08-18-2010 10:21 AM
08-18-2010 12:10 PM
10-12-2011 05:04 PM
Note that this approach makes the endpoints of the range half as likely to be realized as the internal points.
Rick Wicklin's blog post looks at continuous ranges and integer ranges; here we have non-integer but discrete ranges.
Something like this would work.
To generalize it you would need to create a macro or use proc fcmp.
09-21-2013 02:35 AM
Hi Rick! Do you a sample simulation code to generate multilevel data (nested data)? An example of this is a cluster-randomized trial (CRT) design where the treatment is implemented at the group level while the outcome is measured at the individual level?