Sample 24998: Compute 95% confidence limits for correlation coefficients using Fisher's Z
Compute 95% confidence limits for correlation coefficients using Fisher's Z
NOTE: Beginning in SAS 9.1, this macro is obsolete. Use the FISHER option
in PROC CORR.
- PURPOSE:
-
Compute 95% confidence intervals for the Pearson correlation
coefficients using the estimates computed by PROC CORR.
- REQUIREMENTS:
-
Base SAS Software Release 6.08 or later.
- USAGE:
-
Follow the instructions in the Downloads tab of this
sample to save the %CORRCI macro definition. Replace the text within quotes in the following statement with the location of the %CORRCI macro definition file on your system. In your SAS program or in the SAS editor window, specify this statement to define the %CORRCI macro and make it available for use:
%inc "<location of your file containing the CORRCI macro>";
Following this statement, you may call the %CORRCI macro. See the Results tab for an example.
The following parameters may be specified:
- DATA=
- SAS data set which contains the original data used
as input to PROC CORR.
- CORR=
- SAS data set of TYPE='CORR'. Most often this is an
OUTP= data set generated by PROC CORR.
- OUT=
- SAS data set which contains 9 variables (10 variables
if BY-group processing is desired):
-
Rirst variable in a pair of variables.
-
Second variable in a pair of variables.
-
The Pearson correlation coefficient for
the pair of variables.
-
The number of observations used to compute
the correlation.
-
The p-value for a significance test for
the correlation coefficient.
-
The BY variable (in the case of BY-group
processing).
-
Fisher's Z value.
-
Sigma value for Fisher's Z.
-
Lower 95 percent confidence limit.
-
upper 95 percent confidence limit.
- BYVAR=
- Variable name used in BY-group processing. Omit this parameter if BY-group processing is not needed.
- REFERENCES:
-
Snedecor & Cochran (1980), Statistical Methods, Iowa State
University Press, pp. 185-188.
These sample files and code examples are provided by SAS Institute
Inc. "as is" without warranty of any kind, either express or implied, including
but not limited to the implied warranties of merchantability and fitness for a
particular purpose. Recipients acknowledge and agree that SAS Institute shall
not be liable for any damages whatsoever arising out of their use of this material.
In addition, SAS Institute will provide no support for the materials contained herein.
These sample files and code examples are provided by SAS Institute
Inc. "as is" without warranty of any kind, either express or implied, including
but not limited to the implied warranties of merchantability and fitness for a
particular purpose. Recipients acknowledge and agree that SAS Institute shall
not be liable for any damages whatsoever arising out of their use of this material.
In addition, SAS Institute will provide no support for the materials contained herein.
- EXAMPLE:
-
The following statements create data set ONE containing variables X1, X2, and X3. This data set is then analyzed by PROC CORR and an output data set, TWO, is created containing the matrix of Pearson correlation coefficients. The %CORRCI macro reads the correlation data set and creates data set THREE containing correlations and 95% confidence intervals for each.
data one;
input x1 x2 x3;
datalines;
11.84707 16.93533 30.30385
7.59473 15.75173 29.40286
10.63967 20.48042 27.80475
12.82885 23.87893 31.90823
9.17214 19.37775 30.86424
9.37882 20.29886 31.44052
8.91870 20.58530 28.72000
10.67217 17.43738 26.87034
9.04225 20.03456 26.29315
9.23498 15.70513 28.45162
7.28257 19.00768 25.49542
9.42358 20.35338 25.70803
8.99361 20.23768 30.99658
9.48139 19.93371 31.81558
13.41978 27.21234 33.53439
10.38762 16.71561 32.75510
9.63953 20.25250 26.78733
9.67609 23.03686 31.54124
7.97324 20.15488 24.73205
9.94725 18.38008 31.03492
7.47832 15.45577 29.69831
11.17087 23.22185 27.57108
8.98054 18.06027 23.70008
9.85106 22.46030 25.40800
6.83538 20.26074 27.11173
10.43364 21.54151 27.83231
8.95012 17.96959 27.38565
7.68489 21.21179 25.61499
8.76266 18.79180 21.63533
13.56824 22.65316 28.15209
5.97392 16.79257 24.26937
11.55536 19.79710 29.47594
14.46293 23.32019 33.16632
9.31360 21.66007 30.16746
7.33522 18.12996 25.32852
13.01646 24.63966 33.67223
10.46776 21.55677 37.31202
8.74968 26.72172 32.41989
11.05527 22.03290 33.10484
8.36753 22.00099 30.53665
9.31461 17.04604 27.71464
9.71028 14.90492 26.61771
10.50100 15.93331 26.58496
11.59753 18.06033 40.43878
14.56284 25.87312 34.55972
5.31291 19.64416 22.67945
6.65624 14.03495 27.14256
10.08980 22.54291 26.70714
8.44074 22.22102 29.27018
10.05060 20.59674 35.32499
8.56124 20.22675 28.47314
9.10676 19.87403 29.62240
6.61521 21.54044 25.08155
13.28789 15.94131 32.08812
12.06948 16.36298 31.77478
11.33812 18.66243 38.83121
7.47306 14.87482 28.51459
4.43891 19.59826 29.03008
6.07296 16.31533 24.21147
9.41121 25.62246 32.59980
14.28650 21.81332 35.19784
11.23610 19.98522 31.17991
5.52991 18.83095 24.89066
11.21624 16.00669 32.16123
11.41885 17.67534 35.02728
11.77425 20.11171 27.89257
7.36826 14.94006 25.83307
9.99781 18.88736 29.85778
12.04007 12.75452 37.18675
13.12532 18.59637 35.71118
9.30269 21.77130 28.31119
5.40426 19.75766 24.20834
11.30906 24.09407 25.76265
12.20726 19.51309 35.66550
8.93185 17.76375 31.34298
11.24166 15.75918 35.39888
10.37316 19.40015 24.48352
13.73441 23.45906 28.77281
8.53062 18.39039 20.96273
12.13759 21.77956 34.75851
10.09987 19.51518 33.23135
8.23613 17.74195 29.09213
9.45489 14.41677 31.37922
8.95388 19.47108 32.11275
11.80227 16.32790 39.37990
8.22948 18.46778 30.37589
11.58760 20.06535 32.30379
4.49313 14.64329 22.27903
6.16884 21.58907 27.46434
8.75492 16.30769 29.82571
9.43899 20.52954 35.47211
11.81902 20.00596 35.64332
10.49819 24.52172 30.61800
12.15001 17.70018 28.11096
9.71977 12.24687 30.56766
13.24286 16.56150 27.18675
11.07728 17.64190 33.02688
6.20052 20.41377 27.66562
7.57754 27.33327 27.75320
12.77903 20.09754 28.94103
;
proc corr data=one outp=two;
var x1 x2 x3 ;
run;
/* Define the CORRCI macro */
%inc "<location of your file containing the CORRCI macro>";
%corrci(data=one, corr=two, out=three);
proc print data=three noobs;
run;
- RESULTS:
- Following are the results of PROC PRINT from the above example showing the contents of the OUT= data set from the %CORRCI macro.
var1 var2 corr n pvalue fishersz sigmaz l95 u95
x1 x2 0.21943 100 0.02827 0.22306 0.10153 0.02404 0.39867
x1 x3 0.57486 100 0.00000 0.65475 0.10153 0.42661 0.69303
x2 x3 0.07968 100 0.43069 0.07985 0.10153 -0.11860 0.27184
Right-click on the link below and select Save to save
the %CORRCI macro definition
to a file. It is recommended that you name the file
corrci.sas.
Download and save corrci.sas
Compute 95% confidence intervals for the Pearson correlation coefficients using the estimates computed by PROC CORR.
| Type: | Sample |
| Topic: | SAS Reference ==> Procedures ==> CORR
Analytics ==> Nonparametric Analysis
Analytics ==> Descriptive Statistics
|
| Date Modified: | 2007-08-11 03:03:09 |
| Date Created: | 2005-01-13 15:03:10 |
Operating System and Release Information
| SAS System | Base SAS | All | n/a | n/a |
| SAS System | SAS/STAT | All | n/a | n/a |
