Happy Thanksgiving! We’re very thankful to our readers. Oh sure, sometimes we realllllly wish people would focus on the ratings for shows like Chuck and Terminator: The Sarah Connor Chronicles instead of stuff like whether or not Summer Glau is a good actress, or who’s hotter, Summer Glau or Yvonne Strahovski? From my perspective that’s merely a “name two women who won’t be coming over to my house for Thanksgiving” situation, and so the ratings for the shows are more interesting to me.
But there’s a lot of good chatter about ratings as well and Bill and I have certainly learned a lot over the last year and no longer feel that we’re complete newbs when it comes to ratings analysis. A lot of what we’ve learned has been from our readers, so thank you!
Being that it is Thanksgiving and so long as I wasn’t spending the day with Yvonne or Summer, I thought I’d take a few minutes to focus on NUMBERS (since this is TV by the Numbers, not TV by the Who’s Hotter?)
We recently got a comment that suggested that because the ratings for both the 18-34 and 18-49 demographics in the overnight report for House were both similar that you might infer from the data that there were more people 18-34 watching than people 35-49. I understand how the mind can wrap itself around it that way and it’s one of the reasons I wish Nielsen didn’t use ratings points for anything and just stuck with the number of viewers. A ratings point simply means 1% of whatever population is being measured and there are lots of populations Nielsen measures.
Two populations we focus a lot on are the 18-49 year old and 18-34 year old demographics. And at least in our overnight reports the performance in those demos is always reflected in terms of ratings points.
For example, in the overnight report for Tuesday’s numbers (which did change more than a little bit in the final numbers) we showed House with a 5.1 rating among 18-49 year old and a 4.9 rating among 18-34 year olds. But does that mean that House had more viewers under 35, than between 35-49? While we didn’t get enough significant digits to tell you precisely how many viewers there were in each category, we have enough information to find out if it’s really true that more of House’s 18-49 year old viewers are under 35.
The table below shows the total television household populations in each of the age demographics according to Nielsen for the 2008-2009 viewing season. I’ve also broken them out by gender just because the table looked so bare with only one row of data! As you can see, the total population of 18-34 year olds is 68.17 million and the population of 18-49 year olds is 131.97 million.
| Demographic | 18-34 (000) | 18-49 (000) |
| Men | 34,270 | 65,680 |
| Women | 33,900 | 66,290 |
| Total | 68,170 | 131,970 |
So how do House’s demos shake out? In the example used – and I’m sticking with the overnight preliminary data since I lack the “final” data for the 18-34 demo – in round numbers it’s like this:
- 18-49 year old viewers: 6.73 million (.051*131.97 million)
- 18-34 year old viewers: 3.34 million (.049*68.17 million)
So it turns out that it is not true that most of House’s 18-49 year old viewers are under 35. In fact, in this example slightly more viewers were 35-49 than 18-34. I understand how the mistake gets made comparing the demo ratings, but the problem there is while the ratings numbers were very similar, the populations they measure are very different.
I am definitely not picking on the person who made the comment. We appreciate it — it does focus on numbers and it gives us a chance to clear up how things work.
Considering it was me that made this point originally, I will just add myself to this.
While I appreciate these numbers and in fact they are correct in addressing my mistake, I still believe that my point stands.
NCIS does skew older and this is shown by the massive jump in the demos.
HOUSE 18 – 34: 3.34 million
HOUSE 35 – 39: 3.39 million
HOUSE older (or younger): 5.76 million
NCIS 18 – 34: 1.43 million
NCIS 35 – 49: 3.45 million
NCIS older (or younger) : 13.28 million
So assuming that most of the non demographic are older, then nearly 17 million of NCIS’s viewers are over 34, as opposed to just over 9 for House.
So in fact, NCIS does have more of the older viewer
Tony, I was only addressing your comment regarding House, because we had enough data to respond to that. It seems very likely to me that most of the NCIS viewing is 50+, but lacking the 50+ and under 18 data, I can’t state that definitively. Sure, given NCIS’ 18-34 numbers, the chances that it is doing gangbusters with the under 18 crowd is zero — but, I still lack the data.
As a percentage, the 50 and over viewing for almost every show (including House) is quite large, but it’s easy (and correct) to assume NCIS skews older than House.
Robert, thanks for the post. I’m sure most of your readers didn’t need the tutorial, but it’ll certainly help me make more since of the numbers. Do you publish the Nielsen demographic numbers prior to the start of each season, or is there another site from which they can be obtained.
Maybe it’s just wishful thinking on my part, but I expect T:SCC’s numbers to rebound next week. No DWTS, CBS is airing repeats, and the new episode is stand-aloneish. So, I think 5.2M is a fair expectation. Agree?
ljo, I think I did publish some of the “Nielsen Universe” estimates somewhere before the start of the season, but it would probably be something good to make more visible on the site or at least link to them via “numbers 101″. At some point soon I’ll do a more comprehensive post on the Nielsen estimates and make them more visible. Thanks.
The reason they use ratings points is actually quite simple – because it’s easier to compare success in different targets. If you reach 5.1 in the 18-49 but only 4.9 in the 18-34, you know that you have been slightly more successful in the 18-49. Because, you know, 5.1 is more than 4.9.
If it were the other way around, and you just knew that you had 6.7 million in the 18-49 and 3.3 million in the 18-34, that’s hard to compare because, obviously it’s easier to accumulate viewers the larger the population.
In fact, in Europe, share is used much more frequently than ratings (which is likely to add another layer of confusion to some) to account not just for target group size, but also for overall target group viewing.
Corey — No DWTS, CBS repeats and I’m not sure the Monday Night Game (Jacksonville/Houston) will be one of the bigger MNF games. If TSCC is not over 5 million, that would be very bad.
I’m not very interested in analyzing the TSCC ratings at this point. It seems likely to get its full season two run regardless of the ratings. It also seems very unlikely it will get a third season, but I don’t see the point in worrying about that until March or April.
Ray, that’s a great point about it being easier to compare relative success in different targets using ratings, thanks. We report the share data, and in some instances it’s helpful additional data (in the case of House vs. NCIS, for example).
I have to say it’s not even close Yvonne Strahovski is much hotter than Summer Glau case closed oh and Chuck is a better show than TSCC as well
im surprised that the biggest key-measure for a show isn’t the 14-34 measure. I’m pretty sure that High Schooler’s ratings would mean more than someone who is in college or starting a new life out of college since High Schoolers rely on their parents to buy things from advertisers. I’m 16 and I don’t understand why it seems my view of a show doesn’t really count. I have tons and tons of friends who watch Heroes and haven’t stopped watching since season 1 (between 14-18 ages) yet the ratings suggest that over 30% of the viewers and ratings have dropped. I haven’t seen that. I don’t understand how Nielson can say absolute numbers without having an absolute measuring system.
Joseph–you are right that Yvonne is hotter than Summer. But Summer is a true scifi genre actress and Yvonne is not. And in the scifi genre, Summer is clearly in the top 10 in hotness. It’s all relative, my friend.
One thing I think would be neat to study is the relative success of the “lead-in” among different age brackets. I know some people around here are skeptical of the value of a lead-in and, more broadly, of the value of clustering high quality shows on the same nights in order to generate spillover effects and maximize viewer returns. But I wonder if the effectiveness of lead-ins varies by age bracket. Now that we know that the networks are focusing more on the 18-49 and 18-34 brackets, I for one would be interested to see if lead-ins are more or less important among those brackets.
I want to make sure I’m understanding this, so can someone tell me if I’m calculating this right?
I’m using November 3 Gossip Girl numbers because there was a press release listed for them, and I could find a gender breakdown.
If total viewers was 3.16 million, and it scored a 3.9 in w 18-34, then I can multiply 33,900,000 by 3.9% and find out that there were 1.32 million viewers in that demographic. Almost 42% of total veiwers fall into this category
If w 18-49 was a 2.6 then there were 1.72 million viewers in demo. 400,000 female viewers were then between 35-49.
Aline, your calculations are correct. Assuming your original numbers are correct (I didn’t check them), it’s all correct.
Thanks Bill!
Joseph:
1) Reread the first paragraph of this article.
2) Kindly take your unwanted opinions elsewhere.
I see why Robert would rather focus on ratings than stuff like this. It’s easier to discuss facts than argue over opinions.
Here’s a different number problem I have begun figuring on: when you add up all the 18-49 and 18-34 for all four nets hour by hour for each day, they ( esp. 18-34 ) drop off, on some nights significantly, after the seven to eight central hour. They fluctuate differently each day from nine to ten – some days they continue their decline, other days, oddly enough, they return to higher than eight to nine but never as high again as seven to eight. Something – out to movies, to dinner, dinner at home, miniatur golf, cable, movie rental? – lowers the numbers after eight central on a very regular and significant basis. I only noticed this a week ago and bc I often save weekly ratings to see trends I may do an exact number breakdown to see any patterns. What it seems to say first is that the seven to eight so-called family hour is actually the best time to catch 18-34 and 18-49 viewers, who after eight are doing something else on most nights of the week until after prime time. I would guess GM, P and G, MacDonalds, Whirlpool etc would favor this time period.
Imagine strahovski playing a terminator LMFAO! That would be hilarious and I don’t mean that as a compliment. I can easily see Glau or any other actress for that matter playing her role on Chuck. Glau all the way here.
And yes, I read the first paragraph.
Looks like there are more visitors who are T:SCC fans than of Chuck here, yet it doesn’t project the ratio to real ratings numbers ratio. Shame.