[EMAIL PROTECTED] (Manola) wrote in message news:<[EMAIL PROTECTED]>... > I work for a processing company. > > We have widgets that we believe to be out of specification. Using > sampling tables, we determined that our sampling size should be 800 > widgets. > > Ahead we went measuring and measuring, and after going through 200 > widgets we have seen that every single one is out of spec. At this > point we question if it is necessary to continue or if it is OK to > stop. If we draw a conclusion now, how confident can we be?
What were the sampling tables for? Presumably they are for assessing some particularly rare rate that you're obviously well in excess of. You don't need statistics here. /Obviously/ it's no good. Glen . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
