posted 11-23-2006 03:20 PM         

I agree with you. I like owning and driving a high performance car. I could not tell you the first thing about the interior/mechanical workings! I just know how to make them operate correctly and often at speeds that violate most state laws! I am not a mechanic, just a "need for speed" practicioner!

I too have in-laws on the way over so I need to go and get my blood alcohol to an appropriate level.

Happy Thanksgiving to all of you guys!

posted 11-27-2006 03:39 PM            

now I'm Birkenstocks and ganga on the mountaintop.

go figure.

Actually drugs an altitude don't mix.

I've pulled a handful of idiot college students off some of the 14k-foot peaks here. Mountains create their own weather, and things can change quickly. On the hike down from this peak we were in snow again. I prefer snow to lightning, but it is more slippery.

My first trip to that mountain was Oct. '93 chasing a runnaway youth. I found him by goddam-luck in the dark, off trail, about 500 vertical feet below that summit, and off trail - jacketless, dehydrated, and voiceless. We had been searching since just before dusk, I had consumed all of three quarts of water, and four sets of flashlight batteries had frozen out (before I discovered lithium batteries). Thankfully, I had the foresight to instruct the folks at our 10k-foot basecamp at treeline to build a big fire so I could at least orient our way back to basecamp.

It generally gets a little risky being up on mountaintops after about noon or at the very latest 3:00 PM.

So, no birkenstocks, but a 6 dollar pair of sandals from wallyworld
http://www.raymondnelson.us/rn/chillin_and_cookin.jpg

By the way, I'm the one on the right.

Rasta-boy with the hair is my 14 year-old, and doesn't seem to know who Bob Marley is. You wouldn't know it by looking at him, but he's a bigger nerd than I. He attends an International Baccalaureate High School where he earns straight As. My home is a nintendo-free zone - including all variants. So he plays a mean game of chess. He also runs an interesting web discussion forum frequented by other teenagers who like to write computer games. For some damn reason, he thinks he want to grow up to study psychology.

r

-----------

On another matter,

I've had some more communication with Shawn, who has obtained additional information from Dr. Kircher.

quote:

---

It looks like that the Stoelting polygraph thread has died out but here are some comments that Dr. Kircher had regarding your criticism of the CPS:

I can’t believe we're being criticized for making measurments that are too precise. I tested 16 and 24 bit ADs and found that I had underestimated the error. I digitized a constant source voltage using a 16-bit AD and a 24-bit AD. The standard deviation of several thousand samples at 1000 Hz was 75 on the 16-bit AD and 360 on the 24-bit AD (the respective SD of differences between successive samples were slightly greater). The relative error for the 16-bit 75/65536 = .1% was greater than the error for the 24-bit 360/16777216 = .002%. If an observed change from the subject that exceeds the inherent noise of the electronics is considered reliable, then a 16-bit AD divides the range of 100 microSiemens into 100/.1 = 1000 reliable units when we need 10000 to measure to the nearest .01 microSiemens. The 16-bit AD is insufficient (without an offset and gain adjustment). A 24-bit AD provides 100/.002 = 50000 reliable units to cover the same range, which is more than adequate.

---

I am corrected.

I didn't mean to be so critical of the CPS system - I noticed the decimal error and slipped in to aggressive-examiner-know-it-all mode.

I feel honored to enjoy even indirect communication with the guru (Dr. Kircher).

I think I'm beginning to understand the ATD issues.

I have a couple more questions on this. First, is that how we calculate a coefficient of variation? I thought we divide the SD by the mean to achieve a dimensionless value. Anyway, I understand dividing by range, as that is a practical representation of precision. Second, the numbers at the top of the first screenshot you sent appear to be quantiles - a representation of the proportion of items in the data set equal to or smaller than a given value or observation (the most common quantiles are the quartiles 25%, 50%, and 75% - the 50th percentile is also the median). So, my question pertains to the fact that you indicated, unless I misunderstood, that examiners could compute OSS ratios from those values - but quantiles have different variability than the actual measurements, so couldn't we expect OSS ratios based on quantiles to be different than the ratios based on measurements?

Most people probably aren't interested, but some might like to know that there may be good reasons for the claims you've made.

Peace,

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)