Nielsen Reports Declines In Fault Rates.

Nielsen Media Research reported that fault rates, the percentage of sample homes that do not report complete television ratings information on any given day, have declined markedly from November to May in all four markets where Local People Meters (LPMs) were introduced in 2004.

Nielsen said continued improvements in fault rates would be difficult over the next several months because they usually increases during the summer, when children are home from school, viewing patterns are different, power outages are more common, and families go on vacation or otherwise spend time away from home and Nielsen has difficulty gaining access to sample homes.

LPMs are an advanced system of electronic meters that electronically and continuously record demographic viewing in sample homes. Nielsen is introducing them in the top ten markets, replacing the previous system of paper diaries and set-top meters.

Fault rates are one measure of sample quality. From November 2004 to May 2005, overall LPM fault rates declined in three of the four markets:

New York -10.4%
Los Angeles +1.6%
Chicago -24.7%
San Francisco -22.4%

Fault rates declined among African American households in all four markets and among Latino/Hispanic households in three of the four markets. From November 2004 to May 2005, African American fault rate declines were:

New York -22.0%
Los Angeles -14.4%
Chicago -18.1%
San Francisco -22.9%

Among Latino/Hispanics, fault rate results during this same period were:

New York +1.7%
Los Angeles -14.8%
Chicago -14.3%
San Francisco -33.7%

“Faults” occur for many reasons, including the installation of new equipment (television, DVD, or VCR, etc), electrical failure, machine failure, telephone failure, out-of-sync tuning of a television, and, in a people meter, a failure to log in properly. Faulting is generally higher among larger or younger households, households with children, households that watch more television and households with more VCRs, video games and other devices. The more television activity there is in a home, the greater possibility there is for activities that cause faulting.

While fault rates are a component of a sample’s quality, they are not the most important one. Other factors that are more critical to a sample’s accuracy are the size of a sample, the percentage of targeted households that agree to accept a people meter, or the demographic representation of the sample. Faulting does not significantly affect the accuracy of ratings because, on a daily basis, Nielsen “weights” the contributing sample households in the demographic, thereby moving closer to the true estimate of viewing by that demographic.

Over the past few months, Nielsen has introduced several quality initiatives to reduce fault rates as part of its commitment to continuous improvement in all aspects of sample quality. These initiatives include increased field personnel, personal coaching for demographic groups with the highest fault rates, and the installation of dedicated phone lines.

Fault Rates Lower Under People Meters than Set Meters

Nielsen also announced that in all markets (including Philadelphia and Washington DC, which are about to be launched in July), fault rates are considerably better in the new LPM samples than they were in the set meter samples they are replacing (see Table 2 for more detail).

In the four most recent LPM markets, set faults were lower for all audiences in May 2005 than they were in May 2004, when the set meter was being used (these are direct comparisons and as such do not include personal faulting, such as a failure to log in correctly or fill out a diary completely, since set meter samples do not have persons faults.) The same is
true in Philadelphia and Washington DC, where LPM data will become commercially available in July. In these two markets, LPM and set meter were compared for May 2005, when both systems were operational. With lower fault rates representing improvements, the magnitude of overall improvement in these markets was:

New York -31.7%
Los Angeles -17.9%
Chicago -34.5%
San Francisco -19.1%
Philadelphia -22.2%
Washington -8.0%

Among African Americans, LPM fault rates in May were lower than set meter fault rates a year ago in five of the six new LPM markets:

New York -44.1%
Los Angeles +3.7%
Chicago -12.5%
San Francisco -39.2%
Philadelphia -22.1%
Washington -0.9%

Among Latino/Hispanics, LPM fault rates in May were lower than set meter fault rates a year ago in five of the six new LPM markets:

New York -39.4%
Los Angeles -26.5%
Chicago -35.7%
San Francisco +28.3%
Philadelphia -36.0%
Washington -16.1%

For more information at http://www.vnu.com