![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
I wish I had interactive tv
May. 2nd, 2007 09:15 pmI really wanted to slap the announcer upside the head, and ask him for the missing details in his report
It was a simple enough thing. First he had a pie chart showing that at some point, 25% said they had something and 42% said they didn't have it. The unlabeled gray segment? About 33%? No clue given.
Then he had the next pie chart, with 23% now saying that they had whatever it is, and 52% saying that they don't have it. No mention made of the gray 25% (by my calculation).
Instead, the focus of the announcer's talk was on the incredible growth in those who don't have it. Yappity-yap.
Not once did he provide a clue about how big the sample was. Nor about why the unlabeled grey folks had shifted from grey to announcing their partisanship. Nor any explanation for why a yes/no question had one third abstaining in the first place, and one fourth in the second.
I did think that maybe I'd challenge some of my students to determine how big the sample needs to be to make the shift significant. E.g., a small sample may mean that the 10% growth is within the range of noise - when does this growth actually become significant?
Actually, the more I think about it, this would be a good exercise for statistics students. But it certainly isn't the basis for a news report!
It was a simple enough thing. First he had a pie chart showing that at some point, 25% said they had something and 42% said they didn't have it. The unlabeled gray segment? About 33%? No clue given.
Then he had the next pie chart, with 23% now saying that they had whatever it is, and 52% saying that they don't have it. No mention made of the gray 25% (by my calculation).
Instead, the focus of the announcer's talk was on the incredible growth in those who don't have it. Yappity-yap.
Not once did he provide a clue about how big the sample was. Nor about why the unlabeled grey folks had shifted from grey to announcing their partisanship. Nor any explanation for why a yes/no question had one third abstaining in the first place, and one fourth in the second.
I did think that maybe I'd challenge some of my students to determine how big the sample needs to be to make the shift significant. E.g., a small sample may mean that the 10% growth is within the range of noise - when does this growth actually become significant?
Actually, the more I think about it, this would be a good exercise for statistics students. But it certainly isn't the basis for a news report!
