PROC UNIVARIATE output explanation

[Tsegazaab]

Hello;

I coded a proc univariate to analyze the variable FEV for smokers and non-smokers. Let me write the code here and the output as well.

PROC MEANS DATA=pulfun;
CLASS smoke;
VAR fev;
RUN;


PROC SORT DATA=pulfun;
BY smoke;
RUN;

PROC UNIVARIATE DATA=pulfun PLOT;
By smoke;
VAR fev;

And, I got the output attached as PDF file. My question is:

"Based on the output of PROC UNIVARIATE, describe the differences and similarities in the shapes of the FEV distributions for smokers and non-smokers.  Specifically identify the features of the output that you used."

Help me please. Thank you!

[Reeza]

It's easier to make a comparison if they're overlayed in some manner. 

Try the histogram statement. 

[Tsegazaab]

So, could you answer my question please? I am not clear with that concept. I am waiting.

Thank you!

[Reeza]

Distributions have specific features that are used for comparison:

measure of center such as median/mean

Tails

Outliers

You can compare the distributions using the graphs and/or the box plot. I find a box plot easier to write up a comparison.

So how do smokers compare for the median and mean for the FEV variable? Are the tails longer in one than the other? In a specific direction? What about outliers between the two?

Khan Academy has a full chapter on this topic (Comparing features of a distribution). They do a better job of explaining this than I would. 

https://www.khanacademy.org/math/statistics-probability/displaying-describing-data/comparing-feature...

This is clearly homework so this as much of an answer I'm willing to provide (my personal rule). Others may (or may not) choose to provide more assistance. Good Luck. 

[Rick_SAS]

Look at the comparative histograms. Also use

ods graphics on;

and get rid of the PLOT option.