OPTIONS LS=75 NODATE;

DATA big;
   INFILE "heightdat.lst" FIRSTOBS=5;
   INPUT parent son;
RUN;

PROC MEANS;
   VAR son parent;
RUN;

DATA questions;
   /* these values used for prediction or estimation at x=68,x=72 */
   INPUT parent son;
   CARDS;
   68 .
   72 .
;

DATA big;
   SET big questions;

PROC REG DATA=big;  
*PROC REG SIMPLE; 
   MODEL son=parent;
   *MODEL son=parent/CLB;
   OUTPUT OUT=out1 RESIDUAL=r P=yhat UCL=pihigh LCL=pilow UCLM=cihigh LCLM=cilow STDP=stdmean;


DATA questions;
   SET out1;
   IF son=.;
RUN;

PROC PRINT;  /* default dataset here is "questions" */
   TITLE "questions regarding prediction, estimation when x=68, x=72";
RUN;

PROC CORR DATA=big; /* we had to switch back to "big" data*/
   VAR son parent;

DATA fisherz;
   n=928;
   r=SQRT(0.2105);
   rratio=(1+r)/(1-r);
   z=PROBIT(0.975);
   expon=PROBIT(0.975)/SQRT(n-3);
   rlow=(rratio*exp(-2*expon)-1)/(rratio*exp(-2*expon)+1);
   rhigh=(rratio*exp(2*expon)-1)/(rratio*exp(2*expon)+1);
RUN;

PROC PRINT;
   TITLE "computations of 95% confidence interval for rho";
RUN;

OPTIONS ps=50;

/*PROC UNIVARIATE DATA=out1 NORMAL PLOT; 
   TITLE "residual analysis";
   VAR r;
RUN;

*GOPTIONS DEV=pslepsf COLORS=(BLACK);

SYMBOL1 I=RL
   VALUE=DOT;

PROC GPLOT DATA=out1;
   PLOT son*parent;
RUN;

PROC GPLOT DATA=out1;
   PLOT r*yhat;
RUN;
QUIT;*/