SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution
Locked
 
|
I am pleased to inform that a new program Confidence Interval
for Cohen's d Based on the Noncentral t-Distribution has been added to SPSS Syntax Files http://gaius.fpce.uc.pt/niips/spss_prc/index.htm#meta Valentim R. Alferes |
Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution
Administrator
|
Valentim,
Do you care to elaborate on what/why is going on with this beginning of the code? Sometimes comments in code can be useful! Why all those EXECUTE statements? I'm surprised it works at all without an exe after the end loop prior to the GET FILE. other than that it looks like you are generating 2 million cases of constant data? Later sorting it twice. Seriously? Why? Maybe I'm missing something? Surely there are more direct/efficient ways to achieve this? OTOH, since we don't know WTF the reason why it was written this way/intention it is difficult to make any cogent proposals. As currently coded I am left feeling compelled to extract my eyeballs with a sharpened spoon! David begin data. 0.84 20 15 end data. * Enter confidence level (by default, CL=95%). compute CL=95. execute. loop NCP=0 to 2*10**6. xsave outfile='%temp%\file1'. end loop. get file='%temp%\file1'. compute df=N1+N2-2. compute tobs=d/sqr(1/N1+1/N2). compute NCP=NCP-10**6. compute NCP=NCP/10**4. compute p1=((100-CL)/100)/2. compute p2=1-p1. compute prob=ncdf.t(tobs,df,NCP). compute dif1=abs(p1-prob). execute. sort cases by dif1(a). if ($casenum=1) NCPupper=NCP. execute. compute dif2=abs(p2-prob). execute. sort cases by dif2(a). if ($casenum=1) NCPlower=NCP. execute. compute w=1. execute. ........ more bizarre stuff.... ----
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
In reply to this post by valferes
David, Given that SPSS has
not the “Inverse Noncentral t Cumulative Distribution Function”, there is only one way to obtain the accurate
value of a noncentrality
parameter (NCP) for a given probability:
to use the “Noncentral t Cumulative Distribution
Function” and iteratively search for the NCP.
This is what the code does! The number of cases
generated is a
simple matter of precision (with
the current numbers, we obtain accurate
NCPs up to the fourth decimal place).
The two “sort commands” are needed because there are two NCP: for p = .025 and p = .975. I hope that SPSS 23 will include the “Inverse Noncentral t cdf”, as is the case of MATLAB (see http://www.mathworks.com/help/stats/nctinv.html) Best regards, Valentim |
Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution
|
|
I needed an inverse, back in the days when computers were more finite, so
=====================
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
I found the other way. The other way to get the inverse from the cumulative distribution is to interpolate (using transformed scales that produce approximate linearity) and iterate. Two iterations should give you the 4th decimal. -- Rich Ulrich Date: Mon, 28 Jul 2014 19:01:28 +0000 From: [hidden email] Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution To: [hidden email] David, Given that SPSS has not the “Inverse Noncentral t Cumulative Distribution Function”, there is only one way to obtain the accurate value of a noncentrality parameter (NCP) for a given probability: to use the “Noncentral t Cumulative Distribution Function” and iteratively search for the NCP. This is what the code does! The number of cases generated is a simple matter of precision (with the current numbers, we obtain accurate NCPs up to the fourth decimal place). The two “sort commands” are needed because there are two NCP: for p = .025 and p = .975. I hope that SPSS 23 will include the “Inverse Noncentral t cdf”, as is the case of MATLAB (see http://www.mathworks.com/help/stats/nctinv.html) ... |
|
|
Dear Rich Ulrich,
=====================
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
I would be very grateful if you could show me your solution to the same problem: * My data: t observed and degrees of freedom. data list list/tobs(f12.10) df(f8.0). begin data. 2.4592681838 33 end data. * Problem: Find the noncentrality parameters for p1=.025 and p2=.975. My code is here and the illustrative output is here. Sincerely, Valentim R. Alferes http://gaius.fpce.uc.pt/niips/ Date: Tue, 29 Jul 2014 00:45:37 -0400 From: [hidden email] Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution To: [hidden email] I needed an inverse, back in the days when computers were more finite, so ===================== 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 REFCARDI found the other way. The other way to get the inverse from the cumulative distribution is to interpolate (using transformed scales that produce approximate linearity) and iterate. Two iterations should give you the 4th decimal. -- Rich Ulrich Date: Mon, 28 Jul 2014 19:01:28 +0000 From: [hidden email] Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution To: [hidden email] David, Given that SPSS has not the “Inverse Noncentral t Cumulative Distribution Function”, there is only one way to obtain the accurate value of a noncentrality parameter (NCP) for a given probability: to use the “Noncentral t Cumulative Distribution Function” and iteratively search for the NCP. This is what the code does! The number of cases generated is a simple matter of precision (with the current numbers, we obtain accurate NCPs up to the fourth decimal place). The two “sort commands” are needed because there are two NCP: for p = .025 and p = .975. I hope that SPSS 23 will include the “Inverse Noncentral t cdf”, as is the case of MATLAB (see http://www.mathworks.com/help/stats/nctinv.html) ... |
Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution
|
Administrator
|
Why not build some loops rather than going at it with a sledge hammer by building 2 million cases?
Also get rid of the EXECUTE statements! --
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
David and Rich,
=====================
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
The code is entirely correct and provides reliable results (compare with ESCI
or MOTE outputs).
Moreover, it is done so that any novice user can get
the CI via a simplified input: d, N1
and N2.
Sincerely, Valentim
http://gaius.fpce.uc.pt/niips/
> Date: Tue, 29 Jul 2014 05:46:56 -0700 > From: [hidden email] > Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution > To: [hidden email] > > Why not build some loops rather than going at it with a sledge hammer by > building 2 million cases? > Also get rid of the EXECUTE statements! > -- > > valferes wrote > > Dear Rich Ulrich,I would be very grateful if you could show me your > > solution to the same problem: > > > > * My data: t observed and degrees of freedom. > > data list list/tobs(f12.10) df(f8.0). > > begin data. > > 2.4592681838 33 > > end data. > > > > * Problem: Find the noncentrality parameters for p1=.025 and p2=.975. > > My code is here and the illustrative output is here. > > Sincerely, > > Valentim R. Alferes > > http://gaius.fpce.uc.pt/niips/ > > Date: Tue, 29 Jul 2014 00:45:37 -0400 > > From: > > > rich-ulrich@ > > > Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on the > > Noncentral t-Distribution > > To: > > > SPSSX-L@.UGA > > > > > > > > > I needed an inverse, back in the days when computers were more finite, so > > I found the other way. > > > > The other way to get the inverse from the cumulative distribution is to > > interpolate (using transformed scales that produce approximate linearity) > > and iterate. Two iterations should give you the 4th decimal. > > > > -- > > Rich Ulrich > > > > Date: Mon, 28 Jul 2014 19:01:28 +0000 > > From: > > > valferes@ > > > Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on the > > Noncentral t-Distribution > > To: > > > SPSSX-L@.UGA > > > > > > > > > David,Given that SPSS has not the “Inverse Noncentral t Cumulative > > Distribution Function”, there is only one way to obtain the accurate value > > of a noncentrality parameter (NCP) for a given probability: to use the > > “Noncentral t Cumulative Distribution Function” and iteratively search for > > the NCP. This is what the code does!The number of cases generated is a > > simple matter of precision (with the current numbers, we obtain accurate > > NCPs up to the fourth decimal place). The two “sort commands” are needed > > because there are two NCP: for p = .025 and p = .975.I hope that SPSS 23 > > will include the “Inverse Noncentral t cdf”, as is the case of MATLAB (see > > http://www.mathworks.com/help/stats/nctinv.html) ...===================== > > To manage your subscription to SPSSX-L, send a message to > > > 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 > > > > ===================== > > To manage your subscription to SPSSX-L, send a message to > > > 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 > > > > > > ----- > Please reply to the list and not to my personal email. > Those desiring my consulting or training services please feel free to email me. > --- > "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." > Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" > -- > View this message in context: http://spssx-discussion.1045642.n5.nabble.com/SPSS-program-Confidence-Interval-for-Cohen-s-d-Based-on-the-Noncentral-t-Distribution-tp5726831p5726844.html > Sent from the SPSSX Discussion mailing list archive at Nabble.com. > > ===================== > 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 > > > > |
Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution
|
Administrator
|
40 seconds?
This runs in about 1/10th of a second and gives as much accuracy as could be desired! ----- DEFINE NCPMacro (d !TOKENS(1) / N1 !TOKENS(1) /N2 !TOKENS(1) /CL !TOKENS(1) !DEFAULT (95) ). INPUT PROGRAM. LOOP #=1 TO 1. END CASE. END LOOP. END FILE. END INPUT PROGRAM. COMPUTE CL=!CL. COMPUTE df=!N1+!N2-2. COMPUTE tobs=!d/sqr(1/!N1+1/!N2). COMPUTE p1=1-((100-!CL)/100)/2. SET MXLOOPS=100. COMPUTE epsilon=.0000001. VECTOR NCPs(2). LOOP #=1 TO 2. + COMPUTE minabsdif=1. + COMPUTE bottomNCP=-10. + COMPUTE topNCP=10. + COMPUTE newdelta=1. + COMPUTE incr=1. + IF #=1 P=P1. + IF #=2 P=1-P1. + LOOP. + LOOP NCPvalue= bottomNCP TO topNCP BY incr. + COMPUTE delta=newdelta. + COMPUTE prob=ncdf.t(tobs,df,NCPvalue). + COMPUTE newdelta=abs(p-prob). + DO IF newdelta LT minabsdif. + COMPUTE minabsdif=newdelta. + COMPUTE minNCP=NCPvalue. + END IF. + END LOOP. + COMPUTE bottomNCP=minNCP-incr. + COMPUTE topNCP=minNCP+incr. + COMPUTE incr=incr/10. + END LOOP IF delta LE epsilon. + COMPUTE NCPs(#)=minNCP. PRINT minNCP (F10.6). END LOOP. PRINT /NCPs1 NCPs2 (2F12.8). EXECUTE. !ENDDEFINE. NCPMacro d=0.84 N1=20 N2=15 . * result .39506810 4.48934000
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution
Administrator
|
David, I'm getting the following error messages when I run that:
>Error # 4700 in column 7. Text: minNCP >An output command contains an unrecognized keyword. The recognized keywords >are OUTFILE, RECORDS, TABLE, and NOTABLE. >Execution of this command stops. >Error # 4702 in column 14. Text: ( >An unexpected symbol was encountered on an output command. >Error # 4700 in column 15. Text: F10.6 >An output command contains an unrecognized keyword. The recognized keywords >are OUTFILE, RECORDS, TABLE, and NOTABLE. >Error # 4702 in column 20. Text: ) >An unexpected symbol was encountered on an output command. .39506810 4.48934000 Adding a slash to your first PRINT command fixes it. I.e., changing this... PRINT minNCP (F10.6). to this... PRINT /minNCP (F10.6). With that change, I get the following error-free output: NCPMacro d=0.84 N1=20 N2=15 . .395068 4.489340 .39506810 4.48934000 My version info: v 21.0.0.2 for Windows, 64-bit, Windows 7 Professional, SP1.
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution
|
Administrator
|
I'd say get rid of that line altogether.
I'm not sure how I missed that (I must have accidently dislodged the / at one point). Anyhow, it could probably be made a bit more efficient (no need probably to search the entire lo to hi range once one finds a minimum, but I got lazy). --
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
David,
=====================
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
Thank you very much for your Macro. I also made
some changes in Syntax
and now it runs about
4 times faster: http://gaius.fpce.uc.pt/niips/spss_prc/meta/esCIdNCT.sps If I further reduce the range of the possible NCP values, the Syntax is even faster: http://gaius.fpce.uc.pt/niips/spss_prc/meta/esCIdNCTalt.sps My initial goal was to condense the
calculations into three or four lines of code in order to insert them in a broader program that
includes different measures
of effect size. For
this, I still have to wait for
the inverse non-central t distribution in SPSS 23.
Valentim http://gaius.fpce.uc.pt/niips/ > Date: Tue, 29 Jul 2014 13:23:52 -0700 > From: [hidden email] > Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution > To: [hidden email] > > I'd say get rid of that line altogether. > I'm not sure how I missed that (I must have accidently dislodged the / at > one point). > Anyhow, it could probably be made a bit more efficient (no need probably to > search the entire lo to hi range once one finds a minimum, but I got lazy). > -- > > Bruce Weaver wrote > > David, I'm getting the following error messages when I run that: > > > >>Error # 4700 in column 7. Text: minNCP > >>An output command contains an unrecognized keyword. The recognized > keywords > >>are OUTFILE, RECORDS, TABLE, and NOTABLE. > >>Execution of this command stops. > > > >>Error # 4702 in column 14. Text: ( > >>An unexpected symbol was encountered on an output command. > > > >>Error # 4700 in column 15. Text: F10.6 > >>An output command contains an unrecognized keyword. The recognized > keywords > >>are OUTFILE, RECORDS, TABLE, and NOTABLE. > > > >>Error # 4702 in column 20. Text: ) > >>An unexpected symbol was encountered on an output command. > > .39506810 4.48934000 > > > > Adding a slash to your first PRINT command fixes it. I.e., changing > > this... > > > > PRINT minNCP (F10.6). > > > > to this... > > > > PRINT /minNCP (F10.6). > > > > With that change, I get the following error-free output: > > > > NCPMacro d=0.84 N1=20 N2=15 . > > .395068 > > 4.489340 > > .39506810 4.48934000 > > > > > > My version info: v 21.0.0.2 for Windows, 64-bit, Windows 7 Professional, > > SP1. > > > > David Marso wrote > >> 40 seconds? > >> This runs in about 1/10th of a second and gives as much accuracy as could > >> be desired! > >> ----- > >> DEFINE NCPMacro (d !TOKENS(1) / N1 !TOKENS(1) /N2 !TOKENS(1) /CL > >> !TOKENS(1) !DEFAULT (95) ). > >> INPUT PROGRAM. > >> LOOP #=1 TO 1. > >> END CASE. > >> END LOOP. > >> END FILE. > >> END INPUT PROGRAM. > >> COMPUTE CL=!CL. > >> COMPUTE df=!N1+!N2-2. > >> COMPUTE tobs=!d/sqr(1/!N1+1/!N2). > >> COMPUTE p1=1-((100-!CL)/100)/2. > >> SET MXLOOPS=100. > >> COMPUTE epsilon=.0000001. > >> VECTOR NCPs(2). > >> LOOP #=1 TO 2. > >> + COMPUTE minabsdif=1. > >> + COMPUTE bottomNCP=-10. > >> + COMPUTE topNCP=10. > >> + COMPUTE newdelta=1. > >> + COMPUTE incr=1. > >> + IF #=1 P=P1. > >> + IF #=2 P=1-P1. > >> + LOOP. > >> + LOOP NCPvalue= bottomNCP TO topNCP BY incr. > >> + COMPUTE delta=newdelta. > >> + COMPUTE prob=ncdf.t(tobs,df,NCPvalue). > >> + COMPUTE newdelta=abs(p-prob). > >> + DO IF newdelta LT minabsdif. > >> + COMPUTE minabsdif=newdelta. > >> + COMPUTE minNCP=NCPvalue. > >> + END IF. > >> + END LOOP. > >> + COMPUTE bottomNCP=minNCP-incr. > >> + COMPUTE topNCP=minNCP+incr. > >> + COMPUTE incr=incr/10. > >> + END LOOP IF delta LE epsilon. > >> + COMPUTE NCPs(#)=minNCP. > >> PRINT minNCP (F10.6). > >> END LOOP. > >> PRINT /NCPs1 NCPs2 (2F12.8). > >> EXECUTE. > >> !ENDDEFINE. > >> NCPMacro d=0.84 N1=20 N2=15 . > >> > >> * result .39506810 4.48934000 > >> valferes wrote > >>> David and Rich, > >>> > >>> The code is entirely correct and provides reliable results (compare with > >>> ESCI > >>> or MOTE outputs). > >>> Moreover, it is done so that any novice user can get > >>> the CI via a simplified input: d, N1 > >>> and N2. > >>> > >>> The time it takes to run the syntax is negligible (on my computer, no > >>> more > >>> than 40 seconds). > >>> > >>> > >>> > >>> > >>> Anyway, I'm waiting for one of two solutions: > >>> > >>> (a) Inverse Noncentral t cdf available in > >>> IBM SPSS 23; > >>> > >>> (b) The solution found > >>> by Rich. > >>> > >>> Sincerely, > >>> > >>> > >>> Valentim > >>> > >>> > >>> > >>> http://gaius.fpce.uc.pt/niips/ > >>> > >>> > >>> > >>> > >>> > >>> > >>> > >>>> Date: Tue, 29 Jul 2014 05:46:56 -0700 > >>>> From: > > >>> david.marso@ > > >>>> Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on > >>>> the Noncentral t-Distribution > >>>> To: > > >>> SPSSX-L@.UGA > > >>>> > >>>> Why not build some loops rather than going at it with a sledge hammer > >>>> by > >>>> building 2 million cases? > >>>> Also get rid of the EXECUTE statements! > >>>> -- > >>>> > >>>> valferes wrote > >>>> > Dear Rich Ulrich,I would be very grateful if you could show me your > >>>> > solution to the same problem: > >>>> > > >>>> > * My data: t observed and degrees of freedom. > >>>> > data list list/tobs(f12.10) df(f8.0). > >>>> > begin data. > >>>> > 2.4592681838 33 > >>>> > end data. > >>>> > > >>>> > * Problem: Find the noncentrality parameters for p1=.025 and p2=.975. > >>>> > My code is here and the illustrative output is here. > >>>> > Sincerely, > >>>> > Valentim R. Alferes > >>>> > http://gaius.fpce.uc.pt/niips/ > >>>> > Date: Tue, 29 Jul 2014 00:45:37 -0400 > >>>> > From: > >>>> > >>>> > rich-ulrich@ > >>>> > >>>> > Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on > >>>> the > >>>> > Noncentral t-Distribution > >>>> > To: > >>>> > >>>> > SPSSX-L@.UGA > >>>> > >>>> > > >>>> > > >>>> > > >>>> > I needed an inverse, back in the days when computers were more > >>>> finite, so > >>>> > I found the other way. > >>>> > > >>>> > The other way to get the inverse from the cumulative distribution is > >>>> to > >>>> > interpolate (using transformed scales that produce approximate > >>>> linearity) > >>>> > and iterate. Two iterations should give you the 4th decimal. > >>>> > > >>>> > -- > >>>> > Rich Ulrich > >>>> > > >>>> > Date: Mon, 28 Jul 2014 19:01:28 +0000 > >>>> > From: > >>>> > >>>> > valferes@ > >>>> > >>>> > Subject: Re: SPSS program: Confidence Interval for Cohen's d Based on > >>>> the > >>>> > Noncentral t-Distribution > >>>> > To: > >>>> > >>>> > SPSSX-L@.UGA > >>>> > >>>> > > >>>> > > >>>> > > >>>> > David,Given that SPSS has not the “Inverse Noncentral t Cumulative > >>>> > Distribution Function”, there is only one way to obtain the accurate > >>>> value > >>>> > of a noncentrality parameter (NCP) for a given probability: to use > >>>> the > >>>> > “Noncentral t Cumulative Distribution Function” and iteratively > >>>> search for > >>>> > the NCP. This is what the code does!The number of cases generated is > >>>> a > >>>> > simple matter of precision (with the current numbers, we obtain > >>>> accurate > >>>> > NCPs up to the fourth decimal place). The two “sort commands” are > >>>> needed > >>>> > because there are two NCP: for p = .025 and p = .975.I hope that SPSS > >>>> 23 > >>>> > will include the “Inverse Noncentral t cdf”, as is the case of MATLAB > >>>> (see > >>>> > http://www.mathworks.com/help/stats/nctinv.html) > >>>> ...===================== > >>>> > To manage your subscription to SPSSX-L, send a message to > >>>> > >>>> > 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 > >>>> > > >>>> > ===================== > >>>> > To manage your subscription to SPSSX-L, send a message to > >>>> > >>>> > 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 > >>>> > >>>> > >>>> > >>>> > >>>> > >>>> ----- > >>>> Please reply to the list and not to my personal email. > >>>> Those desiring my consulting or training services please feel free to > >>>> email me. > >>>> --- > >>>> "Nolite dare sanctum canibus neque mittatis margaritas vestras ante > >>>> porcos ne forte conculcent eas pedibus suis." > >>>> Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff > >>>> in abyssum?" > >>>> -- > >>>> View this message in context: > >>>> http://spssx-discussion.1045642.n5.nabble.com/SPSS-program-Confidence-Interval-for-Cohen-s-d-Based-on-the-Noncentral-t-Distribution-tp5726831p5726844.html > >>>> Sent from the SPSSX Discussion mailing list archive at Nabble.com. > >>>> > >>>> ===================== > >>>> To manage your subscription to SPSSX-L, send a message to > >>>> > > >>> 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 > >>>> > >>>> > >>>> > >>>> > >>> > >>> > >>> ===================== > >>> To manage your subscription to SPSSX-L, send a message to > > >>> 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 > > > > > > ----- > Please reply to the list and not to my personal email. > Those desiring my consulting or training services please feel free to email me. > --- > "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." > Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" > -- > View this message in context: http://spssx-discussion.1045642.n5.nabble.com/SPSS-program-Confidence-Interval-for-Cohen-s-d-Based-on-the-Noncentral-t-Distribution-tp5726831p5726851.html > Sent from the SPSSX Discussion mailing list archive at Nabble.com. > > ===================== > 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 > > > > |
Re: SPSS program: Confidence Interval for Cohen's d Based on the Noncentral t-Distribution
|
Administrator
|
"My initial goal was to condense the calculations into three or four lines of code in order to insert "
Er how about 2 lines: INSERT filename. NCPMacro d=....N1=....N2=....
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |