notes3-23.doc

J201 3/23 CREATEDATE 3/23/09 9:52 AM The future of polling Polling as a tool or horserace odds? Does polling demonstrate the decline of the Bush presidency? Citizen support of state policy Approve or disapprove of handling the war in Iraq Break down of percentages Early after 9/11 50% thought that Saddam Hussein was directly related to the attack (never proven he had any thing to do with it) 2006 Zogby International poll of US troops stationed in Iraq 72% US should withdraw all troops from Iraq within a year 42 unclear about US?s role in Iraq Approval/disapproval Bush: started at 50, went up to 90 after 9/11, went down to 50 by 2003 but then went up to 75 when declare war, then goes down again then up slightly with Saddam, finally start sliding down until end of term to lower 30% Pollster.com: Obama and Clinton close primary race Seeing this has effects who goes out and votes for their candidate Obama in office Favorable and unfavorable ratings, first week after election, fav goes up, then starts to go down once reality sinks in Who voted for dems vs who voted for reps Different topics How polls affect economic factors Why engage in political opinion polling Identify market segments for particular messages? Determine whether messages are effective? Estimate policy support as part of a news story? Ascertain the will of the people fot governance A good poll Locates people who fairly represent the population we?re interested in Asks them fair questions, comprehensible and useful questions (push polling, directing question in a certain way to get out info that the caller wants oublic to know) Calculates results fairly, without 4 ways to screw up polling sampling error non-random (personal biases) coverage errors cell phones (are the means you?re using effective) response rate errors answering machines (people don?t respond) measurement errors bad questions Margin of error (one over the square root of n (number of voters) Thousand people means margin of error is about 3%