Here is another one of her papers that explains how POLYGRAPHS are instrumental in the containment approach.

http://dcj.state.co.us/ors/pdf/docs/eoc51.pdf

Oh yeah, here is a snippet from her bio:
Ms. English has been the principal investigator on a number of studies funded by the National Institute of Justice, including two national surveys of probation and parole management practices pertaining to adult sexual offenders. One of the research products, “Managing Adult Sex Offenders in Community Settings: A Containment Approach,” was published in 1996 by the American Probation and Parole Association. Current research includes assessing the impact of sex crime disclosures by offenders participating in specialized post-conviction polygraph exams.

Wonder_Woman wrote on Oct 4^th, 2007 at 10:50pm:
 

Ah, thanks for bringing it back to my expertise...

As I pointed out in prior posts and to Kim several years ago, accuracy in diagnostic testing has 5 statistics that are needed to measure it. So these comparisons are incomplete and do not show the whole picture.

Once again, these are the five things needed to assess "accuracy":

1. Sensitivity, which is the probability that the test is positive given that the person has the underlying condition.
2. Specificity, which is the probability that the test is negative given that the person does not have the underlying condition.
3. Positive predictive value (PPV), which is the probability that the person has the condition given that the test is positive. Its complement is the false positive rate.
4. Negative predictive value (NPV), which is the probability that the person does not have the condition given that the test is negative. Its complement is the false negative rate.
5. Base rate, which is the prevalence of the condition in the population or the probability that you've got someone with the condition. This is also the "chance" of just guessing correctly whether the person has the condition or not.

Sensitive and specificity are usually measured in lab studies. 

The base rate, at least for diseases, is estimated from medical records. 

PPV and NPV are both functions of sensitivity, specificity, and the base rate. These measure how the diagnostic tests perform when the true status of the underlying condition is unknown, which is the case in screening applications. 

Basically, they are a measure of how well the test performs when compared against a gold standard. Any PPV or NPV less than 90% has no usefulness in my opinion. 

Additionally, when the base rate is low (deception in employment screening), your PPV goes down and your false positive rate goes up. Conversely, when your base rate is high (deception in sex offending), your NPV goes down and your false negative rate goes up.

And without showing PPV, NPV, and base rates of the conditions those test measure, the numbers above are incredibly misleading and completely without context...

Additionally, unlike like the polygraph, all of those medical listed tests can be run sequentially or in concert to reduce the probability of error. Once a person has been polygraphed, because of the nature of the test there is a high likelihood of prior tests affecting future tests...

Wonder_Woman wrote on Oct 4^th, 2007 at 10:50pm:


I'll leave it up to others to determine the strength of your argument.

Wonder_Woman wrote on Oct 4^th, 2007 at 10:50pm:

No need to ask, her study does use it...

Did you just make that up, or have been readin' the NRC report again?

Digit,

You must know that its not quite that simple, and that you do have some options about limiting ourselves to bayesian models. There are a lot of tests built on good 'ole inferentials. Take IQ tests, for example: is there a base rate for IQ?

The point is: when you limit things like this you are really engaging in a straw man argument, not a real discussion.

Certainly there is much to learn, we simply have to open the mind to do it.

and this

Quote:


OK, maybe not ad hominem, but certainly straw man - because anyone who would be a true believer is certainly a fool. Right?

Funny, I thought Kim English's data indicated polygraph does contribute to the containment process.

'till then

Quote:


You sir, are engaging in misdirection, when you lose sight of the fact that tests are just tests, and focus only on the voices of the idiots.

Intelligence, like cholesterol, is a normally occurring phenomena (though amorphous). Telling lies is also a normal occurence for humans, as is telling the truth at times. The questions that tests seek to answer are things like how much is too much (high blood pressure, for example), and what is normal (both blood pressure and IQ). Other questions, of interest to things like mental health test measures, include more specific questions like what does the test protocol of a depressed person look like, or what does the protocol of a narcissistic person look like? Similarly, what does the test data of a deceptive or truthful person look like? It is the role of bayesian and inferential mathematics, and signal detection models, to provide probability estimates as answers to those questions.

Its just testing.

EOS

You have lost it my friend. Welcome to the Twilight Zone.

Regards,

MM

Paradiddle, MM,

Again we shall let the audience decide ....

Mind theatre ass kicking !!! Just another quality service I offer ....

Regards ....

Eos,

Could you please send a copy of the CM trainer and the necessary key to:

4thetruth@swarmail.com

Thanks!

Nonombre...