posted 05-09-2005 08:33 AM            

I enjoyed Mr. Lynch's article in the relevant issue, but found myself in disagreement with some details.

>Evidentiary polygraph examinations are experiments in which
>we, the experimenters, attempt to eliminate as many
>variables as possible in our testing format so that what
>remains is the null-hypothesis which we intended to prove.
>The null-hypothesis states there is no significant
>difference between the relevant and comparison questions
>in a ZC protocol. By including the outside issue question,
>we add a variable.

Actually I believe you are incorrect on this. That hypothesis underlies only that CQT polygraphs Can be scored.

The null hypothesis for NDI tests is that there is no difference (no statistically significant absence) in reactions to relevant questions while reacting to comparison questions.

The alternative hypotheses (that which we would investigate) for NDI is that there is a statistically significant absence of reactions to relevant questions while reacting to comparison questions.

That is the problem, to cause and measure (and statisiticaly evaluate the significance) - using a two sample Z-test or other common method - the statistically significant absence of reactions.

The null nypotheses for DI tests is that there is again no difference (no statistically significant absence) in reactions to relevant questions while reacting to comparison questions.

The alternative hypothesis for DI tests is that there are statisitically significant differences in reactions to relevant questions compared with reactions to comparison questions.

In research, the goal is to reject the null hypothesis - usually that there is no difference. However, rejecting the null hypothesis does not "prove" the hypothesis - but is said to "support" the hypothesis.

This of course begs another question for investigation - what factors cause one to react to relevant questions while not reacting to comparison questions.

r

------------------

posted 05-09-2005 01:14 PM            

I think you're right, but you might want to back up and explain what this means or you're going to leave a lot of people with no background in statistics (or like I've mentioned somewhere along the way about myself, statistics courses were some time ago) and social science terms feel like the water is getting too deep.