Breaking Down The PIA Ratios --Dickeson

For starters, these PIA numbers are "survey" data, not random sample results. Does that make a difference? Of course it does. We can't say that these results are "representative" of the 45,000-printer universe. Nine hundred printers, if randomly chosen, would provide a valid random statistic, but 900 volunteers do not. They're 900 companies willing to open their kimonos—sorta.