Activists Fault Nielsen Fault Rates

The Don't Count Us Out Coalition is charging that Nielsen Media Research's "Local People Meter" fault rates for African-American and Hispanic users are "staggering."

Nielsen isn't challenging the fault rates (when Nielsen fails to get data from a home), but counters that the coalition, which Nielsen labels a "News Corp. front group," has "conveniently" picked a single, aberrant, week (the week ending Nov. 21) out of a pattern of declining fault rates. Nielsen also says fault rates don't tell the whole story.

According to the coalition, which is renewing its call for federal oversight of Nielsen, there are "gaping discrepancies" between the fault rates for Blacks and Hispanics vs. the general population in New York, Los Angeles, Chicago and San Francisco.

Nielsen has been rolling out its local people meters in those markets, while at the same time battling the coalition every step of the way. The coalition charges that the meters undercount minority viewing. Nielsen says they instead more accurately count viewing to alternatives, such as cable, though it has also conceded problems with fault rates and pledged to correct them.

News Corp., whose stations have suffered viewing drops post-people meter, has backed the coalition campaign

"In all four cities surveyed," says the coalition, minority fault rates were more than 20% higher for Black and Hispanic households, according to Josh Lahey, DCUO campaign manager.

In San Francisco, says Lahey, fault rates in Black and Hispanic homes were "43% and 123% higher than White homes, respectively."

"Any system in which Blacks and Hispanics are consistently being counted out at such a high rate needs to be re-evaluated," said Lahey.  "It has been seven months since Nielsen launched the LPMs.  By this point, the inconsistencies and fault rates should have been alleviated, not growing."

Nielsen spokesman Matt Tatham responds: "Mr. Lahey, who cherry-picked one random week in November to make his point, is wrong.  In fact fault rates for these groups are declining, not only for the most recent week of Dec. 12, but also over the long term.

"In New York, Chicago and Los Angeles, faulting for African Americans and Latinos is well under 20% and far closer to overall fault rates than Mr. Lahey indicated.  Even more important than the weekly results, however, is the overall trend, in which African American, Latino and overall fault rates are converging.

Mr. Lahey and News Corp. have been using fault rates as a red herring. In fact, measurement quality depends on many things, including sample size, the percentage of targeted households that agree to accept a people meter, and investment in technology, all of which are more important to overall accuracy than fault rates."

One News Corp. executive's unofficial response: Bull. Official response: Check the numbers.

News Corp. referred B&C to Lahey, who said that although the coalition cited one week, he disputed the suggestion that the overall numbers were improving. "If this is improvement, I'd hate to see what they were before."

Lahey supplied a series of Nielsen-sourced charts, including for Chicago, that showed fault rates fluctuating slightly but demonstrating no dramatic changes. Nielsen also supplied a chart for Chicago.

Lahey's chart shows a Hispanic fault rate of 21.5% ending with the week of Nov. 21, pointing out that it is 49% higher than the non-Hispanic household fault rate.

Nielsen's Chicago chart also shows about a 21.5% fault rate (no comparison to non-Hispanics) for that week, but extends the line graph to Nov. 28, showing that the fault rate had fallen to 18%.

Contributing editor John Eggerton has been an editor and/or writer on media regulation, legislation and policy for over four decades, including covering the FCC, FTC, Congress, the major media trade associations, and the federal courts. In addition to Multichannel News and Broadcasting + Cable, his work has appeared in Radio World, TV Technology, TV Fax, This Week in Consumer Electronics, Variety and the Encyclopedia Britannica.