• by Joe Mandese  @mp_joemandese, December 23, 2009

Nielsen Tuesday released the latest in an ongoing series of "in-panel" validation studies measuring the compliance of people living in Nielsen households to properly push buttons on their people meters. The studies, which are somewhat controversial, because they are being conducted on Nielsen's "live" ratings panel and theoretically could influence viewing behavior, showed the same pattern as previous studies: a relatively high overall sample performance index, but lingering problems with households in which multiple viewers are present when watching TV.

The most recent release, which was conducted during July and August, showed slightly lower overall agreement rates between Nielsen's actual ratings and the telephone surveys it conducted to validate their results, but Nielsen indicated the differences were not statistically significant and are in line with the overall trend it has been observing, including problems with multiple viewer situations.

"When there is only one person in the audience, net agreement is high, at 94.1%," Nielsen wrote in a report sent to clients Tuesday. "In contrast to the overall results, the viewing index is 1.06, indicating 6% more viewing is recorded by the people meter than is indicated by the coincidental. With two persons in the audience, both measures decline to 82.7% and .85, respectively. As there are more persons in the audience, net agreement between the stated number of persons in the audience and the people meter records declines to a cumulative low of 71.3% with four or more persons in the audience, and a viewing index of .79. This is being addressed with urgency in our remediation program."

Nielsen has been working with ratings households to develop better methods to ensure they push their buttons properly, and said it would continue with that program.