MarcR said:
Sounds like what we have is a choice between:
(1) A valid sample and a notoriously unreliable instrument.
The diary has deficiencies, since it measures cume, TSL and memory. Thus the rounding (usually upwards) into even blocks of time and the loss of short incidents of listening to P4 and below stations.
The first enhances or exaggerates TSL, as we know. The second is irrelevant, as light listening cumers are not reachable using reach and frequency goals that are economically realistic since it would talke 100 spots a week to reach most of them.
So the unreliability is basically limited to the rounding factor.
(2) An inadequate sample and a reliable instrument.
The PPM has other issues. It measures TSL, cume and the willingness of panelists to carry it every day, all day. Women show as lighter users of radio, and some of this is suspected to be related to the inability of women in some types of clothing to clip the PPM to theirself... causing loss of carriage, and thus, listening time.
The much smaller size of the panel vs. total quarterly diary count means that the panel must be perfectly sized and balanced at all times or individual cells will not be statistically reliable. The disaster in houston from April to July and August shows that recruiting and keeping a panel are harder than Arbitron envisioned and keeping panelists active is also a major issue.
So, while the meter, when being carried, measures more accurately precise tune in times, it has the defect of panel behaviour, which is being shown to be extremely difficult to maintain within tolerance.
Arbitron's monopoly and their continued use of diaries is indefensible.
There is no plan to implement the PPM outside the top 50 markets at this time, and the top markets will not be rolled out until the end of 2010. The diary will continue to be used, as the cost of the PPM in the smaller markets is at present not affordable given the billings levels and economy of the other 250 markets. The diary is approved in all but 1 US market by the MRC, and agencies will keep using diary data for the forseeable future in non-PPM markets.
Agencies don't trust Arbitron numbers and so they avoid buying radio.
On every call I have made to explain PPM to media departments at agencies with accounts like McDonalds and Ralphs, buyers are used to the diary data and uncertain about PPM and what it means. I never heard any criticism of the diary method, while the PPM was liked because the data would come out monthly within days of the end of each month, a boon to buyers doing seasonal buys and wishing to avoid seasonal bumps in radio listening.
Arbitron gets their feet held to the fire and finally moves toward PPM but early numbers suggest accurate numbers might go against some segments heretofore over-counted. So they play the race card and the industry, as usual, caves in.
No segment was overcounted. The percentage of weighting to acheive proporitonality in the diary is usually very small, and the exceptions transitory. The PPM in New York showed that the meted panel did not approach proportionality, and was off by as much as 40% in some areas. MRC accreditation was the real issue, and general market groups like Clear and Cox and Cumulus demanded what they had been sold. Between the MRC and the major subscribers, Arbitron realized they could not get the panel into compliance by the time of the December planned currency release...
Incidentally, it says in David Hinckley's column in yesterday's Daily News that "Arbitron will continue to release PPM data side-by-side for the next six months with renewed 'diary' ratings" (or at least until the ethnic minority stations succeed in their efforts to suppress that information too).
We will all be looking to see if the sample report shows anything close to 100% DDI's in all cells. Since more ethnic listening is in 18_34 due to population concentrations, the fact that 18-34 is Arbitron's poorest performing broad cell is the first thing we will all look at It is doable. In the last diary survey in Houston the total share for Spanish language stations was 19.5, and in the October PPM it is 19.7, so nobody was being "gifted" with free shares.
