Venusian

You know how you always know someone who... or know someone who knows someone who... well I dont know anyone or know anyone who knows anyone who is a neilson family. I dknow this is supposed to be secret but i think it'd be obvious when you go to someones house and there is a little box on the tv that they have to punch their code into when they watch tv. Does anyone know about how mnay Neilson houses there are supposed to be?

Well, I'm bored at work so I decided to rant...

Eternia

Two different times in the past oh, 10 years or so my family here have been sent a packet from Neilson. They asked us to fill out what we watched for a week. I remember clearly because one weekend TBS was having a Saved By The Bell marathon and I put it down every hour for like Saturday hehehe. I think they enclosed $1 too as a 'prize' for filling it out.

I'd love to be a regular or whatever with however they do it (a box on the tv I guess). Maybe then some of my favorite shows would have not received the axe...

renaldow

Yeah, I've done the Nielson diary too, but that was years ago. Now they don't have to do that so much. With digital cable boxes the cable co can tell exactly what someone watches and they sell the data to Nielson. Even some older boxes that are non-digital do that too.

That's how they can say things like '18 million people watched xxxxxxx last night.'

Venusian

this is actually what i was thinking. My main reason for asking is that could i be a neilson person without actually knowing it.

Scheherazade

Funny you mention this, one of my friends' families was just made a neilson family for last week. They got these cards or something they need to fill out for each TV and a dollar with each card to compensate them or something.

Apparently there's such a thing as neilsons for the internet too as I've known someone who had a box on their computer monitoring the sites they've visited.

~Scheherazade

McHawkson

But they don't know how MANY people actually viewed. If you got 20 TVs in your house and you're alone, they'll count it as 20 people viewed.

Scheherazade

It works like that though, I forget the exact numbers but it works something like 1 neilson viewer counts as 2000 people. Until they can monitor what we all watch, which I really don't see why they can't besides maybe it being cost prohibitive, then this is as good as it can get.

It really isn't all that fair, especially to shows that have loyal viewers as more of a cult following, but I imagine it's fairly accurate some of the time if not most.

~Scheherazade

milo bloom

I doubt it's highly accurate. It's all based on statistics. Like if you interview ten people on the street, and only one likes chocolate, then you could say that 90% of American's don't like chocolate. Now, statistically that's correct. But realistically, I seriously doubt it.

The problem is the sampling pool of only 10 people, you may have just gotten nine people walking out of a support group for people allergic to chocolate. However, if you up the sampling pool to 100, then you would have a percentage closer to reality. 1000, 10,000... you get the point. The problem is, there's not enough Neilson families. That's why stuff like Seinfeld stays on for 10 years (which admittedly has some funny bits, but not ten years worth) and an intelligently written comedy like Sportsnite lasts a year and half.

If you'll recall, the original Star Trek was cancelled twice due to low ratings. Granted Spock's Brain was a stinker, but the rest of the series *sorta* tried for cerebral storytelling.

Bobubott

I really don't think the problem is with the number of people who participate. I seem to remember from a statistics class that you need just a surprisingly low number of participants in order to accurately generalize about the entire population.

Here are two bigger problems regarding those who keep TV diaries:
1) People are lazy and have bad memories. They'll put off filling out their diaries for a day or two, and by then they can't remember what they watched, or who was watching.
2) People lie. People will write down what they think are the "right answers." They might have been watching Jerry Springer, but for fear of looking like inbred hicks, they'll write down that they were watching PBS and the Discovery Channel.

And I disagree... Seinfeld did have enough funny material to last its entire run.

das Monkey

I like Chocolate.

das

renaldow

Only if you have 20 cable boxes that are turned on. Yeah, it's innacurate, but it's close enough now days to get reliable ratings. You've got to think that Nielson has been doing this for years and have it down to a science. They aren't someone asking 10 people if they like chocolate. Are they infallible? Of course not. But they do give a good representation, at least good enough to extrapolate who likes the show and where in the country it's most popular.b

Venusian

How diversified are Neilson families? for example are they just randomly picked or randomly picked per area or what? If they are randomly spread out and California runs out of power maybe an unproportionally small number of neilson families will be out. I'm sure Neilson knows what they are doing by now, either that or they are fooling eveyrone pretty good.

Orson

What most people fail to realise is that networks and their advertisers are less interested in overall viewers than they are in demographic penetration. For example, Seinfeld didn't really set the ratings alight in it's first few seasons, but the networks and advertisers kept with it because the people who did watch were middle-class urbanites with large disposable incomes.

In other words, a bunch of trailer trash, poor-ass hillbillies can watch a show in their millions and it can still get cancelled. But if rich (white) people watch a show in just half the same numbers, it can be almost guaranteed a renewed contract.

Not saying it's right or wrong. Just saying it happens.

milo bloom

I didn't think of that Orson, very good point.

However, Bobubott, I seem to recall from my statistics classes that sampling "could" give you a fairly accurate view, but it's not sure thing. I also have to stand by my belief that there's not enough families in their sample pool.

I just wish there were some way to insure honest internet voting for shows. It'd be a lot easier, and sci-fi/fantasy might actually have a chance. Or we could end up with 24/7 Buffy vs 7 of 9 matchups

Orson

In the UK, Sky - our only digital satelite operator (5m households in a population of 50m) - is giving away free digital recievers and satellite dishes, on condition that you connect them to your phoneline.

At night-time, the digibox phones up Sky and tell them what you watched and for how long. (I live in a very poor, inner-city area, so I like to buck the trend and leave BBC News/CNN running for long periods of time when I'm out of the house. It also helps the ratings for these information channel, which might face the axe if they cannot justify their existence with advertisers.)
Interestingly, I believe the digibox also downloads software upgrades via the phoneline.

I'm still not sure whether this is an invasion of privacy, or a positive and healthy technological advance. I saw a recent news report saying that the law is trailing behind the technology, and is not prepared - in case the technology is used to monitor us too closely.

Bobubott

I'm not trying at argue that the Neilsens are accurate by any means. However, I stand by my belief that in terms of just the quantity of people surveyed, the Neilsens are surveying enough people.

milo bloom, to shed a little light on how random sampling and confidence intervals work, I did a little online searching. On one website, they said that in order to generalize to a population of 5000+ people with 95% confidence, you only need to randomly sample about 400 people (I think this is the figure I remember from my stats class, which we all had a hard time believing was true).

I found another website which you might find interesting... http://www.ibiblio.org/journalism/pollcomd.txt. In fact, they use ice cream flavors as their example. Here's a snippet from the webpage:

--------------------
3. Thou shalt know the poll's error margin. Here's how to calculate a poll's error margin using any calculator. 1. Enter the number .25. 2. Press the division key. 3. Enter the number of people interviewed for the poll. 4. Press the "equal" key. 5. Press the "square root" key. 6. Press the multiplication key. 7. Enter the number 1.96. 8. Press the "equal" key. 9. Press the multiplication key. 10. Enter the number 100. 11. Press the "equal" key. The resulting number is the poll's accuracy at the 95 percent level of confidence. For example, a result of 4.2 means that if 60 percent of the people interviewed for the survey said they preferred chocolate ice cream over vanilla, then one can be 95 percent confident that the actual percentage of people favoring chocolate over vanilla in the population the poll purports to represent lies somewhere between 55.8 (that's 60 minus 4.2) and 64.2 (which is 60 plus 4.2). What's a good error margin? It depends. An error margin of 4.2 will let you predict the winner of an election if the poll shows one of two candidates has 80 percent of the vote. Subtracting the error margin from 80 percent reveals that, at the very worst, the candidate has 75.8 percent of the vote and is clearly still ahead. If, however, the poll shows one of two candidates with 53 percent of the vote, then you can't predict a winner (without using some really hairy statistics that you'll need a computer and some fancy software to calculate) because the candidate could have anywhere from 57.2 percent of the vote to only 48.8 percent of the vote.
--------------------

(For reference, the only data I could find on how many people participate in Neilsen research estimated the number at around 13,000 people.)

So, it seems that in terms of quantity, the Neilsens are probably doing things right. As others have said, for a company that has been doing this for this long, I'm sure they know a thing or two about statistics and all that.

That being said about quantities, I have little confidence in the qualitative side of the research. I already mentioned the problems with diaries. As someone else mentioned, the problem with the set-top box is that it doesn't tell you how many people are watching, or if anyone is watching at all. And, with something like the Neilsens, it's questionable if you can ever really get a random sample. How are participants picked? If it's by randomly picking names from the phone book, that excludes people without phones or with unlisted numbers. If it's done by having Neilsen employees randomly walking up to people's houses and asking them to participate, it's less likely that people in rural areas or people in dangerous neighborhoods will be picked. You get the idea.

Internet voting would seem like a nice idea, but a few problems seem obvious. One is the possibility of hacking and deliberate tampering of the data. Another is that internet voting would skew the results toward younger, more tech-savvy people and wouldn't be representative of some older people who don't even know how to use their toaster.

I guess my point is this... I don't trust the Neilsens. It's not because of how many are studied, but because of how they are studied.

Hoc

OK, I'm breaking protocol a bit by telling you this. But I'm a Nielsen family. I don't have to keep a diary or anything -- my habits are recorded by computer and sent online to Nielsen.

Venusian

are you joking or are you serious?

Shonn

I just got done doing the week long diary for Nielson. They even gave me 2 dollars.

Venusian

what do you record if you flipped back and forth between two shows. or you channel surfed for half an hour?

Shonn

If I flipped back and forth, I would write down the show I liked more. If I was channel surfing, I would write down the show I watched for the most amount of time.

Hoc

I get a $50 savings bond 4 times a year.

Stats

The Neilsen company knows what it's doing. I've been to many professional conferences with technical presentations by Neilsen statisticians, on topics like:

* incentives: Hoc's $50 savings bonds or the $2 cash and their effects on honesty and response.

* sample size: ensuring enough families are in the sample to satisfy Neilsen's customers, the networks, who pay for the ratings.

* sample design: allocating their sample size among different types of families (young/old, White/Black/Hispanic/others, etc.) to obtain broad coverage of the population.

* sample weighting: comparing the demographics of their families to known demographic benchmarks, like the census, to accurately give a weight (e.g., 2000 people per Neilsen person) to each person or family in their sample.

* variance estimation: estimating the possible range of values in which the true viewership might lie, b/c all samples are subject to error.

I even know a few people who work(ed) at Neilsen and have nothing but respect for their background and commitment to the science of statistics. And yes, it's true that a survey requires a sample of only about 1500 people to accurately represent the entire population.

big whoppa

I filled out the surveys years ago. They enclosed $2 in with the booklets but I haven't received any since.

The $50 savings bonds are nice. I'd like to know how they determine to be regular Nielsen families.

Are cable companies allowed to share viewing habits with a company like Nielsen? That could be an invasion of privacy.

Eplicon

I participated in the Neilsen Ratings thingy myself not too long ago, although it was the diary format (i.e., writing down what I watched on what channel at a particular time).

I'm not sure how well I would have represented my viewing area or my age group, since I don't watch that much network programming. And I don't have cable, to boot. I watch a lot of newsmagazines (60 Minutes, 48 Hours, 20/20, PrimeTime, Dateline, etc.) and various other stuff, but not much else.

They paid me $5 (they included five crisp $1 bills in my packet!).