The script begins by creating a `SocEcon` table containing five socio-economic variables via a DATA step. It then proceeds to calculate a correlation matrix with `PROC CORR`. A second CORR type table is also created manually. The core of the analysis relies on `PROC REG`, which is used twice: first to fit two regression models and save their estimates and covariances (`OUTEST=regest COVOUT`), then to generate the matrix of sums of squares and cross products (`OUTSSCP=regsscp`). The intermediate result tables (`corrcorr`, `datacorr`, `regest`, `regsscp`) are displayed with `PROC PRINT` for inspection.
Data Analysis
Type : CREATION_INTERNE
The two data tables used, `SocEcon` and `datacorr`, are created within the script using the `datalines` instruction. No external data source is necessary.
1 Code Block
DATA STEP Data
Explanation :
This DATA STEP block creates the `SocEcon` working table. It reads 12 observations for 5 socio-economic variables (Pop, School, Employ, Services, House) directly from the code using the `datalines` instruction.
This DATA STEP block creates the `SocEcon` working table. It reads 12 observations for 5 socio-economic variables (Pop, School, Employ, Services, House) directly from the code using the `datalines` instruction.
Copied!
| 1 | title 'Five Socioeconomic Variables'; |
| 2 | title2 'Harman (1976), Modern Factor Analysis, Third Edition'; |
| 3 | |
| 4 | DATA SocEcon; |
| 5 | INPUT Pop School Employ Services House; |
| 6 | DATALINES; |
| 7 | 5700 12.8 2500 270 25000 |
| 8 | 1000 10.9 600 10 10000 |
| 9 | 3400 8.8 1000 10 9000 |
| 10 | 3800 13.6 1700 140 25000 |
| 11 | 4000 12.8 1600 140 25000 |
| 12 | 8200 8.3 2600 60 12000 |
| 13 | 1200 11.4 400 10 16000 |
| 14 | 9100 11.5 3300 60 14000 |
| 15 | 9900 12.5 3400 180 18000 |
| 16 | 9600 13.7 3600 390 25000 |
| 17 | 9600 9.6 3300 80 12000 |
| 18 | 9400 11.4 4000 100 13000 |
| 19 | ; |
| 20 |
2 Code Block
PROC CORR Data
Explanation :
`PROC CORR` procedure calculates the correlation matrix for all numeric variables in the `SocEcon` table. The `noprint` option suppresses display in the standard output, and `out=corrcorr` saves the results in a new table named `corrcorr`. This table is then displayed with `PROC PRINT` and its structure is inspected with `PROC CONTENTS`.
`PROC CORR` procedure calculates the correlation matrix for all numeric variables in the `SocEcon` table. The `noprint` option suppresses display in the standard output, and `out=corrcorr` saves the results in a new table named `corrcorr`. This table is then displayed with `PROC PRINT` and its structure is inspected with `PROC CONTENTS`.
Copied!
| 1 | PROC CORR noprint out=corrcorr; |
| 2 | RUN; |
| 3 | |
| 4 | PROC PRINT DATA=corrcorr; |
| 5 | RUN; |
| 6 | |
| 7 | PROC CONTENTS DATA=corrcorr; |
| 8 | RUN; |
3 Code Block
DATA STEP Data
Explanation :
This DATA STEP manually creates a special SAS CORR type table named `datacorr`. This table type can be used as input by other statistical procedures. The correlation matrix data is read from the `datalines`. The resulting table is displayed with `PROC PRINT`.
This DATA STEP manually creates a special SAS CORR type table named `datacorr`. This table type can be used as input by other statistical procedures. The correlation matrix data is read from the `datalines`. The resulting table is displayed with `PROC PRINT`.
Copied!
| 1 | title 'Five Socioeconomic Variables'; |
| 2 | |
| 3 | DATA datacorr(type=corr); |
| 4 | INFILE CARDS missover; |
| 5 | _type_='corr'; |
| 6 | INPUT _Name_ $ Pop School Employ Services House; |
| 7 | DATALINES; |
| 8 | Pop 1.00000 |
| 9 | School 0.00975 1.00000 |
| 10 | Employ 0.97245 0.15428 1.00000 |
| 11 | Services 0.43887 0.69141 0.51472 1.00000 |
| 12 | House 0.02241 0.86307 0.12193 0.77765 1.00000 |
| 13 | ; |
| 14 | |
| 15 | PROC PRINT DATA=datacorr; |
| 16 | RUN; |
4 Code Block
PROC REG Data
Explanation :
This `PROC REG` procedure performs two linear regression analyses on the `SocEcon` table. The results (parameter estimates) are stored in the `regest` table using `outest=regest`. The `covout` option includes the covariance matrix of the estimates. Both models, 'full' and 'empser', are specified. The `regest` result table is then printed.
This `PROC REG` procedure performs two linear regression analyses on the `SocEcon` table. The results (parameter estimates) are stored in the `regest` table using `outest=regest`. The `covout` option includes the covariance matrix of the estimates. Both models, 'full' and 'empser', are specified. The `regest` result table is then printed.
Copied!
| 1 | PROC REG DATA=SocEcon outest=regest covout; |
| 2 | full: model house=pop school employ services / noprint; |
| 3 | empser: model house=employ services / noprint; |
| 4 | QUIT; |
| 5 | |
| 6 | PROC PRINT DATA=regest; |
| 7 | RUN; |
5 Code Block
PROC REG Data
Explanation :
A second execution of `PROC REG` is used to generate the matrix of sums of squares and cross products (SSCP) for the model variables. The result is stored in the `regsscp` table via the `outsscp=regsscp` option. The `regsscp` table is then displayed for verification.
A second execution of `PROC REG` is used to generate the matrix of sums of squares and cross products (SSCP) for the model variables. The result is stored in the `regsscp` table via the `outsscp=regsscp` option. The `regsscp` table is then displayed for verification.
Copied!
| 1 | PROC REG DATA=SocEcon outsscp=regsscp; |
| 2 | model house=pop school employ services / noprint; |
| 3 | QUIT; |
| 4 | |
| 5 | PROC PRINT DATA=regsscp; |
| 6 | RUN; |
This material is provided "as is" by We Are Cas. There are no warranties, expressed or implied, as to merchantability or fitness for a particular purpose regarding the materials or code contained herein. We Are Cas is not responsible for errors in this material as it now exists or will exist, nor does We Are Cas provide technical support for it.
