[Sigia-l] Prominence in Short Lists

[Matthew Rehkopf]

Dan asked:
"Figuring that Items X and Y will be chosen much more often than the
"with complications" items, what is the best way to list them?"

My stats class covered this topic fairly well. The text (Statistics and
Research Methods for Managerial Decisions, Davies/Utts/Simon, 2002) cites
M.Parten in "Surveys, Polls, and Samples (New York: Harper & Row, 1950);
pp.210-211 as stating that if alternatives in the measurement instrument are
lengthy or complex, or if the respondent does not feel strongly about the
subject, the alternative *presented last* will be chosen more often than
others. 

It goes on to say that if you expect this bias, you should use the
"split-ballot technique", where separate instruments are developed, rotating
the order of the alternatives so that each alternative appears in the final
position as equal number of times. 

Matt