*promote

Promoting this video and sending it back into the queue for one more try; last queued Thursday, July 21st, 2011 8:26am PDT - promote requested by dystopianfuturetoday.

*fear

Adding video to channels (Fear) - requested by NetRunner.

The "half of schools in district are below average" probably makes sense. Most likely the average in question is a national or state average. If you expect that your district's schools are above (a state or national) average, finding out that half of them aren't would be news, and while the headline could have been clearer that's a journalistic (rather than a math) failing.

Even if that isn't the case, and the average was for the district, it's not a tautology when considering how schools are likely to be measured (ie. test scores, rather than by ranking or something). Suppose you have 5 schools, and their "average tests scores" are (40%,40%,40%,90%,90%). In this case, the "average school" would have a score of 60% and 3 of the 5 schools are, thus, below average. This is not terribly useful information, but it's not nonsense - and the percent of schools that were below average could reasonably expected to vary. It even tells you something about the underlying data that may be interesting. If only 20% of schools are below average, you know that they're likely far below average (and thus likely legitimately require attention).

In order to have a silly tautology, you'd have to say something like "5 of 10 schools scored below the median" (or something similar) - and I'd be very surprised to find that kind of headline.

>> ^jmzero:

Suppose you have 5 schools, and their "average tests scores" are (40%,40%,40%,90%,90%). In this case, the "average school" would have a score of 60% and 3 of the 5 schools are, thus, below average.


I heart statistics. Superintendent Chalmers would sandbag one school in the district so that all the other schools are above average.

I think we fear the Metric system more.

I was there in the front row. Jealous?

Since the average is always made up from 100%
Half of the schools will always be 50% (or depending on the numbers, very close).

And the "Better" the schools are doing, the margin gets slimmer that it could be less than 50%

Some schools would have to be doing VERY bad indeed if you wanted to say "60% of the schools are ABOVE average."

>> ^Sagemind:

Since the average is always made up from 100%
Half of the schools will always be 50% (or depending on the numbers, very close).
And the "Better" the schools are doing, the margin gets slimmer that it could be less than 50%
Some schools would have to be doing VERY bad indeed if you wanted to say "60% of the schools are ABOVE average."


I could have 99 students in a class getting 90% and 1 student getting 89%, and I can say that almost all (but not all) students are above class average. The same could also be said if the two percentages I used were 10% and 9% respectively.

>> ^sadicious:

>> ^Sagemind:
Since the average is always made up from 100%
Half of the schools will always be 50% (or depending on the numbers, very close).
And the "Better" the schools are doing, the margin gets slimmer that it could be less than 50%
Some schools would have to be doing VERY bad indeed if you wanted to say "60% of the schools are ABOVE average."

I could have 99 students in a class getting 90% and 1 student getting 89%, and I can say that almost all (but not all) students are above class average. The same could also be said if the two percentages I used were 10% and 9% respectively.


There are three kinds of average: mean, median, and mode.
If he means median or assumes a normal distribution, he's fine.

Every time I see this guy's name, I expect 3 things: a boxer, chicken and Degrassi Junior High, which I've never seen.

to be fair: "-1" is very very rarely used in germany. usually it's "UG" for "Untergeschoss" (and continuing 1. UG, 2. UG, etc.), which is the exact same as "basement", "subbasement" etc.

Any way you look at it - statistics are used to lie and can never be trusted.
http://en.wikipedia.org/wiki/How_to_Lie_with_Statistics