Hello,
I have questions about statistical method regarding to my research;
I used four document papers, two were computer generated and two from actual
human generated. Subjects were randomly assigned to conditions wherein they
received either two computerized papers, two human papers, or one of each.
The order in which they were given the papers and which of the two instances
they received (when they received one of each) were also randomized.
Subject were asked to read the paper declare whether each of their two paper
were human generated or computer generated, and to rate the certainty of
each judgment on a seven point scale, form very certain to very uncertain.
In addition they were asked to give their reasons for making the judgments
they did.
1. Can the Subjects perceive as different the computerized from human paper
at better than chance rates?
2. Are subjects who have one of each type of papers more accurate in their
judgments than participants who have two papers of the same type? (i.e.,
does the ability to compare a computer to a human resume improve judgment
accuracy?)
3. Is certainty related to accuracy of judgment?
The problem I am having is that, each of my subject has two observations,
the error term is correlated. I have found out that the Mcnemar test can
answer question#2. But my ultimate goal is to answer overall in question #1.
Does anybody have any suggestion? I am greatly appreciated and Thank you in
advance
Bob L. Sookvong
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|