Fuzzy math

[GratefulCitizen]

I do not find it suprising that liberal academia eschews mathematics.

http://city-journal.org/2008/18_1_snd-ed_schools.html

Results in mathematics are rarely malleable and generally are uncooperative towards furthering political ends.

(I suppose political ends can be furthered by mathematics if it is used in the science and industry of weapons design and production. ;) )

[warrottjr]

Results in mathematics are rarely malleable and generally are uncooperative towards furthering political ends.

Lies, damned lies, and statistics.

[GratefulCitizen]

Lies, damned lies, and statistics.

You just hit on a major pet-peeve of mine: statistical survey sampling (polling).

Liberals would love to keep people ignorant of what inferences can reliably be drawn from these polls.
(or, more specifically, what inferences cannot be drawn)

Media outlets love to spout "margin of error".

However,
-they usually omit the actual questions used in the poll;
-the existence of a poll is reported after the fact, so there's no check against underreporting;
-to my knowledge, no media outlet has ever indicated the confidence level of a poll;
-furthermore, media outlets have not explained what a confidence level is nor the significance thereof.

They have also neglected to explain that projection errors like what happened in Florida during the 2000 Presidential election are guaranteed to occur.


Ignorance is the friend of the propagandist.

[warrottjr]

You just hit on a major pet-peeve of mine: statistical survey sampling (polling).

O.K. Let's do the math!

In digital signal processing there is a phenomenon known as "aliasing."

Basically, you can't measure a frequency more than half the sampling frequency. (For you purists, there's a "plus one" in there somewhere.)

So, when you sample something, you're only getting a taste.

So, let's switch to cooking!

If the flavors aren't properly "marianated" (as they say in N'Awlins), the sample is not representative.

[GratefulCitizen]

warrottjr:

After re-reading my own post, I realized it may have been unclear.

My response was in no way meant to be a disagreement or accusation.

Rather, it is an agreement with the theme of "lies, damn lies, and statistics."

Political ends can be furthered when the media misrepresents what statistical science can and cannot do.

Widespread knowledge of how the mathematics actually work would undermine many "bandwagon" techniques.

[warrottjr]

Political ends can be furthered when the media misrepresents what statistical science can and cannot do.

Widespread knowledge of how the mathematics actually work would undermine many "bandwagon" techniques.

Lt. Uhura, I thought you said there was no intelligent life on this planet.