An argument favoring a sequential experiment design over a predetermined sample size in the approach to HUD's plan to test the effects on households of various housing allowance schemes. An informal sequential approach can be expected to be more cos...
