Numbers 101: Why Your Favorite Shows' Ratings Get Adjusted Up & Down in the Final Ratings

Categories: Help,TV Ratings Reference

Written By

December 15th, 2012

We get this question a lot: what's the difference between the preliminary numbers you post ~11:30a ET every day and the final numbers that are usually posted ~5 hours later? Why are their differences in the ratings between the preliminary and final ratings and why do some shows get adjusted when others don't?

Good question!

The preliminary ratings we post (the "Nielsen Fast National Ratings" in Nielsen jargon) just measure whatever aired on local affiliates in the time period. Most of the time that works pretty well.  But there are a few fairly common situations that increase the likelihood of adjustments to the preliminary Fast National Ratings in the National Final Ratings:

1. Live events: the Fast National ratings only measure what was on in the local time period. So if something airs coast-to-coast at 8pm ET and 5pm PT (like a sporting event) the 5pm viewing on the west coast isn’t included in the fast nationals. It is added to the final program ratings though.

2. Local preemptions: the fast nationals measure whatever aired in the time period. So if NBA basketball or MLB baseball aired on the CW affiliate in Chicago, the basketball or baseball viewing *is* included in the fast national ratings but it is stripped out in the finals. When this happens, the numbers in the preliminary (Fast National) ratings are often inflated, and then appropriately "deflated" in the National Finals.

3. Overruns: the fast nationals are time period measures only, so if Modern Family runs from 9:00-9:31, in the fast nationals, the extra minute goes into the show airing at 9:30p. In the final national ratings that minute gets stripped out of whatever aired at 9:30 and is added into Modern Family where it belongs.

It's major local preemptions that usually cause the most false hope and confusion. To our consternation, despite posting red, bold-text, italicized disclaimers at the top of the preliminary (Fast National) ratings noting major preemptions and that there is a greater likelihood of adjustments, many fans completely ignore the warning.

So even when we disclaim that the Thursday Night Football game aired on the ABC affiliate in the Philadelphia market and to take the preliminary numbers with a grain of salt, there are still plenty of "Woo hoo, Scandal at season highs, baby!" comments on the preliminary post only to have Scandal adjusted down three tenths, to just below season highs with adults 18-49 in the finals.


  • Grey

    @ azzman

    I hear ya. I know it has a “big brother” element to it but I think it would be better/more accurate if every TV made came with a pre-installed chip (or something) that tracked what’s being watched. That way, in theory, every TV set (not just a certain household or group of households) would be counted. If one person in the house is watching “Masterpiece Theater” in one room while another person in the house is watching “Honey Booboo” they both get counted. *And* I’d bet the demo standard, 18-49 years old, for advertisers would be proven faulty as well. If you’re selling a better mouse trap, you’re selling a better mouse trap. Everyone will buy it not just the 18-49 year olds.

  • Percysowner

    On that, I agree with you and we did not always do that. But the practice of comparing the prelims to prior originals’ final ratings is what Nielsen does and what the networks themselves do too.

    As it usually does, “When in Rome” ultimately won out.

    Thanks for telling us this. I’m sure you’ve posted it before, but either I didn’t read the comments when you did or I was skimming to fast. It’s always driven me crazy, but I just figured it was what it was. Now that I know that you didn’t decide on your own to compare prelims to originals, but that you are following standard practice I completely understand.

  • Percysowner

    @Azzman First I think most fans of reality shows don’t consider them garbage. I watch a few, but mostly I watch scripted stuff. Still people like what they like and I’m certainly not so superior that the networks should bow to my tastes. Second, fan fervor and involvement do not ratings make. Believe me 3 of my favorite shows are on the CW and boy are the fans invested (read bat**** crazy at times). But realistically, we are proud and loud and few. Finally, I don’t someone making money off me without my knowledge and consent. Heck, I don’t want my viewing habits checked without my knowledge or consent either. Plus, the only way to get the numbers advertisers want is to be able to tell who is in the room watching the show at any time. They don’t care if 60 year old me watches something, but they do care if my 25 year old does. Outside of cameras watching what I do with my TV I can’t think of an accurate way for current technology to track who is watching as opposed to what is being watched.

  • Terryjackson1NYC

    I really like the example used at the end of this post (Scandal/Philly/ABC/Thursday). The reason why is silly, but it’s important. If a sporting event preempts a show in Philadelphia, it will have a much bigger impact on the preliminary ratings, since Philly is the fourth largest market in the nation (behind the obvious first three: New York, LA and Chicago). If sports pre-empt a show in, say, the Buffalo market, it won’t be as big of a deal because Buffalo is at 51 in the list of market size.

    So, if a sold-out game between an LA and a New York team preempts ANYTHING, that will amount to ~12% of the entire national average. If the shows being preempted by such a game are typically low-rated, you can see bumps of two or three (and maybe even much, MUCH more – maybe even five or six – depending on the records of the teams and how far into that sport’s season the airing is).

    At this time of the year, preliminary ratings on Mondays and Tuesdays are fun to read, but I do typically ignore all of the comments, since they aren’t based in reality, even for those shows that weren’t preempted. The only numbers you can really count on are which network won the night and, for the most part, if it’s toward the end of sports season (other than hockey), the network with the broadcast rights will win the night.

  • eridapo

    @ John K

    One thing not mentioned in this post is if, for example, there is an NBA game on a CW affiliate, and The Vampire Diaries airs at, 10pm for example, after the game finished (or any other titme outiside of when it was regularly scheduled, often can be at midnight), do those ratings (in this case The Vampire Diaries numbers at 10pm) count in the Finals? Or do the finals basically represent ratings only for markets where there was no pre-emption

    As long as it airs before 3AM on that day, it will be in the finals. If it airs after 3AM, the adjustment will only show up on the weekly reports.

  • eridapo

    “So even when we disclaim that the Thursday Night Football game aired on the ABC affiliate in the Philadelphia market and to take the preliminary numbers with a grain of salt, there are still plenty of “Woo hoo, Scandal at season highs, baby!” comments on the preliminary post only to have Scandal adjusted down three tenths, to just below season highs with adults 18-49 in the finals.”

    True… When people look at a gain of a 1/10 one week and a drop of a 1/10 the next, they are ignoring the statistical error built into the Nielsen sample. It could well have been that there was no change whatsoever and that the shows performance was steady from week to week.

    What people should be looking for is trends over a period of time (4 to 5 weeks given a 22 episode season). If the show hovers around a certain average, chances are the show is stable.

    ex… 2.1, 1.9, 2.0, 2.,0, 1.9 (Chances are the show is really a 2.0 for the five week period).

    If, however, you see a continue rise or fall each week during a 5 week period, you should either be rejoicing or worrying about the fate of your favorite show.

    ex… 2.1, 1.9, 1.8, 1.8, 1.6 (Chances are the show is losing viewers).

    For shows that have history (shows that have aired the previous season), you might want to compare the prior season average to each weeks ratings (but take into account the overall decline that might be occurring throughout television).

    Ex… Season 2 of show X had a rating of 2.3, but this season it is performing as follows;

    2.1, 2.0, 2.1, 2.1, 2.0…. When you compute the average, you see the show probably lost 2/10s of a point from the prior year. Is this reason to panic? Yes and no. If the entire Network (i.e. all other shows are down by similar numbers), there is no reason to panic. If, however, your show is the only one down, you panic….

    In stats, you normally want to recalculate your average when something significant happens that would cause your results to shift… An example of this is if your show shifts time slot or nights. You wouldn’t want to use the prior averages because the conditions are not the same… An example of this would be The Mentalist….

    The shift in the Mentalist from Thursday at 10PM to Sundays at 10PM impacted the shows viewers and demos if we were to compare them year over year. For the Mentalist at least it makes no sense to use The Mentalist numbers from last year to see how the show is doing this year. The better solution is to to re-compute its average using the current years data (take the first six shows of the season) and compare its performance to this average. When you do that for the Mentalist, you find that the show has been holding steady around the 1.8 or better demo for sometime.. The days when it has gone way below that are the days in which Football pushed the airing to nearly 11PM.

  • bruce

    the method of rating a show seems to need a change in the way it collects data. I am glad that Nielsen is proposing a change in that area because I believe that with computers , laptops etc, there is a large segment of viewers that are not counted

  • tjw


    I think you would be surprised at what data Nielsen actually collects. They actually do track the amount of television watched via computers, video game consoles, mobile devices, and the like. I don’t know if they track show-by-show numbers, but they definitely know how much television people watch on various devices (spoiler alert – it’s actually not that much when compared to old-fashioned television).

    Ultimately, though, those numbers don’t really matter, since they’re not what advertisers pay for. The only ratings that matter (and will continue to matter) are the ratings that advertisers care about, and for the time being, that’s the C+3 rating.

© 2015 Tribune Digital Ventures