Butterfly ballot study flawed

Nov. 19, the IDS reported on a "study" performed here concerning the butterfly ballot ("Butterfly ballots put to the test"). A spokesperson claimed that "one could infer that about 12,000 ballots were intended for Gore by taking the 4.64 percent difference between Bush and Gore found in the study and applying it to Palm Beach County" but that it would "only be an estimate." \nThat disclaimer is quite an understatement! I am only in the middle of taking first-year statistics, and even I can tell that making that inference is on shaky ground. Let me point out some flaws: \n1. A study like that should be based on randomly selected participants. It wasn't. \n2. Even if it were randomly selected, the results could only be generalized to the population, i.e. psychology students, not extrapolated to Florida voters. \n3. Some of the difference might be attributed to the strange sampling method. If someone told you to pretend vote for a candidate you hated based on your age, would you still follow directions? \n4. No data analysis whatsoever was reported, only flat percentages. \nI say, show us the real data from the study and let us decide. Don't make wild inferences.