In the key points this week, you have read and watched videos about statistical concepts, especially applied to polling. You read a fictitious poll example about people's opinion on the constitutionality of prayer in schools. Here is another fictitious example:

Suppose that on a specific date the Huffington Post published the following headline:

"Non-believers now outnumber believers in most states."

They based this claim on their national poll report (claiming a margin of error 5 percentage points) that asked, "How would you classify yourself, as a) believing b) agnostic or c) atheist?" (RESULTS: a) believing: 21% agnostic: 35% atheist: 44%).

Elsewhere in the article, it said that the survey data also show that three out of four people in the country also believe that any church who speaks out about politics should lose its religious exemption status.

But, after some extensive research elsewhere you find out that the survey only called people in large cities in just four states: California, New York, Oregon, and Washington. You also find out that to begin the survey, one of the demographic questions asked was whether the respondent was affiliated with a specific church. If the respondent answered yes, the interviewer asked a few more perfunctory questions, but then closed the survey. All such responses were not counted in the survey results. If respondents did not claim affiliation with a church, the interviewer asked several other questions. Only people's answers who were not associated with a particular church were counted in the results. A total of 304 surveys were counted toward the findings.

Action Items

1. In a Microsoft Word document, your own fictitious example of a poll. Use the example above as a guide. Keep it simple but make it as real as possible. In your document, also identify what is wrong with the polling information you made up.