Author
|
Topic: Bugs
|
stat Member
|
posted 12-03-2009 07:06 PM
The last thread posed an interesting topic with a host of discussion angles. At what point do we call off the test when a component either .....well....wait a minute. Like in Buster's previous post, say a component seems to be malfunctioning---the op word being "seems." To simply ask if you cease a test because a component is malfunctioning presupposes that we know for absolute certainty that it is the component and not some elderly persons eroded sweat glands or some other wild artifact. APA is pretty unambiguous regarding instrument funtionality, but the real world holds challenges that such rigidity loves to scoff. Some arguments could be made that the fact that say, the pneumo can be predictably useless on a morbidly obese person, or that during extreme cold months, an outdoor construction worker's skin will be so non-conductive that lotion simply disappears into their dermis like a wind through a chain link fense---shows us how un-uniform test subjects and their particulars can present. Is that a means for test termination? The tests are still scored, but with many zeroes, and a sense of anxiety sets in knowing that the charts may be scrutinized in a hearing or peer review. Having a bum component is an aweful disability----but is it a deal breaker in an investigation (non-PE)test? I suppose, but would any examiner run a test if at the last minute a pneumo were bugged/gliched? The testing lab can be just as critical as we all know. Ever run a test with hammering down the hall, or someone's nearby door keep opening and shutting? I watched a trophy video from a touring examiner at a conference that had a truck jake braking every minute or so in the background--by well thought out standards, wasn't that a deal breaker----the fact that the examiner knew he was testing subjects (it was his office)very near a highway. It is a real menace when we are at such mercy of our gear and environment, but such things as bugs and near deal-breakers should be freely discussed in this forum without fear of punitive recourse. The fine examiner who ran the test with his shoelaces untied in the previous post did a service to the profession by describing a shitty situation that presented itself and begged to be dealt with at a moment's notice. I honestly don't know what I would do in any given situation which would take me by surprise----though I like to pretend to have such steely foresight.[This message has been edited by stat (edited 12-03-2009).] IP: Logged |
sackett Moderator
|
posted 12-03-2009 09:43 PM
I like controversy, so I'll go first...The question really is, do we need 3 channels to divine the truth, or can it be accomplished with 2? I don't know about anyone else, but ASTM (E-52 2439) REQUIRES all three channels for a valid examination. I too, have had examinations wherein the tracings were less than optimal, but they existed. I have also started tests wherein a channel did not work, malfunctioned or broke. If I can't fix it; the test ends! I explain the misfortune of the situation and request they return to help prove their truthfulness. Purposefully, running an examination without an operational channel is, in my opinion, inappropriate (Sorry Buster) and potentially unethical and most certainly invalid... Fire in the hole!!! Jim
P.S. On a side note for discussion. What happens when you get in a hurry and ask questions out of their appropriate order. Does that not invalidate the test as well?
[This message has been edited by sackett (edited 12-03-2009).] IP: Logged |
Poly761 Member
|
posted 12-04-2009 03:23 AM
I could not adequately support an opinion regarding an exam I conducted with only two components as I was instructed three functioning components are required. I'm not familiar with any literature citing validity of exams using only 2of3 components. In viewing a chart tracing just prior to the start of a test, if I don't observe "average" tracings being produced in each component by the examinee I'll stop. In most cases I believe tracing problems can be distinguished between instrument malfunctions and those attributed to an examinee. If the problem is not identified and resolved, no test. In my experience, distortions/responses due to outside noises are recorded, noted and in most cases not scored during analysis. If the noise(s) cause continual problems, we wait for the noise to subside or reschedule. Obviously there are many situations presented during an exam that are not as easily resolved. For all situations, including those less than ordinary that surface during an exam, we hope our training & experience guides us to a sucessful conclusion. END..... IP: Logged |
Barry C Member
|
posted 12-04-2009 09:32 AM
Well, I've run discriminant analysis on several options, adding and removing features. The short version: polygraph is pretty good even if you remove an entire component from the mix. Unfortunately, when you remove EDA, you get the worst results, but they still aren't terrible. Would I suggest you skip a component? No. Is it a fatal flaw? It's nowhere near as bad as many will make it out to be. (As an aside, many don't score pneumo data in DLCQTs.)IP: Logged |
Buster Member
|
posted 12-04-2009 05:53 PM
I did stop the test after two charts and no called it. It wasn't totally because of the GSR, I also got really unstable cardio readings. I will post those charts tonight, I have been buried at work. See in my spot, I am asked to do a test and my bosses expect a test. Also (this wass lenghty conversation before on this site) we get one shot to interview these people. Not in this case particularly, but in general. I was in a task force meeting in Philly yesterday and they offered two people polygraphs to be conducted today during an interview. They both no showed. Detectives felt if I was there yesterday we may have a closed case right now. [This message has been edited by Buster (edited 12-04-2009).] IP: Logged |
stat Member
|
posted 12-05-2009 10:00 AM
I like Sackett's post script. A goofed up question sequence is also a point of uncertainty and even---if the test is high profile---quiet terror. Material construction and physics tells us that probably half of everyone's spygmanometer cuff has a slow leak and air pressure resistance is trailing off----does anyone do the soup can test anymore? Probably not due to corrective line tracing---but the compononet fails like any thing else mechanical. Like a transmission that steadily fails over time, at what point does an average examiner replace it----without waiting for a zipper tear? PS I ask about the cuff as I have had a few cardio cuffs leak over time, and the calibration test (1 per chart) would never reveal such, nor would it reveal gsr problems. To protect the identity of the manufacturer, I shall refer to the brand as Schmoelting.
[This message has been edited by stat (edited 12-05-2009).] IP: Logged |
Buster Member
|
posted 12-05-2009 08:19 PM
Test w/o GSR: No Call/Stopped after 2 chartsDLC Chart #1
Chart #2
[This message has been edited by Buster (edited 12-05-2009).] IP: Logged | |