A client wanted to migrate a randomization being done in Oracle to a Microsoft SQL Server 2008. Does anyone have any experience with the adequacy and best randomization approach? Efficiency is less important than the quality, although multiple million random numbers are required. Any subroutine better than a linear congruential approach? or subroutine with known validation? A commercially available program is perfectly acceptable. The output of this SQL randomization will be used for state governmental work.
-------------------------------------------
Allen Fleishman, PhD, PStat®
President
Allen Fleishman Biostatistics
-------------------------------------------