likelihood ratio and the Jeffreys
prior interval are comparable in both coverage and length, although the Jeffreys
interval is a bit shorter on average. If we also take simplicity of presentation
and ease of computation into account, the Agresti–Coull interval, although a bit
too long, could be recommended for use in this problem. If simplicity is not a
paramount issue, either the Wilson, the likelihood ratio, or the Jeffreys interval
may be used, depending on taste.


Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: [phone]




From:        Marta García-Granero <[hidden email]>
To:        [hidden email],
Date:        11/19/2012 07:52 AM
Subject:        Re: [SPSSX-L] Confidence Interval
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

Hi Bruce:

Long time no see. Knee deep in Stata (all my class notes need BIG
modifications to switch from SPSS to Stata). I thought I might indulge
in a relapse into old adictions ;)

See below the adaptation of Wilson's method to include CL (95 or 99).
Just for fun... big sight. I turned it to DE (*) some time ago, just in
case anyone wants it too.

(*) Deprecated Excel ;)

------------------------------------------------------------
BTW, if you like confidence intervals as measures of effect size, I
recommend you Robert Newcombe's last book (a bit heavy on maths, I kind
of like that, but I agree it could be too much for non-teachers/non-math
addicts): "Confidence intervals for proportions and related measures of
effect size", CRC Press. ISBN 978-1-4398-1278-5. I convinced the
University library to buy one, an it is all mine (until January 4th,
when I'm supposed to return it and let other teachers get a glimpse at
it). His effect size measures for Mann-Whitney and Wilcoxon tests are great!

Best regards,
Marta GG
Stata-rookie

DEFINE IC1PROP(!POSITIONAL !TOKENS(1) /!POSITIONAL
!TOKENS(1)/!POSITIONAL !DEFAULT(95) !TOKENS(1)).
DATASET NAME Data.
DATASET DECLARE Results WINDOW=HIDDEN.
PRESERVE.
SET ERRORS=NONE RESULTS=NONE.
MATRIX.
DO IF (!3 EQ 95) OR (!3 EQ .95). *Just in case people use the fraction *.
- COMPUTE z = 1.959964.
ELSE IF (!3 EQ 99) OR (!3 EQ .99).
- COMPUTE z = 2.575829.
END IF.
COMPUTE num= !1.
COMPUTE den= !2.
COMPUTE p=num/den .
COMPUTE zsq = z*z.
COMPUTE x1 = 2*num+zsq .
COMPUTE x2 = z*(zsq+4*num&*(1-p))&**0.5 .
COMPUTE x3 = 2*(den+zsq) .
COMPUTE x4 = (x1 - x2) / x3 .
COMPUTE x5 = (x1 + x2) / x3 .
COMPUTE vnames={'X','N','Percent','Lower','Upper'}.
SAVE {num,den,100*p,100*x4,100*x5} /OUTFILE=Results /NAMES=vnames.
END MATRIX.
RESTORE.
DATASET ACTIVATE Results.
FORMAT X N(F8) Percent TO Upper (PCT4.2).
OMS /SELECT TABLES
 /IF COMMANDS='Summarize' SUBTYPES='Case Processing Summary'
 /DESTINATION VIEWER=NO.
SUMMARIZE
 /TABLES=ALL
 /FORMAT=LIST NOCASENUM NOTOTAL
 /TITLE="Confidence interval for one proportion (Wilson's method) (*)"
 /CELLS=NONE.
OMSEND.
DATASET ACTIVATE DatA.
DATASET CLOSE Results.
ECHO !QUOTE(!CONCAT('(*) Confidence level: ',!3,'%')).
!ENDDEFINE.

IC1PROP 16 48 95.
IC1PROP 16 48 99.

=====================
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

[Rubier Perez]

Thank you all very much for your contribution, I was really searching
for the possibility of through the SPSS interface to estimate a
confidence interval of 95% for the variance and the proportion.
Thanks

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.



--

Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

[Ryan]

You can use the genlin procedure to construct confidence limits about a proportion.

Ryan

On Nov 21, 2012, at 3:42 PM, Rubier Perez <[hidden email]> wrote:

> Thank you all very much for your contribution, I was really searching
> for the possibility of through the SPSS interface to estimate a
> confidence interval of 95% for the variance and the proportion.
> Thanks
>
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
>
>
>
> --
>
> Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas
>
> Infomed: http://www.sld.cu/
>
> =====================
> 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

=====================
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

[Rubier Perez]

Hi, Ryan. And where is the option in SPSS for the genlin procedure?
Thank you


  Ryan Black <[hidden email]> escribió:

> You can use the genlin procedure to construct confidence limits
> about a proportion.
>
> Ryan
>
> On Nov 21, 2012, at 3:42 PM, Rubier Perez <[hidden email]> wrote:
>
>> Thank you all very much for your contribution, I was really searching
>> for the possibility of through the SPSS interface to estimate a
>> confidence interval of 95% for the variance and the proportion.
>> Thanks
>>
>> ----------------------------------------------------------------
>> This message was sent using IMP, the Internet Messaging Program.
>>
>>
>>
>> --
>>
>> Este mensaje le ha llegado mediante el servicio de correo
>> electronico que ofrece Infomed para respaldar el cumplimiento de
>> las misiones del Sistema Nacional de Salud. La persona que envia
>> este correo asume el compromiso de usar el servicio a tales fines y
>> cumplir con las regulaciones establecidas
>>
>> Infomed: http://www.sld.cu/
>>
>> =====================
>> 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
>
> =====================
> 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
>

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


--

Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

[Ryan]

Rubier,

You have a couple of options using the pull-down menu:

Analyze-->Generalized Linear Models-->Generalized Linear Models
Analyze-->Nonparametric Tests-->One Sample

The nonparametric tests procedure provides several options for
constructing confidence limits about a binomial proportion.

Ryan

On Nov 21, 2012, at 4:36 PM, Rubier Perez <[hidden email]> wrote:

> Hi, Ryan. And where is the option in SPSS for the genlin procedure?
> Thank you
>
>
> Ryan Black <[hidden email]> escribió:
>
>> You can use the genlin procedure to construct confidence limits
>> about a proportion.
>>
>> Ryan
>>
>> On Nov 21, 2012, at 3:42 PM, Rubier Perez <[hidden email]> wrote:
>>
>>> Thank you all very much for your contribution, I was really searching
>>> for the possibility of through the SPSS interface to estimate a
>>> confidence interval of 95% for the variance and the proportion.
>>> Thanks
>>>
>>> ----------------------------------------------------------------
>>> This message was sent using IMP, the Internet Messaging Program.
>>>
>>>
>>>
>>> --
>>>
>>> Este mensaje le ha llegado mediante el servicio de correo
>>> electronico que ofrece Infomed para respaldar el cumplimiento de
>>> las misiones del Sistema Nacional de Salud. La persona que envia
>>> este correo asume el compromiso de usar el servicio a tales fines y
>>> cumplir con las regulaciones establecidas
>>>
>>> Infomed: http://www.sld.cu/
>>>
>>> =====================
>>> 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
>>
>> =====================
>> 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
>
>
>
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
>
>
> --
>
> Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas
>
> Infomed: http://www.sld.cu/
>
> =====================
> 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

=====================
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

[Rubier Perez]

Hi, Ryan. Ok, thank you very much.
Rubier


R B <[hidden email]> escribió:

> Rubier,
>
> You have a couple of options using the pull-down menu:
>
> Analyze-->Generalized Linear Models-->Generalized Linear Models
> Analyze-->Nonparametric Tests-->One Sample
>
> The nonparametric tests procedure provides several options for
> constructing confidence limits about a binomial proportion.
>
> Ryan
>
> On Nov 21, 2012, at 4:36 PM, Rubier Perez <[hidden email]> wrote:
>
>> Hi, Ryan. And where is the option in SPSS for the genlin procedure?
>> Thank you
>>
>>
>> Ryan Black <[hidden email]> escribió:
>>
>>> You can use the genlin procedure to construct confidence limits
>>> about a proportion.
>>>
>>> Ryan
>>>
>>> On Nov 21, 2012, at 3:42 PM, Rubier Perez <[hidden email]> wrote:
>>>
>>>> Thank you all very much for your contribution, I was really searching
>>>> for the possibility of through the SPSS interface to estimate a
>>>> confidence interval of 95% for the variance and the proportion.
>>>> Thanks
>>>>
>>>> ----------------------------------------------------------------
>>>> This message was sent using IMP, the Internet Messaging Program.
>>>>
>>>>
>>>>
>>>> --
>>>>
>>>> Este mensaje le ha llegado mediante el servicio de correo
>>>> electronico que ofrece Infomed para respaldar el cumplimiento de
>>>> las misiones del Sistema Nacional de Salud. La persona que envia
>>>> este correo asume el compromiso de usar el servicio a tales fines y
>>>> cumplir con las regulaciones establecidas
>>>>
>>>> Infomed: http://www.sld.cu/
>>>>
>>>> =====================
>>>> 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
>>>
>>> =====================
>>> 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
>>
>>
>>
>> ----------------------------------------------------------------
>> This message was sent using IMP, the Internet Messaging Program.
>>
>>
>> --
>>
>> Este mensaje le ha llegado mediante el servicio de correo
>> electronico que ofrece Infomed para respaldar el cumplimiento de
>> las misiones del Sistema Nacional de Salud. La persona que envia
>> este correo asume el compromiso de usar el servicio a tales fines y
>> cumplir con las regulaciones establecidas
>>
>> Infomed: http://www.sld.cu/
>>
>> =====================
>> 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
>
> =====================
> 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
>

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


--

Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

[Ryan]

Hi Rubier,

 

Below is SPSS syntax that generates binary (0/1) data (n=100), and then constructs 95% confidence intervals using the nonparametric procedure. I wrote the syntax to generate the data, but then used the pull-down menu to obtain the syntax for the nonparametric procedure via the PASTE button. Using the nonparametric procedure on the simulated data, you will obtain the following 95% confidence intervals:

 

Clopper-Pearson--> 95% CI: [.675, .848]

Jeffreys--> 95% CI: [.681, .844]

Likelihood--> 95% CI: [.681, .845]

 

Hope this illustration helps.

 

Ryan

--

 

*Beginning of Data Generation Code.
 set seed 98765432.
new file.
 inp pro.
loop #obs= 1 to 100.
  comp x = rv.bernoulli(.70).
   end case.
  end loop.
 end file.
 end inp pro.
 EXECUTE.
*End of Data Generation Code.

 

*Use NPTESTS to obtain confidence limits about a single proportion using Clopper-Pearson, Jeffreys, and Likelihood.
NPTESTS
  /ONESAMPLE TEST (x) BINOMIAL(TESTVALUE=0.5 CLOPPERPEARSON JEFFREYS LIKELIHOOD 
    SUCCESSCATEGORICAL=LIST(1) SUCCESSCONTINUOUS=CUTPOINT(MIDPOINT))
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05 CILEVEL=95.

On Thu, Nov 22, 2012 at 10:44 AM, Rubier Perez <[hidden email]> wrote:

Hi, Ryan. Ok, thank you very much.
Rubier


R B <[hidden email]> escribió:

Rubier,

You have a couple of options using the pull-down menu:

Analyze-->Generalized Linear Models-->Generalized Linear Models
Analyze-->Nonparametric Tests-->One Sample

The nonparametric tests procedure provides several options for
constructing confidence limits about a binomial proportion.

Ryan

On Nov 21, 2012, at 4:36 PM, Rubier Perez <[hidden email]> wrote:

Hi, Ryan. And where is the option in SPSS for the genlin procedure?
Thank you


Ryan Black <[hidden email]> escribió:

You can use the genlin procedure to construct confidence limits
about a proportion.

Ryan

On Nov 21, 2012, at 3:42 PM, Rubier Perez <[hidden email]> wrote:

Thank you all very much for your contribution, I was really searching
for the possibility of through the SPSS interface to estimate a
confidence interval of 95% for the variance and the proportion.
Thanks

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.



--

Este mensaje le ha llegado mediante el servicio de correo
electronico que ofrece Infomed para respaldar el cumplimiento de
las misiones del Sistema Nacional de Salud. La persona que envia
este correo asume el compromiso de usar el servicio a tales fines y
cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

=====================
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

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


--

Este mensaje le ha llegado mediante el servicio de correo
electronico que ofrece Infomed para respaldar el cumplimiento de
las misiones del Sistema Nacional de Salud. La persona que envia
este correo asume el compromiso de usar el servicio a tales fines y
cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

=====================
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

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


--

Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

[Ryan]

One more point--After seeing that the nonparametric procedure produces Likelihood confidence intervals, I tried to replicate the results using GENLIN. As expected, the results were the same [to at least the 3rd decimal place]. See the code below my name. (Note that the values inserted into the COMPUTE statement were obtained from the output).

 

Ryan

--

 

* Generalized Linear Models.
GENLIN x (REFERENCE=FIRST)
  /MODEL INTERCEPT=YES
 DISTRIBUTION=BINOMIAL LINK=LOGIT
  /CRITERIA METHOD=FISHER(1) SCALE=1 COVB=MODEL MAXITERATIONS=100 MAXSTEPHALVING=5
    PCONVERGE=1E-006(ABSOLUTE) SINGULAR=1E-012 ANALYSISTYPE=3(WALD) CILEVEL=95 CITYPE=PROFILE(.0001)
    LIKELIHOOD=FULL
  /MISSING CLASSMISSING=EXCLUDE
  /PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION.

 

COMPUTE Lower_Limit=exp(.7598884094034025) / (1 + exp(.7598884094034025)).
COMPUTE Upper_Limit=exp(1.6959805306633795) / (1 + exp(1.6959805306633795)).
EXECUTE.

On Sun, Nov 25, 2012 at 7:00 PM, R B <[hidden email]> wrote:

Hi Rubier,

 

Below is SPSS syntax that generates binary (0/1) data (n=100), and then constructs 95% confidence intervals using the nonparametric procedure. I wrote the syntax to generate the data, but then used the pull-down menu to obtain the syntax for the nonparametric procedure via the PASTE button. Using the nonparametric procedure on the simulated data, you will obtain the following 95% confidence intervals:

 

Clopper-Pearson--> 95% CI: [.675, .848]

Jeffreys--> 95% CI: [.681, .844]

Likelihood--> 95% CI: [.681, .845]

 

Hope this illustration helps.

 

Ryan

--

 

*Beginning of Data Generation Code.
 set seed 98765432.
new file.
 inp pro.
loop #obs= 1 to 100.
  comp x = rv.bernoulli(.70).
   end case.
  end loop.
 end file.
 end inp pro.
 EXECUTE.
*End of Data Generation Code.

 

*Use NPTESTS to obtain confidence limits about a single proportion using Clopper-Pearson, Jeffreys, and Likelihood.
NPTESTS
  /ONESAMPLE TEST (x) BINOMIAL(TESTVALUE=0.5 CLOPPERPEARSON JEFFREYS LIKELIHOOD 
    SUCCESSCATEGORICAL=LIST(1) SUCCESSCONTINUOUS=CUTPOINT(MIDPOINT))
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05 CILEVEL=95.

On Thu, Nov 22, 2012 at 10:44 AM, Rubier Perez <[hidden email]> wrote:

Hi, Ryan. Ok, thank you very much.
Rubier


R B <[hidden email]> escribió:

Rubier,

You have a couple of options using the pull-down menu:

Analyze-->Generalized Linear Models-->Generalized Linear Models
Analyze-->Nonparametric Tests-->One Sample

The nonparametric tests procedure provides several options for
constructing confidence limits about a binomial proportion.

Ryan

On Nov 21, 2012, at 4:36 PM, Rubier Perez <[hidden email]> wrote:

Hi, Ryan. And where is the option in SPSS for the genlin procedure?
Thank you


Ryan Black <[hidden email]> escribió:

You can use the genlin procedure to construct confidence limits
about a proportion.

Ryan

On Nov 21, 2012, at 3:42 PM, Rubier Perez <[hidden email]> wrote:

Thank you all very much for your contribution, I was really searching
for the possibility of through the SPSS interface to estimate a
confidence interval of 95% for the variance and the proportion.
Thanks

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.



--

Este mensaje le ha llegado mediante el servicio de correo
electronico que ofrece Infomed para respaldar el cumplimiento de
las misiones del Sistema Nacional de Salud. La persona que envia
este correo asume el compromiso de usar el servicio a tales fines y
cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

=====================
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

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


--

Este mensaje le ha llegado mediante el servicio de correo
electronico que ofrece Infomed para respaldar el cumplimiento de
las misiones del Sistema Nacional de Salud. La persona que envia
este correo asume el compromiso de usar el servicio a tales fines y
cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

=====================
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

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


--

Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

[Art Kendall]

I cobbled together a data generation I had with Ryan's syntax.
I intended to use a few Ns and OMS to grab the results and graph them but It is now time for some clients.  If I have a chance I'll add the OMS and graph in the next few days.
Meanwhile here is something to eyeball to help develop a feel for the comparison.

*compare methods confidence intervals for proportions.
set seed 20121125.
new file.
input program.
   loop ID= 1 to 100.
      vector x (100,f1).
      loop #i = 1 to 100 by 1.
         compute proportion= #i/100.
         compute x(#i) = rv.bernoulli(proportion).
      end loop.
      end case.
   end loop.
   end file.
end input program.
EXECUTE.
*End of Data Generation Code.

*Use NPTESTS to obtain confidence limits about a single proportion
   using Clopper-Pearson, Jeffreys, and Likelihood.
* Surprise this does not work    /ONESAMPLE TEST (x1 to x100) .
NPTESTS
   /ONESAMPLE TEST(
  x1  x2  x3  x4  x5  x6  x7  x8  x9 x10 x11 x12 x13 x14 x15 x16 x17 x18 x19 x20
 x21 x22 x23 x24 x25 x26 x27 x28 x29 x30 x31 x32 x33 x34 x35 x36 x37 x38 x39 x40
 x41 x42 x43 x44 x45 x46 x47 x48 x49 x50 x51 x52 x53 x54 x55 x56 x57 x58 x59 x60
 x61 x62 x63 x64 x65 x66 x67 x68 x69 x70 x71 x72 x73 x74 x75 x76 x77 x78 x79 x80
 x81 x82 x83 x84 x85 x86 x87 x88 x89 x90 x91 x92 x93 x94 x95 x96 x97 x98 x99 x100)
 BINOMIAL(TESTVALUE=0.5
      CLOPPERPEARSON JEFFREYS LIKELIHOOD
      SUCCESSCATEGORICAL=LIST(1) SUCCESSCONTINUOUS=CUTPOINT(MIDPOINT))
   /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
   /CRITERIA ALPHA=0.05 CILEVEL=95.
On 11/26/2012 7:27 AM, R B wrote:

One more point--After seeing that the nonparametric procedure produces Likelihood confidence intervals, I tried to replicate the results using GENLIN. As expected, the results were the same [to at least the 3rd decimal place]. See the code below my name. (Note that the values inserted into the COMPUTE statement were obtained from the output).

 

Ryan

--

 

* Generalized Linear Models.
GENLIN x (REFERENCE=FIRST)
  /MODEL INTERCEPT=YES
 DISTRIBUTION=BINOMIAL LINK=LOGIT
  /CRITERIA METHOD=FISHER(1) SCALE=1 COVB=MODEL MAXITERATIONS=100 MAXSTEPHALVING=5
    PCONVERGE=1E-006(ABSOLUTE) SINGULAR=1E-012 ANALYSISTYPE=3(WALD) CILEVEL=95 CITYPE=PROFILE(.0001)
    LIKELIHOOD=FULL
  /MISSING CLASSMISSING=EXCLUDE
  /PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION.

 

COMPUTE Lower_Limit=exp(.7598884094034025) / (1 + exp(.7598884094034025)).
COMPUTE Upper_Limit=exp(1.6959805306633795) / (1 + exp(1.6959805306633795)).
EXECUTE.

On Sun, Nov 25, 2012 at 7:00 PM, R B <[hidden email]> wrote:

Hi Rubier,

 

Below is SPSS syntax that generates binary (0/1) data (n=100), and then constructs 95% confidence intervals using the nonparametric procedure. I wrote the syntax to generate the data, but then used the pull-down menu to obtain the syntax for the nonparametric procedure via the PASTE button. Using the nonparametric procedure on the simulated data, you will obtain the following 95% confidence intervals:

 

Clopper-Pearson--> 95% CI: [.675, .848]

Jeffreys--> 95% CI: [.681, .844]

Likelihood--> 95% CI: [.681, .845]

 

Hope this illustration helps.

 

Ryan

--

 

*Beginning of Data Generation Code.
 set seed 98765432.
new file.
 inp pro.
loop #obs= 1 to 100.
  comp x = rv.bernoulli(.70).
   end case.
  end loop.
 end file.
 end inp pro.
 EXECUTE.
*End of Data Generation Code.

 

*Use NPTESTS to obtain confidence limits about a single proportion using Clopper-Pearson, Jeffreys, and Likelihood.
NPTESTS
  /ONESAMPLE TEST (x) BINOMIAL(TESTVALUE=0.5 CLOPPERPEARSON JEFFREYS LIKELIHOOD 
    SUCCESSCATEGORICAL=LIST(1) SUCCESSCONTINUOUS=CUTPOINT(MIDPOINT))
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05 CILEVEL=95.

On Thu, Nov 22, 2012 at 10:44 AM, Rubier Perez <[hidden email]> wrote:

Hi, Ryan. Ok, thank you very much.
Rubier


R B <[hidden email]> escribió:

Rubier,

You have a couple of options using the pull-down menu:

Analyze-->Generalized Linear Models-->Generalized Linear Models
Analyze-->Nonparametric Tests-->One Sample

The nonparametric tests procedure provides several options for
constructing confidence limits about a binomial proportion.

Ryan

On Nov 21, 2012, at 4:36 PM, Rubier Perez <[hidden email]> wrote:

Hi, Ryan. And where is the option in SPSS for the genlin procedure?
Thank you


Ryan Black <[hidden email]> escribió:

You can use the genlin procedure to construct confidence limits
about a proportion.

Ryan

On Nov 21, 2012, at 3:42 PM, Rubier Perez <[hidden email]> wrote:

Thank you all very much for your contribution, I was really searching
for the possibility of through the SPSS interface to estimate a
confidence interval of 95% for the variance and the proportion.
Thanks

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.



--

Este mensaje le ha llegado mediante el servicio de correo
electronico que ofrece Infomed para respaldar el cumplimiento de
las misiones del Sistema Nacional de Salud. La persona que envia
este correo asume el compromiso de usar el servicio a tales fines y
cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

=====================
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

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


--

Este mensaje le ha llegado mediante el servicio de correo
electronico que ofrece Infomed para respaldar el cumplimiento de
las misiones del Sistema Nacional de Salud. La persona que envia
este correo asume el compromiso de usar el servicio a tales fines y
cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

=====================
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

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


--

Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas

Infomed: http://www.sld.cu/

=====================
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

===================== 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