Lying with statistics or the Aaron Brown Got Better Ratings meme
Aaron Brown's contract with CNN has officially expired and now he's talking. (here, here and here.) Some people - but not, to his credit, Mr. Brown himself - are using the opportunity to dust off the old "Aaron Brown got better ratings than Anderson Cooper gets now" urban legend. We were saddened to see even Rachel Sklar at Eat the Press, whom we usually adore, propagating this data atrocity (admittedly in a footnote), so we thought we would address it here.
People are getting these bogus figures by looking at the program ranker figures, which lists viewers by program, rather than the hourly figures. AC360 spans two hours, and replaced both NewsNight and a repeat of Lou Dobbs. BOTH hours outscore the shows they replaced, but because the 11PM hour is always lower, it brings down the average of the whole show, and thus makes it appear (to the stupid) that Anderson Cooper's numbers are lower than Aaron Brown's were.
Another 'how to lie with statistics' technique is to compare a datapoint to an average. Take Larry King's 'Paris' rating, for instance, and act like he gets that every night. Or the ratings from, say, right after 9/11. Yet another technique is to compare unlike months, say, November to May. People's viewing varies across the year; some months ALL ratings are lower and some months more people are home and watching. You have to compare like to like to get VALID comparisons.
Here's some real data, drawn from TVNewser's archives.
June '05
Aaron at 10PM - P2 623, Demo 208
Lou at 11PM - P2 314, Demo 114
June '06
AC360 at 10PM - P2 784, Demo 252
AC360 at 11PM - P2 481, Demo 191
June '07
AC360 at 10PM - P2 790, Demo 283
AC360 at 11PM - P2 488, Demo 205
Now please. We don't have a thing against Mr. Brown, who seems to be a nice and astute gentleman, but can we please put a rest to the Aaron's ratings were better than Anderson's shtick? It's simply not true.
People are getting these bogus figures by looking at the program ranker figures, which lists viewers by program, rather than the hourly figures. AC360 spans two hours, and replaced both NewsNight and a repeat of Lou Dobbs. BOTH hours outscore the shows they replaced, but because the 11PM hour is always lower, it brings down the average of the whole show, and thus makes it appear (to the stupid) that Anderson Cooper's numbers are lower than Aaron Brown's were.
Another 'how to lie with statistics' technique is to compare a datapoint to an average. Take Larry King's 'Paris' rating, for instance, and act like he gets that every night. Or the ratings from, say, right after 9/11. Yet another technique is to compare unlike months, say, November to May. People's viewing varies across the year; some months ALL ratings are lower and some months more people are home and watching. You have to compare like to like to get VALID comparisons.
Here's some real data, drawn from TVNewser's archives.
June '05
Aaron at 10PM - P2 623, Demo 208
Lou at 11PM - P2 314, Demo 114
June '06
AC360 at 10PM - P2 784, Demo 252
AC360 at 11PM - P2 481, Demo 191
June '07
AC360 at 10PM - P2 790, Demo 283
AC360 at 11PM - P2 488, Demo 205
Now please. We don't have a thing against Mr. Brown, who seems to be a nice and astute gentleman, but can we please put a rest to the Aaron's ratings were better than Anderson's shtick? It's simply not true.
posted by The Staff @ Unanchored at
8:39 AM
<< Home