Hi there

It takes about 1.5-2 seconds on my PC. Thank you so much for that syntax, I will be using it in the future!!!

Cheers Paul

> Asesoría Bioestadística <bioestadistica@eresmas.net> wrote: > > Hi Richard > > > A. A lot of the time, the sort provided with the system is good. > Could > > you write the elements of your vector out as cases (VARSTOCASES or > > LOOP/XSAVE), use SORT CASES, and re-assemble the vector? > > Mmmm. Right now I'm modifying the code to leave the original dataset > untouched, so I'd rather not write anything into it. > > > B. I do not have this article anymore, but recall it as a clear > > explanation of a fast sorting algorithm, not requiring (I think) > > pointers or other techniques absent from SPSS: > > > > Robert Sedgewick: Implementing Quicksort Programs. > > *Communications of the ACM* 21(10): 847-857 (1978). > > Thanks for the information. I made a Google search for "quicksort > sorting > algorithm", and I found Java code for that fast sorting method. > Unfortunately, I have problems trying to translate it to SPSS MATRIX > code. > I think I would need a complex macro for that. Meanwhile, I sort of > cheated > and found a way of avoiding the sorting step by using a pointer > created > with the MATRIX function GRADE. > > The final version for the code works reasonably fast and in any SPSS > version. I have tried it with W98SE/SPSS 11.5 in a Pentium at 350 Mhz > with > 64 Mb RAM with sample sizes up to 100 each one and it takes less than > 3 > seconds. > > ¿Could anybody give me some feedback about the speed in different > systems? > > Thanks > > Marta Garcia-Granero > > * CI for a difference in 2 medians by Hodges-Lehman method: > Hodges JR and Lehman EL (1963) "Estimates of location based on rank > tests" > Annals of Mathematical Statistics 34, 598-611. > > * Although the example provided has equal sample sizes, > it will work with unequal sample sizes, too. > > * Example dataset & Mann-Whitney test *. > DATA LIST FREE /group(F4.0) wghtgain(F8.0). > BEGIN DATA > 1 175 1 149 1 132 1 187 1 218 1 123 1 151 > 1 248 1 200 1 206 1 219 1 179 1 234 1 206 > 2 142 2 214 2 311 2 249 2 337 2 176 2 262 > 2 211 2 302 2 216 2 195 2 236 2 253 2 199 > END DATA. > VAR LABEL group 'Treatment group' /wghtgain 'Weight gain (pounds)'. > VALUE LABELS group 1 'Control' 2 'Vit. A'. > > * Median & IQR calculation *. > EXAMINE > VARIABLES=wghtgain BY group > /PLOT NONE > /PERCENTILES(25,50,75) > /STATISTICS NONE > /NOTOTAL. > > * Mann-Whitney test *. > NPAR TESTS /M-W= wghtgain BY group(1 2). > > ************************************************ > * CI CODE * > ************************************************ > * If n1*n2>10000, change MXLOOPS accordingly. > * There is a temporary listwise deletion of missing data before > running > MATRIX *. > > SET MXLOOPS=10000. > TEMPORARY. > SELECT IF((~MISSING(group)) AND (~MISSING(wghtgain))). > MATRIX. > PRINT /TITLE ' CONFIDENCE INTERVAL FOR THE DIFFERENCE OF TWO > MEDIANS'. > * Get data: replace variable names by your own *. > GET group /var=group. > GET depvar /var=wghtgain. > * Count sample sizes & separate data in two groups *. > * Assumes group values are 1 & 2, change code values if different*. > COMPUTE n=nrow(group). > COMPUTE n1=0. > COMPUTE n2=0. > LOOP i=1 TO n. > DO IF group(i) EQ 1. > COMPUTE n1=n1+1. > ELSE IF group(i) EQ 2. > COMPUTE n2=n2+1. > END IF. > END LOOP. > COMPUTE group1=MAKE(1,n1,0). > COMPUTE group2=MAKE(1,n2,0). > COMPUTE n1=0. > COMPUTE n2=0. > LOOP i=1 to n. > DO IF group(i) EQ 1. > COMPUTE n1=n1+1. > COMPUTE group1(n1)=depvar(i). > ELSE IF group(i) EQ 2. > COMPUTE n2=n2+1. > COMPUTE group2(n2)=depvar(i). > END IF. > END LOOP. > RELEASE group,depvar,n. > * Compute vector of all differences *. > COMPUTE linear=MAKE(1,n1*n2,0). > LOOP i=1 to n1. > LOOP j=1 to n2. > COMPUTE linear(n2*(i-1)+j)=group1(i)-group2(j). > END LOOP. > END LOOP. > COMPUTE pos=GRADE(linear). > RELEASE group1,group2. > * Computes median of differences *. > COMPUTE half=(n1*n2)/2. > COMPUTE pair=(TRUNC(half)=half). > DO IF (pair eq 0). > COMPUTE medpos=(n1*n2+1)/2. > LOOP i=1 TO n1*n2. > DO IF medpos=pos(i). > BREAK. > END IF. > END LOOP. > COMPUTE median=linear(i). > ELSE IF (pair NE 0). > COMPUTE medpos1=n1*n2/2. > COMPUTE medpos2=1+medpos1. > LOOP i=1 TO n1*n2. > DO IF medpos1=pos(i). > BREAK. > END IF. > END LOOP. > COMPUTE lmedian=linear(i). > LOOP i=1 TO n1*n2. > DO IF medpos2=pos(i). > BREAK. > END IF. > END LOOP. > COMPUTE umedian=linear(i). > COMPUTE median=(lmedian+umedian)/2. > RELEASE lmedian,umedian. > END IF. > PRINT {n1;n2} > /FORMAT="f5.0" > /TITLE="Sample sizes" > /RLABELS=" N1=" " N2=". > PRINT median > /TITLE="Point estimation (*)" > /rlabels="Median" > /FORMAT="f8.1". > PRINT /TITLE="(*) Remember it is NOT the same as the difference of > the 2 > sample medians". > * Exact or asymptotic 95% & 99% CI *. > DO IF ((n1 LE 20) AND (n2 LE 20)). > COMPUTE data95={0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; > 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; > 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; > 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; > 0,0,0,0,2,3,5,6,7,8,9,11,12,13,14,15,17,18,19,20; > 0,0,0,0,3,5,6,8,10,11,13,14,16,17,19,21,22,24,25,27; > 0,0,0,0,5,6,8,10,12,14,16,18,20,22,24,26,28,30,32,34; > 0,0,0,0,6,8,10,13,15,17,19,22,24,26,29,31,34,36,38,41; > 0,0,0,0,7,10,12,15,17,20,23,26,28,31,34,37,39,42,45,48; > 0,0,0,0,8,11,14,17,20,23,26,29,33,36,39,42,45,48,52,55; > 0,0,0,0,9,13,16,19,23,26,30,33,37,40,44,47,51,55,58,62; > 0,0,0,0,11,14,18,22,26,29,33,37,41,45,49,53,57,61,65,69; > 0,0,0,0,12,16,20,24,28,33,37,41,45,50,54,59,63,67,72,76; > 0,0,0,0,13,17,22,26,31,36,40,45,50,55,59,64,69,74,78,83; > 0,0,0,0,14,19,24,29,34,39,44,49,54,59,64,70,75,80,85,90; > 0,0,0,0,15,21,26,31,37,42,47,53,59,64,70,75,81,86,92,98; > 0,0,0,0,17,22,28,34,39,45,51,57,63,69,75,81,87,93,99,105; > 0,0,0,0,18,24,30,36,42,48,55,61,67,74,80,86,93,99,106,112; > 0,0,0,0,19,25,32,38,45,52,58,65,72,78,85,92,99,106,113,119; > 0,0,0,0,20,27,34,41,48,55,62,69,76,83,90,98,105,112,119,127}. > COMPUTE data99={0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; > 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; > 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; > 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; > 0,0,0,0,0,1,1,2,3,4,5,6,7,7,8,9,10,11,12,13; > 0,0,0,0,1,2,3,4,5,6,7,9,10,11,12,13,15,16,17,18; > 0,0,0,0,1,3,4,6,7,9,10,12,13,15,16,18,19,21,22,24; > 0,0,0,0,2,4,6,7,9,11,13,15,17,18,20,22,24,26,28,30; > 0,0,0,0,3,5,7,9,11,13,16,18,20,22,24,27,29,31,33,36; > 0,0,0,0,4,6,9,11,13,16,18,21,24,26,29,31,34,37,39,42; > 0,0,0,0,5,7,10,13,16,18,21,24,27,30,33,36,39,42,45,48; > 0,0,0,0,6,9,12,15,18,21,24,27,31,34,37,41,44,47,51,54; > 0,0,0,0,7,10,13,17,20,24,27,31,34,38,42,45,49,53,57,60; > 0,0,0,0,7,11,15,18,22,26,30,34,38,42,46,50,54,58,63,67; > 0,0,0,0,8,12,16,20,24,29,33,37,42,46,51,55,60,64,69,73; > 0,0,0,0,9,13,18,22,27,31,36,41,45,50,55,60,65,70,74,79; > 0,0,0,0,10,15,19,24,29,34,39,44,49,54,60,65,70,75,81,86; > 0,0,0,0,11,16,21,26,31,37,42,47,53,58,64,70,75,81,87,92; > 0,0,0,0,12,17,22,28,33,39,45,51,57,63,69,74,81,87,93,99; > 0,0,0,0,13,18,24,30,36,42,48,54,60,67,73,79,86,92,99,105}. > COMPUTE u95=data95(n1,n2). > COMPUTE u99=data99(n1,n2). > PRINT /TITLE"Exact confidence intervals calculated (N1 and N2=<20)". > ELSE IF ((n1 GT 20) OR (n2 GT 20)). > COMPUTE u95=TRUNC(n1*n2/2-0.5-1.96*sqrt(n1*n2*(n1+n2+1)/12)). > COMPUTE u99=TRUNC(n1*n2/2-0.5-2.576*sqrt(n1*n2*(n1+n2+1)/12)). > PRINT /TITLE"Asymptotic confidence intervals calculated (N1 and/or > N2>20)". > > END IF. > LOOP i=1 TO n1*n2. > DO IF pos(i) EQ u95. > BREAK. > END IF. > END LOOP. > COMPUTE low95=linear(i). > LOOP i=1 TO n1*n2. > DO IF pos(i) EQ n1*n2-u95. > BREAK. > END IF. > END LOOP. > COMPUTE high95=linear(i). > LOOP i=1 TO n1*n2. > DO IF pos(i) EQ u99. > BREAK. > END IF. > END LOOP. > COMPUTE low99=linear(i). > LOOP i=1 TO n1*n2. > DO IF pos(i) EQ n1*n2-u99. > BREAK. > END IF. > END LOOP. > COMPUTE high99=linear(i). > PRINT {low95,high95;low99,high99} > /FORMAT="f8.1" > /TITLE="CI for difference of medians" > /RLABELS="95%Level" "99%Level" > /CLABELS="Lower" "Upper". > END MATRIX.