Nielsen Probe Finds People Not Pushing Buttons Properly: The Bigger The Audience, The Bigger The Discrepancy

Nielsen Wednesday began informing advertisers, agencies and television networks that rely on its national TV ratings that a significant number of people don't use its so-called "people meter" system properly. In fact, new data being released by Nielsen shows that the more people that are present in those TV ratings households, the more likely it is that they are not reporting their viewing accurately, and that the overall affect has been understating the national TV ratings by about 8%.

Nielsen did not release information showing what the effect was on specific TV outlets, programs or dayparts, and the average is the net result of people who both overstate and understate their viewing either by pushing people meter buttons when they are not actually watching TV, or by forgetting to push them when the are watching.

The findings are the first results to be released by Nielsen from a somewhat controversial research initiative it began last year, and which it will continue conducting quarterly on an ongoing basis in order to better understand how people comply with the people meter ratings system, and more importantly, how Nielsen can take steps to improve that compliance via better coaching methods, and by focusing on the most problematic households.

advertisement

advertisement

The research is somewhat controversial, because it is being conducted on Nielsen's live "currency" sample, which raises the specter that it could influence how people in the sample actually watch television. The research method being used, involves Nielsen representatives calling ratings households on the phone while they are watching TV, and asking the person who answers the phone what they or other people in the household are watching at the time.

Nielsen has already gotten the blessing of key clients, as well as industry ratings watchdog the Media Rating Council, to conduct the research, but it is being conducted even as Nielsen conducts another, very different test on its live sample to measure their Internet usage alongside their TV viewing (MediaDailyNews April 28).

While there always is some concern when Nielsen conducts separate research on its live TV ratings sample, Nielsen executives said both the studies are important, prudent steps toward improving its ratings, and that they are carefully monitoring how it is affecting the quality of its national TV ratings.

"There is no perfect research tool, but we wanted to find out if there were ways to improve people's people meter performance," Pat McDonough, senior vice president of Planning Policy & Analysis at Nielsen told MDN during a telephone interview explaining why Nielsen embarked on the research.

Nielsen executives said the level of compliance among TV ratings households is consistent with earlier studies Nielsen conducted during the late 1980s when it first introduced people meters to replace its old paper diary and TV set meter system, and during the 1990s. The new research, like the older studies, found that, on average, about 10% of households inaccurately record their viewing via people meters - either by pushing buttons when they are not watching TV, or by not pushing them when they are watching. The net effect is an understatement of 8%, because more people forget to push the buttons than those who push them when they are not watching.

While that range is consistent with previous studies, Nielsen has begun drilling deeper into the behavior of its panelists to find out which ones and what types tend to be the worst offenders. And while the data is still limited, Nielsen's top researcher, Bruce Hoynoski said there appear to be no differences among people based on age or gender of the people in the households, but it has found that Hispanic households, and households that have four or more viewers present in the audience, tend to be the worst offenders. In fact, the more people present in the audience, the more likely it is that the ratings will be inaccurate.

"For one person in the audience the compliance was 93%. It fell to the high 80s when two person are present in the audience. It fell to the mid- to lower 80s with three person in the audience. And it fell into the 70s when you had four or more person in the audience," Hoynoski explained.

He said the insights will enable Nielsen researchers to better deploy their field staff to focus their coaching procedures or adapt them to deal with the most difficult households. So far, that appears to be Hispanics and large audience households, but he said future waves of the phone research could yield insights about other household types that could be addressed later.

The finding could be troubling for media planners and advertisers, because programming that attracts larger audiences - so-called "family-viewing" programming, or water cooler events like the Super Bowl or the Academy Awards - tend to be among the most valuable and expensive for advertisers to buy.

3 comments about "Nielsen Probe Finds People Not Pushing Buttons Properly: The Bigger The Audience, The Bigger The Discrepancy".
Check to receive email when comments are posted.
  1. Stanley Sanders from SWS Inc., April 30, 2009 at 11:05 a.m.

    Is anyone in Broadcasting surprised by this news from Nielsen? People Meters have had issues from the start and Broadcasters have raised concerns about their results.

    It is admirable to see Nielsen acknowledge there are discrepancies, let's hope their research leads to much needed improvement within the People Meter system.

    During these challenging economic times, Broadcasters should not suffer a loss of audience due to a system that failed to measure them properly.

  2. Edward Rincon, April 30, 2009 at 11:14 a.m.

    Interesting that Nielsen "discovered" a problem that has been in existence for years, especially regarding the HIspanic sample. The Latino Television Study that we conducted in 2004 - sponsored by The National Latino Media Coalition and funded by CBS and ABC networks -- identified various problems with Nielsen's methodology regarding Hispanic television households -- including the absence of independent audits to examine the causes of higher non-compliance rates by Hispanic households. Rather than deal with the issues raised, Nielsen hired an Hispanic policy research firm to "debunk" the issues raised in our study, posted criticisms of the study on their web site, and invested heavily in public relations with selected organizations to help the industry forget that these problems existed. A copy of the Latino Television Study can be found at www.rinconassoc.com under the Publications section.
    Edward T. Rincón, Ph.D.
    Rincón & Associates
    Ph: 214-750-0102

  3. John Grono from GAP Research, April 30, 2009 at 8:17 p.m.

    From an outsiders perspective I'm surprised that such compliance checks are not conducted on an on-going basis. Down here in Australia they sure are!

    First, no-one should be 'scared' that this 'research' is being done on the live panel - it should be encouraged so as to produce the most robust possible ratings.

    It should also not be of any surprise that larger homes are less compliant - the more people there are, the more things can go wrong. It is the same with equipment faults - the more TVs, decoder boxes, DVRs etc - the more the home will fault.

    There could however be an underlying issue. It appears as though the entire home may be faulting if there is a single non-compliant household member. Compliance is generally a "people issue". If a people-fault is faulting the entire household this could be looked at.

    Now, given this knowledge, there are corrective measures that can be taken. For example, the panel can be weighted by household size. For example, the 4+ person households can be weighted higher than the 1-person households as we know that they fault at a higher level. I would be STUNNED if this was not already happening - if anyone knows, please respond.

    There are some interesting things that we have found with compliance checking down here in Australia. Many of the "faults" are actually people responding incorrectly over the telephone. Common instances are people claiming to watch a programme (say The Simpsons) on broadcast TV when they were actually watching it on cable TV - this is counted as a system error when it actually isn't! Further, where broadcsat areas overlap, people may be claiming to watch in their "home market" when they are watching 'spill-in' television - again this counts as an error.

    A further benefit from compliance checking calls is that we are finding extra TV sets that have not been metered. That is, the home may have acquired a flat-screen TV, kept the old CRT, but forgot to notify the ratings company. It is then a simple matter to arrange a meter installation.

    A further analysis that can be conducted is to look at both the Top Ten Percent and Bottom Ten Percent of viewers. If a home's viewing is abnormally low it may again be an unmetered TV set. If a home's viewing is abnormally high it may be someone not using the meter correctly and forgetting to log themselves out. As a guideline, between 90%-95% of the panel ae doing everything correctly.

    I'd also be extremely cautious saying that panel issues are race-based. In my experience, most apparent race issues are cohort effects to other more dominant factors - such as household size, the amount of equipment in the home etc. This is not to say that recruitment and retention of NESB (Non English Speaking Background) homes isn't an issue. In Australia we use specialist recruiters fluent in a multitude of languages. All documentation is available in the 12 most dominant languages. We retain "cultural advisers" to help in the "tone" of recruitment and panel management ... and we STILL have issues with certain ethnic groups, simply because they have a greater propensity to refuse to participate in research!

    John Grono, GAP Research, Sydney Australia

Next story loading loading..