From Inside Radio:
Look for less ratings bounce as Arbitron implements new panelist expiration rules. Under existing rules, almost 20% of a market’s PPM panel could turn over during a three-month period. That often causes wide ratings swings from month to month as new sample is ferried in to replace panelists who had reached the expiration point of their tenure. Beginning with the November survey, Arbitron will implement new sample expiration rules intended to smooth out the turnover. “[The current system] creates a lot of turnover in the panel all at once, which creates increased variance in the estimates,” [Arbitron’s Beth] Webb says. “We wanted to reduce the impact by smoothing over the turnover…You won’t get a three-month period where there’s a huge turnover in the panel. This will help the stability of the estimates in that period.”
Put another way, smoothness creates the impression of validity while volatility creates the impression of non-validity.
None of this has anything to do with validity, of course, but it has everything to do with impression.
When the ratings “bounce” and create “wide ratings swings” because the sample changes does that mean the new sample is wrong, the old one was wrong, or the system was wrong? Or all of the above?
Or does it just mean that ratings – like all statistical estimates – are built to “bounce,” no matter how much we pretend they wouldn’t?
This issue throws light on a system where accuracy is ultimately less important than stubborn consistency masquerading as accuracy.
This is so because the samples are only panel-sized.
When we worship predictable numbers rather than true ones, what favors are we doing our clients and ourselves?
Like the ocean itself, sometimes a shiny and smooth surface can hide a tumult below.
Maybe Arbitron – and radio itself – should focus less on the surface and more on the depths.