Dear Experts,   In sas, using "proc reg" produces RSTUDENT = ri / sigma(i) *sqrt(1-hi) where sigma(i) is the sum of squared residuals divided by its degree of freedom.   However, in https://onlinecourses.science.psu.edu/stat462/node/247. The externally studentized residual is calculated as ti = ei / sqrt(MSE(i) * (1-hi)) where ei is yi - y^(i) and MSE(i) is the mean square error based on the estimated model.    So sigma(i) not equal to MSE(i). Which method is correct or better?   Thank you LL     ... View more

Dear Experts,   I read that high leverage point has lower variance. Why is it so? From the formula Var(ei) = s*sqrt(1-hi), I can understand. But I thought points at extreme end tends to influence the regression line alot. Hence intuitively, the variance should be higher.   Thank you L ... View more

 Dear Experts   I have two questions: Q1 https://support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_reg_sect039.htm   what is the square bracket in STDR means? What does the STDR means? Is it same as ri / sqrt( 1/(n-1) * sum ri^2 ) where ri is the ith residual computed as yi - yi hat.   Q2.           proc reg data=work.cfs0x1 noprint;             model &alpname = &regvars/noint;             weight glswgt;             output out=&tmp_ds p=pred_alpha_x stdr=stdr_alpha_x;           run;             data work.cfs0x1;             merge work.cfs0x1(in=in_1)               &tmp_ds(in=in_2);               &alpname = (&alpname - pred_alpha_x) / stdr_alpha_x;           run;   Do you know what is the meaning for calculating  &alpname = (&alpname - pred_alpha_x) / stdr_alpha_x ? I read that people called it regression neutralization. I dont see how it really does. ... View more

Dear Experts,   I have a problem back-filling empty rows or rows with no values on the column, named "Returns". Here are the dataset   Inputs: Dataset A: Ticker Date Returns Appl   200209  0.2 Appl   200210  0.1 Appl   200211  0.21 Appl   200212  0.22 BAC   200209  0.1 BAC   200210  . BAC   200212  0.23 C   200209  0.22 C   200210  0.11   Desired Outputs Date       Appl   BAC C 200209  0.2     0.1   0.22 200210  0.1     0      0.11 200211  0.21   0.23 0 200212  0.22   0      0   My code: %* Extract the full date series; proc sort data=A (keep= yyyymm) nodupkey out=A1 by yyyymm;   %let xtemplist = Appl BAC C;   %do k = 1 %to 3;      %let ticker_name = %scan(&xtemplist,&k);   %* transpose the dataset;   data A;     set A;     if ticker = "&ticker_name";       &ticker_name = returns;     keep yyyymm &ticker_name;  run;  data A1;    merge A1(in=a) A (in=b);    by yyyymm;    if a and b;  run; %end;   But the output A1: yyyymm Appl   BAC C 200209  0.2     0.1   0.22 200210  0.1     .       0.11   Do you have any ideas how to get the desired output?   Thank you for your help L       ... View more

Dear Experts   %let XTEMPLIST  =  BA EN CO CA SV IN UT TRA;     proc iml;     s="&xtemplist";     delims = ' ';     n = countw(s, delims);     xtemplist1 = scan(s, 1:3, delims);     xtemplist2 = scan(s, 4:n, delims);     print xtemplist1;     call symput("t",xtemplist1);     call symput("t1",xtemplist2);   quit; %put &t;   %put &t1;   >> t-->only CO is displayed. How to get BA EN CO?      t1->>TRA is displayed. How to get CA SV IN UT TRA?   Thank you for your help :LL ... View more