A "Gallup poll" is a national public opinion poll run by Gallup, Inc. According to the company's website, a typical sample size for a phone-based Gallup poll is 1000 randomly-selected adults, and poll results are said to be representative of the opinions of all Americans within a margin of error of ±4 percentage points. What methods do you think Gallup, Inc. must use to ensure true random selection of all 1000 phone participants? Give at least 3 methods. Hint: Consider how lists of phone numbers are generated, which types of phone lines are called, the timing of calls, and how participants are selected when one phone number could be associated with various members of a household.

Imagine that you were given the resources to design and conduct a national public opinion poll of your choice. What would be the subject of your survey? What 5 questions would you include? What method would you use to randomly select your participants? How is the language of your survey questions carefully constructed to avoid bias?

Have you seen a recent study or survey whose results seemed hard to believe? What were its claims? What aspects of experimental design do you think might have contributed to these results? What recommendations would you make to improve the experimental design?

To determine whether dogs prefer a specific brand of dog food to that of a competitor, market researchers are working with a local dog shelter to randomly select dogs for participation in a study. All randomly selected dogs are put into a large room with two bowls of food — one with their brand and another with the competitor's brand. The first two dogs immediately go for the dog food containing their ingredient, so the researchers stop the experiment, concluding this is enough evidence that their product will sell the most. Is this a good experiment and conclusion? State reasons for your explanation, and what you would do differently. which situation is an example of observational study.