data sasuser.rwg139t;
set     sasuser.rwg139t;
pval = 2*(probt(-abs(RT_LOG_2), 15));
run;   proc sort data=sasuser.rwg139t; by pval;
proc print; id basin;
var RT_LOG_2 pval D_LOGS_2 blo_IN_2 BLO_LO_2 BLO_L1_2;
run;


                             The SAS System                            

  BASIN     RT_LOG_2    PVAL   D_LOGS_2  BLO_IN_2  BLO_LO_2  BLO_L1_2

  Tsidjior   2.27340  0.03813   0.26957   0.85909   0.70127  -0.78506
  Shotover   1.52489  0.14809   0.15036   0.58899   0.41649  -0.49845
  Karamea   -1.46065  0.16474   0.05543  -0.22503  -0.23686   0.23466
  Inangahu  -1.45600  0.16600   0.05047  -0.21389  -0.13728   0.16821
  Butchers  -1.32672  0.20445   0.04602  -0.22053  -0.19758   0.20902
  L.Hopwoo   1.17649  0.25773   0.04932   0.27116   0.18198  -0.22128
  BullerB   -1.06361  0.30433   0.03985  -0.24352  -0.21206   0.22888
  Arrow      0.88785  0.38864   0.06444   0.12233  -0.09883   0.01445
  BullerA   -0.85221  0.40750   0.02007  -0.13778  -0.08771   0.10831
  Haast      0.71623  0.48485   0.02097  -0.14378  -0.08072   0.11403
  Manuheri  -0.71368  0.48638   0.85892   0.65700   1.31148  -1.09822
  Grey      -0.64580  0.52816   0.01036  -0.08548  -0.07497   0.07941
  Cleddau    0.59112  0.56324   0.01688  -0.14007  -0.08179   0.11255
  Cropp      0.58186  0.56930   0.02671  -0.17424  -0.07229   0.12109
  Hooker    -0.39742  0.69665   0.00593   0.05568   0.01636  -0.03548
  Inangahu  -0.37482  0.71304   0.00329  -0.04189  -0.03435   0.03716
  U.Waitan   0.33817  0.73992   0.00621  -0.08880  -0.05365   0.07230
  Hokitika   0.17981  0.85971   0.00327  -0.07686  -0.05166   0.06532
  Ivory      0.12776  0.90004   0.00100  -0.01030   0.01225  -0.00264

*********************************************************************
Size adjusted cutoff for DFBETAS: 2/sqrt(19) = .4588
Size adjusted cutoff fo Cook's Distance 4/19 = .2105
*********************************************************************
False Discover Rate (FDR) procedure for outliers.

Sort p-values in increasing order as above.

Find the largest value of i where  Pval_i <  alpha*i/n   declare these
cases to be outliers.  Start with i = 19, and then check each.

No cases are declared as outliers.

Notes:  

1) for i = n, compare p-value to alpha  (like in a single hypothesis
test)

2) for i = 1, compare p-value to alpha/n (like the Bonferonni procedure)


more powerful than standard Bonferroni procedure.