HW 2 comments ...

1)  avoid variables that match function names ...

mean = mean(y1, y2, y3);   * it works (in SAS) but not so great;

mean_ct = mean(y1, y2, y3);  * better;

* even though the following works, I suggest avoiding this;

data data;
  input y x;
...

data data;
  set data;

....

2)  data pkg;
      input y1 condition $ / y2 / y3;
      datalines;
    7.66 Plastic
    6.98 Plastic
    7.90 Plastic
    5.26 Vacuum
    ...
    ;

=======>  "pkg" data set

   Plastic  7.66 6.98 7.90
   Vacuum   5.26 ...
   Mixed 
   Co2 

3)  SQL can be used to create multiple tables - you don't have to quit
    and restart each time you want to build a new table

4)  You could do this assignment using SQL or DATA step programming with IN
    statements.  Suggestion:  if you did this assignment using DATA step
    prog, try doing with SQL (and vice versa).

5)  When do you know code didn't work?  
    * if you see multiple rows with the same ID (ID=1 shows up twice)
    * if you have more rows in the table you construct than data

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

HW 3 questions ...


data hw3_data;
  if RestPulse <= 48 then row_cat=1;
  else if RestPulse <= 52 then row_cat=2;
  else if RestPulse <= 59 then row_cat=3;
  else if RestPulse <= 75 then row_cat=4;
 
/* don't do this */
  if RestPulse <= 48 then row_cat=1;
  if RestPulse <= 52 then row_cat=2;
  if RestPulse <= 59 then row_cat=3;
  if RestPulse <= 75 then row_cat=4;

/* could do this */
  if RestPulse <= 48 then row_cat=1;
  if 48 < RestPulse <= 52 then row_cat=2;
  if 52 < RestPulse <= 59 then row_cat=3;
  if 59 < RestPulse <= 75 then row_cat=4;


proc freq data=hw3_data;
  table row_cat*col_cat / chisq trend;
  exact chisq;
run;