Sentry Page Protection

Please Wait...

Data Analysis [3-15]

---

Analyzing Multiple Segments
(Proc Univariate)

Data is often segmented into multiple groups (e.g. Male vs. Female, Treatment vs. Placebo).

Simply add the CLASS statement to Proc Univariate when having to analyze the data by segments.
​
Example

Data School; Input Student $ Subject : $20. Results; Datalines; UTID501 English 95 UTID501 Math 69 UTID501 History 62 UTID502 English 65 UTID502 Math 96 UTID502 History 70 UTID503 English 62 UTID503 Math 60 UTID503 History 82 UTID504 English 80 UTID504 Math 86 UTID504 History 82 UTID505 English 61 UTID505 Math 86 UTID505 History 68 UTID506 English 60 UTID506 Math 89 UTID506 History 99 UTID507 English 92 UTID507 Math 86 UTID507 History 83 UTID508 English 82 UTID508 Math 66 UTID508 History 93 UTID509 English 73 UTID509 Math 64 UTID509 History 61 UTID510 English 91 UTID510 Math 61 UTID510 History 90 ; Run;

Note: there are 30 observations in this data set. Not all of them are shown in the image above.


The SCHOOL data set contains a list of students along with the subjects and results.

A teacher wants to look at the distribution of the exam results for each subject.

Example

​Proc Univariate Data=School;
Class Subject;
Var Results;
Run;

The CLASS statement specifies the variable Subject as the classification variable.

This allows the analysis to be done for the 3 subjects separately.

Is it the same as using the BY statement?

Yes, when running Proc Univariate, the CLASS statement and BY statement generate very similar results. 

Both method allows you to separate the analysis across multiple segments.

Example

Proc Sort Data=School;
By Subject;
Run;

Proc Univariate Data=School;
By Subject;
Var Results;
Run;

This generates very similar results as using the CLASS option. 

(try it!)

---

Note:

1. When using the BY statement, the data set must be sorted prior to the procedure (Proc Univariate).

The CLASS statement, however, requires no such requirement.
​
2. There are situations where the CLASS statement performs completely different tasks than the BY statement. 

This will be explained in later sessions.

---

Exercise

Locate the CARS data set from the SASHelp library.

Compute the Median Horsepower of Audi and BMW. 

Which car maker has a higher median horsepower?

Next

---

Need some help? 

Get Hint

HINT:
You might need to subset the data set when computing the median horsepower.

Get Solution

SOLUTION:
Proc Univariate Data=SASHelp.cars;
Class Make;
Var Horsepower;
Where Make in ("Audi" "BMW");
Run;

The median horsepowers are 220 and 225 for Audi and BMW, respectively. The cars from BMW have a slightly higher median horsepower.

---

Fill out my online form.

---