Our president will keep you safe from terrorists: just keep watching “Dancing with the stars”.

LOL! Well the season finale was last night (which is why I couldn’t really watch the Frontline episode), so now Bush is screwed

The house TV set was also occupied by Dancing with the Stars finale last night. 

One of the two homeowners (namely not SEB nor Mrs. SEB) had control of said TV last night:-)

We live with too many people, lol.  (Yes, and only one tv set connected to the cable available.)

Time to find the program via the net.

Help me out here.  I thought data mining was searching for information WITHOUT identifiers attached.  For example, at one point the CDC was talking about mining drugstores and supermarket pharmacies for sales of OTC cold medications, to track flu symptoms.  If the WH is talking about something like that, I have no problem with it.  If, however, they know WHO is buying aspirin, then that’s another story.
SG

SG: You are normally right. Data mining has nothing to do with looking at names and other sensitive information. Because in Statistics you wouldn’t want to introduce any bias’. But what is the WH going to look for? Their data mining has to consist of looking at names and other sensitive information because they want to know what the likelihood that someone named “Webs”, who spends too much time on SEB and other blogs, has his own blog, and every once in awhile rants about how he hates Bush, is a terrorist.

When the NSA is capturing AT&T;voice and data network packets, I don’t think they care if you use Vick’s or Tylenol.  Which is one reason why this wiretapping case pisses me off.

But if I can get back to the data mining, there are a couple problems with it from a statistical standpoint.

Please note, from this point on I am making the assumption that anyone still reading this comment has a basic understanding of stats and the theory of the Normal Distribution. If not, please read the link above and possibly some of the other posts that blogger has in the series.

For one it’s really hard to not commit a Type I or Type II error when data mining. This topic can be confusing, but basically the larger the sample the better right? Well yes, but there comes a point when your sample size becomes so large you start getting closer and closer to the mean. For anyone interested in a visual representation… for now don’t concern yourself with the pink graph. If you look at the green and red graphs, the green graph would represent a normal distribution or a normal bell curve. The red graph represents what happens when your “n” or sample size becomes too large (think over 1000 generally or too high of a percentage of the population). As you keep drawing more and more samples from the population they will tend to be closer and closer to the mean. When this happens you will be more likely to reject your null hypothesis when you shouldn’t.

In plain English, or in the case of the WH wiretapping, this means the WH will be more likely to flag someone as a terrorist when they shouldn’t have. If you look at the positive number 2 (this represents the number of standard deviations from the mean), you will see in the green graph you can still have individuals from your sample that fit under that graph at 2 standard deviations, but not under the red one. This is why we need officials running this country that understand math and science.

Yes, it is important to be aware of what’s happening. But once you know, what can you do? Elect more politicians? Can you trust them anymore than you trust the current ones?

When you find out that the Bush government was behind the 911 attacks, what is the correct action to take?

When you find out that the Bush government was behind the 911 attacks, what is the correct action to take?

I hope you mean that as a rhetorical example.  We’ve covered that ground before - exhaustively - and it isn’t on our list of places to go again.

Hobbes before you post anything else, please click here. And read through the comments.

Or starting about here and go through the comments, which go quite extensively into the 9/11 thang.

Frontline: Spying on the Home Front (PBS.org)

Windows Media/Quicktime only.

It’s in five parts for your online viewing pleasure. Hope this helps out you guys who missed the episode.

They have 64 episodes in their Frontline Video Archive, including many that deal specifically with the
“War on Terror” and associated issues.