Here is a revised version of the data interpretation survey. I have edited the graph and the survey to (hopefully) get rid of ambiguities. I would be grateful to anyone who anyone who completes the survey, which should take no more than ten minutes.
It's like randomizing the question order in a survey. Sometimes answers are influenced by incidental qualities of the survey and one wants to filter those out. So in an experiment, I might give half the subjects graphs with losses first and half with gains first.
Why reverse gains and losses?
ReplyDeleteIt's like randomizing the question order in a survey. Sometimes answers are influenced by incidental qualities of the survey and one wants to filter those out. So in an experiment, I might give half the subjects graphs with losses first and half with gains first.
ReplyDelete"filter those out" = "control for"
ReplyDelete