Ambiguous test:

[SpaceBadger]

Originally posted by President Elect Shang:
Sorry Fyron, but if they where grouped than that would invalidate the test as now you would have to pick a pair as opposed to choosing one, then choosing another. The very nature of human behavior should allow for some variance. Also remember that correlation does not equal causation. I don’t want to go into much detail or I may start tampering with the results of those who read the Posts before the test.

Well, I see your point that making the subject choose a pair might produce different results than choosing two answers independently, however I also agree with Fyron that any method that does not match up the answers into pairs is meaningless except as to proportion of subjects who consider themselves to be from the country or the city, and proportion of subjects who answer the second question one way or the other. I think it is a problem with the way the UBB polls work, that it won't let you tabulate any correlation such as a percentage that chose 1A & 2A, percentage that chose 1A & 2B, etc.

I suspect that when your professor did this he recorded the data in a manner that allowed such a correlation, else what would be the point?

SpaceBadger

edit: to clarify: The questions should be asked individually rather than in pairs, but the answers need to be correlated after the answering to gain any meaningful results. I don't think the UBB poll allows you to do this.

[ January 13, 2004, 22:54: Message edited by: SpaceBadger ]