OVERVIEW… WHAT’S INVOLVED? Sample Size Sampling Randomization Demographics ** Methodology ** Question ** Sampling Error (calculate?) **
DEMOGRAPHICS Who will you interview/try to interview? What questions (and how many) will you ask them? Fewer increases accuracy, especially in a smaller sample More allows for more nuanced analysis
METHODOLOGY Sample questionnaire & make own spreadsheet Will they record selves or group reads and records Language Who will approach? Who will record?
QUESTIONS Topic Wording—biased wording slants opinions… Are school lunches bad? Should school lunches be improved? Do you like school lunches? What area are you interested in? (Beware of measuring “non- opinions”) How many choices will you give? Fewer increases accuracy, especially in a smaller sample More allows for more nuanced analysis
MARGIN OF ERROR Compare and contrast polls Some survey results that get reported are based on a “poll of polls,” where multiple polls are averaged together. Prominent sites that engage in this practice are FiveThirtyEight (fivethirtyeight.com), Real Clear Politics ( ) and the Cook Political Report (