Featured Course – Statistics: An Activity Based Approach

In recent election cycles Americans have been bombarded with the results of polls designed to gauge voter sentiment and, ultimately, to predict the results of the election. But as the use and importance of polls has grown so has the level of misunderstanding of what the results actually mean. For example, imagine that there’s an election tomorrow and you’ve just heard a news report stating that candidate A is favored by 55% of voters, candidate B is favored by 45%, and that the margin of error is 3 points. **Assuming the poll was well designed and executed (a very big assumption), does this mean that candidate A has a safe margin over candidate B?**

Many would answer yes but, in reality, the lead is safe **except for the 5% chance that it is not**. This 5% chance that the poll is wrong comes from the confidence level and, even though this number is rarely mentioned in news reports, every poll that states a margin of error also has a level of confidence for that estimate, usually 95%. This level of confidence represents the likelihood that chance alone has produced a result that falls outside of the margin of error. So the results reported above actually mean that there is 95% likelihood that between 52% and 58% of voters favor candidate A (and a 5% chance that the actual number is somewhere outside of that range). If the election was held tomorrow, and candidate B won, **would it mean that the statistics didn’t work**? No, it would simply mean that the pollster was “unlucky” and got a sample that was not representative of actual voter sentiment. The confidence level tells us that there is a 1 in 20 chance (5%) that this could happen. Students in this course build a solid understanding of basic statistical principles like this and they learn to apply their knowledge with confidence in a variety of common scenarios.

**Featured Activity: Cancer Mortality – Visualization Using Charts and Maps**

The amount of data available via the Internet for research, exploration, or just “seeing what’s going on” is truly staggering. Governmental bodies at all levels, researchers, interest groups, and non-governmental organizations have all recognized that making their data available is simply good policy. Among the most prominent providers of general interest data are the U.S. Census Bureau, the Unites State Geological Survey (USGS), the Environmental Protection Agency (EPA) and the United Nations. And exploring this data and learning from it doesn’t have have to be dull and academic. If doubt the truth of this statement watch these two videos of the Swedish Doctor, Hans Rosling, talking about global economic and health trends at the TED Conference.

## Hans Rosling shows the best stats you’ve ever seen |
Hans Rosling’s new Insights on Poverty |

Rosling’s research makes use of freely available data from sources including the United Nations, and his presentation demonstrates the value of using effective tools for analyzing data and communicating the results of that analysis. You’ll find much more information about Rosling’s work and the software shown in the videos on the web site, Gapminder.org.

**The National Cancer Institute** (NCI)

NCI provides access to cancer mortality data for all forms of cancer by state and by demographic. Students in this course use the graphing and mapping tools provided by NCI to examine the data and look for trends and information in the data.

For more information, visit the National Cancer Institute website.

Another source for cancer mapping is the CDC Cancer Statistics website.