(04-10-2021 10:24 AM)Memphis Yankee Wrote:  

(04-10-2021 07:08 AM)slhNavy91 Wrote:  

(04-10-2021 01:06 AM)jedclampett Wrote:  

(04-10-2021 12:08 AM)Memphis Yankee Wrote:  

(04-07-2021 10:14 PM)jedclampett Wrote:  True, the data could be described that way, but -

It would be equally accurate to summarize the findings this way:

Appalachian State had as many - or more - 800K+ viewer games than 9 of the 11 AAC teams.

But let's do the comparisons based on the viewership averages (arithmetic means and medians):

.



Not according to my computations[/size][/color], based on the data reported at
https://www.sportsmediawatch.com/college...v-ratings/

Cincy:......2.06 million (data available for 8 games)
Navy:.......1.58 million (data available for 6 games)
UCF:........1.29 million (data available for 8 games)
Memphis:..1.08 million (data available for 6 games)
Houston:...1.02 million (data available for 5 games)
Tulsa:.......1.01 million (data available for 8 games)
USF:.........0.68 million (data available for 3 games)
Tulane:.....0.64 million (data available for 6 games)
ECU:........0.58 million (data available for 3 games)
SMU:........0.56 million (data available for 3 games)
Temple.....No viewership data (therefore, not included in computations)

Average per school in 2020 (computed as the median & arithmetic mean of the mean viewership of the AAC teams):

Median:*.................1.05 million (mid-point; 50th percentile)

Arithmetic mean:^....1.06 million

*The median of the mean viewerships of the AAC teams.

^The arithmetic mean of the mean viewerships of the AAC teams.

.

The only way to obtain an arithmetic mean of 1.257 million viewers would be to divide the grand total of all reported 2020 AAC FB viewership by the total number of AAC schools.

However, that wouldn't have been an appropriate way to compute the mean, due to a violation of the requirement and statistical assumption that each school had data available from an equal number of games.

The fact that some schools had data from 8 games, while others had data from only 3 or 6 games represents what is known as a statistical confound. Another term for this type of problem is statistical "incommensurability".

A "mean" based on unequal numbers of games would be "meaningless," since it would be based on an essentially "apples to oranges" type of comparison, resulting in an estimate that should be assumed to have inadvertently over-weighted the data from teams that had data reported for 8 games.

.


That is true. For example, UCF's average viewership (1.29 million), based on the arithmetic mean was higher than Appalachian State's highest reported viewership (1.25 million).

However, I did not attempt to make the claim that App State's viewership was higher than the viewership of the top 3 AAC teams.

What I wrote was that, based on the number of 800,000+ viewership games, App State's viewership sufficient to match that of most AAC teams. The data based on the number of 800,000+ viewership games supported that statement.

But let's proceed to take your suggestion to a deeper dive into the data and conduct a more comprehensive examination.



My computations indicate that this statement is inaccurate, with respect to both mean and median viewership

Appalachian State's mean viewership of 785,000 viewers per game in 2020 was greater than that of five AAC teams: SMU, Tulane, USF, ECU, and Temple.

Median viewership for each team (mid-point; 50th percentile):*

#1 UCF:........1,274,000 (data available for 8 games)
#2 Cincy:......1,200,000 (data available for 8 games)
#3 Navy:.......1,150,000 (data available for 6 games)
#4 Houston:...1,160,000 (data available for 5 games)
#5 Memphis:..758,468 (data available for 6 games)
#6 App St......742,000 (data available for 6 games)
#7 USF:.........561,000 (data available for 3 games)
#8 Tulsa:.......546,500 (data available for 8 games)
#9 Tulane:.....524,000 (data available for 6 games)
#10 ECU:........525,500 (data available for 4 games)
#11 SMU:........400,500 (data available for 3 games)
......Temple.....No viewership data (therefore, not included in computations)

Appalachian State's median viewership, based on available data, was higher than (#24) Tulsa's viewership, and nearly as high as Memphis's viewership in 2020.

*NOTE: If a team has data for 3 games, the median is the viewership for the game that had the second-highest viewership. If a team has data for 8 games, the reported median is the mid-point of the games with the 4th and 5th highest viewership for that team.


All of us on the board except the George State and GSU fans seem to agree on that point, but the inference that that App State 2020 football viewership of 750,000 per game would have been a significant "drag" on the conference's metrics is not justified by the reported viewership data.

With a median viewership exceeding that of (#24) Tulsa and nearly matching that of Memphis, Appalachian State's viewership can not accurately be described as having been less than that a middle of the pack AAC team.

.

I'm confused by these numbers. Memphis has been near or at the top of the conference for viewership. In one year their numbers are this bad? We were 9-3. These numbers seemed framed to a narrative for app state. Somethings missing.

If you click the link above, you'll find these numbers:

Bowl: Memphis vs. FAU: 1.27M
12/12: HOU-MEM: 235K
10/31: MEM-CIN: 679K
10/17: UCF-MEM: 1.90M
10/3: MEM-SMU: 315K
9/5: ARK ST-MEM: 1.03 M

I've rechecked the computation of Memphis's median viewership across the 6 games, and there was a glitch with the calculator. Their median viewership was 854,500 (not 758,468) in 2020-21.

.

Memphis had much higher viewership in 2019:

Cotton Bowl: 6.22 M

AAC Championship: 2.88 M

CIN-MEM: 2.51 M

MEM-HOU: 292 K

SMU-MEM 2.95 M ...etc. etc.

.

Hey, Memphis Yankee - you really should check two posts after the jedpost to which you replied. https://csnbbs.com/thread-919232-post-17...id17373637
Lot of corrections. Medians were wildly wrong, as was Jed's entering argument of Appalachian State's mean and median viewership. Jed never acknowledged post #367, but he never does say "thank you" when I help him with corrections. Even here where he says he had a glitch, he doesn't even get all the numbers right (HOU-MEM was 239k viewers).

Anyway, back to post #367...Memphis was fifth in the AAC in total viewers, looking at either regular season only or including bowls. Memphis was fourth in the league in average viewers per rated game before the bowls. Memphis was sixth in the conference in average viewers per rated game including the six bowls. That 2020 median viewership number Jed has corrected was fourth best in the AAC.

The games were wildly inconsistent. The Houston/Memphis game only had 292K? Wow. I realize it was a Covid season, but these numbers are all over the place.

Started to reply to you earlier then saw Navy had it covered. Bless his heart meant in the sincerest way possible. I ain’t got time for the messiness. The Houston game was 2020 in a nutshell. Moved multiple times not only in date, but in time and platform. It was an ESPN+ game until a couple of days before kickoff with a time change as well. The game also overlapped with the Tiger/Auburn bball game that got moved so not to interfere with their football game that night. IIRC, it was originally supposed to be an 8 pm tip. Not sure on the exact time just recall hearing on Alabama radio that Auburn moved it due to the fball game. Long story short 2020 was weird and nothing predictive can be determined by one year’s data set, let alone one with so many *****

Using a valid data set (years) shows that teams performing the best get better slots and therefore better ratings. There are some variances within, but by and large it holds true. There’s no brand outside Navy that’s gonna get preferential placement on ESPN networks regardless of record...and even that matters just a bit less than the rest if that makes sense.