These Neilsen Ratings?

[Crewman47]

I do have a small understanding of what these are from what's been said on here and reading it up on Wiki and the way I understand it that people are selected at random to have this box installed in there home in order to check what there watching. With the introduction of Tivo to American homes this would mean a lot more homes to get ratings from, am I right in this assumption?

Anyway what I wondering was if any of our American members in here have one of these boxes (if I'm right in assuming that's how it work) or knows of anyone with one, or even a Tivo? Just asking cause with all the talk of a lot of reasonably good shows being axed lately, despite a lot of people liking them, I'm wondering if the ratings are coming from the wrong target audience and the ratings devices should be given to a larger populace.

Please correct me if I'm totally wrong in this assumption.

 

[JustKate]

Well, the Nielsens use only 25,000 homes across the U.S., which is a very small segment of the population, so the chances that one of them happens to have a member who posts here are not really very high.

And actually, they aren't selected "at random" - they are scientifically selected to be representative. You might like to read about it on Nielsen's website (it's a really long URL, so I've used TinyURL.com to make it less cumbersome): http://tinyurl.com/z3are. Perhaps this will make clear anything that Wiki did not.

 

[Bob The Skutter]

^I believe only around 5000 are actual boxes, the others are detailed diaries, or just forms to fill in.
BARB in here in the UK have a similar system.
Tivo charge for their data, which the advertisers try to have those viewers removed from their fees, because people watching timeshifted recordings are probably not watching the ads, so being popular on Tivo could very well be a disadvantage to a lot of shows.

 

[Mr. B]

Like Kate said they only have 25,000 samples. Their methodology is pathetic no matter how "scientific" their sample selection process is.

Neilson boxes are attached to every TV in a Neilson home and there isn't much interaction required. Each member of the household has their own button which they are supposed to press to indicate who, if anyone is watching the TV when it's on.

At present, Arbitron (ratio rating company) is perfecting personal people meters, a self-contained pager-like device which would be able to detect what if anything the wearer is watching or listening to.

 

[JustKate]

As for "only 25,000 samples," you can get good results with that number, if the households are properly chosen and the system works the way it's supposed to. Not saying it does, but it can. I mean, the boxes are great at figuring out that the TV is on and what happens to be showing on the screen, but as for who, if anybody, in the house is watching it...that's a much trickier thing.

A long time ago, I participated in...something. I can't remember what, though, butt wasn't Nielsen. It required me to keep a log because there was no box. That was really time-consuming to do, and I can promise that the results in turned in weren't all that accurate.

 

[John Picard]

As for "only 25,000 samples," you can get good results with that number, if the households are properly chosen and the system works the way it's supposed to. Not saying it does, but it can. I mean, the boxes are great at figuring out that the TV is on and what happens to be showing on the screen, but as for who, if anybody, in the house is watching it...that's a much trickier thing.

A long time ago, I participated in...something. I can't remember what, though, butt wasn't Nielsen. It required me to keep a log because there was no box. That was really time-consuming to do, and I can promise that the results in turned in weren't all that accurate.

There are approximately what -- 175 million households in the US, and you want to claim that 25,000 is a good statistical sample of what people are actually watching? No way. The Neilsons have been a joke since the 1970's. Even families with the box admit to just leaving the TV on as background noise and not minding what is playing at the time.

 

[Mr. B]

It might have been the Arbitron radio ratings, they still use diaries. Neilson used to do something similar. The term "sweeps" originates from the mass collection of those diaries.

 

[Crewman47]

As for "only 25,000 samples," you can get good results with that number, if the households are properly chosen and the system works the way it's supposed to. Not saying it does, but it can. I mean, the boxes are great at figuring out that the TV is on and what happens to be showing on the screen, but as for who, if anybody, in the house is watching it...that's a much trickier thing.

A long time ago, I participated in...something. I can't remember what, though, butt wasn't Nielsen. It required me to keep a log because there was no box. That was really time-consuming to do, and I can promise that the results in turned in weren't all that accurate.

There are approximately what -- 175 million households in the US, and you want to claim that 25,000 is a good statistical sample of what people are actually watching? No way. The Neilsons have been a joke since the 1970's. Even families with the box admit to just leaving the TV on as background noise and not minding what is playing at the time.

I was thinking that myself. What if for example you have Lost on one channel and something else popular on another channel and for one or two months straight more than 80% of those 25,000 decide to watch the other programme instead of Lost. Does this mean that Lost has to suffer and be cancelled, disappointing the X odd million viewers who do actually watch it just becuase the other programme got more viewers?

 

[John Picard]

^^ Many good programs have been canceled in the last 30 years because of the Neilson Ratings Data capturing. Disgruntled TV viewers have complained for years, but the networks claim it's the only accurate means of collecting data.

 

[Trekker4747]

It's again worth pointing out that only 5000 boxes exsist that compile the Neilsen ratings, the rest of the 'Nielsen Families' just compile diaries.

 

[Michael Chris]

Recently we were supposed to be a nielsen family, but it never came in the mail.

 

[JustKate]

I understand the hostility toward the Nielsens, but believe it or not, there is nothing wrong with the concept of using 25,000 households (or 5,000, if you want to stick to those that have the boxes). I realize this is an uncomfortable thought for many people, but yes, a properly constructed sample of 5,000 can indicate the preferences of 175 million with a fair degree of accuracy.

No, we individuals don't like to think that, and I expect some of you reading this are arguing with me in your heads right this minute, but it's nonetheless true.

However, the emphasis here is on a properly constructed sample, and I don't know that Nielsens' qualifies. But the problem, assuming there is one, is with the way the sample is constructed, not the size of the sample.

 

[zakfa]

I never knew Nielson used such a small sample size. It seems like you could not have a high degree of confidence in your ratings numbers with such a small sampling...

 

[JustKate]

^ But you can. If you do it right.

 

[John Picard]

^ But you can. If you do it right.

No way, 5,000 is just too small of a sample given the number of programs plus peoples' schedules these days.

 

[JustKate]

^ On what do you base that opinion, John? I'm not saying you're necessarily wrong - I'n not an expert in statistics and sampling - but I am saying that there is a very precise science to these things, and it's a lot more complicated than "That doesn't seem like enough samples to me."

It's notable, I think, that despite decades of criticism of the Nielsens, including from all the major networks, nobody's managed to supplant them yet. It could happen, but it hasn't happened yet.

 

[Trekker4747]

^ But you can. If you do it right.

No way, 5,000 is just too small of a sample given the number of programs plus peoples' schedules these days.

Drugs are tested with similar numbers.

 

[zakfa]

I've re-educated myself in statistics the last few minutes and it would seem 5000 well selected samples would be more than sufficient to draw reasonable conclusions. So yeah I guess I was wrong...

I've always had problems in the past with how Nielson didn't include college students but they started including them recently I believe. Certainly it's not perfect but it seems the networks trust the numbers enough.

 

[JustKate]

^ "Enough" is the operative word! They aren't crazy about them, but...right now, there doesn't seem to be anything better, at least not that advertisers will accept.

And if it's done right...it's highly effective. Whether we like to believe that or not.

 

[chrisspringob]

I understand the hostility toward the Nielsens, but believe it or not, there is nothing wrong with the concept of using 25,000 households (or 5,000, if you want to stick to those that have the boxes). I realize this is an uncomfortable thought for many people, but yes, a properly constructed sample of 5,000 can indicate the preferences of 175 million with a fair degree of accuracy.

That's correct. It doesn't matter how big the population is. It could be 175 million or 175 billion or 175 trillion. What matters in determining the accuracy is how big the *sample size* is (25,000 in this case), and the methodology of how they were selected.

People should read up on the methodology of opinion polling. If you understand why polls of California have sample sizes of 1000 people, while polls of Delaware also have sample sizes of 1000 people, then you understand what I'm getting at here.