Perceived Corruption

Thursday 4 December 2014

Today's data flow

Here is another wonderful example of the sort of thing that is flying around the Internet these days that makes teaching statistics so real. Check out this post on the results of survey about perceived corruption . There is a ranked table of 175 countries with scores out of 100 for perceived levels of corruption in the country. A score of zero means most corrupt and 100 most clean. The average scores are given and the results for the previous two surveys are also given! Have a look at the link and the map!


There is so much potential here for statistical analysis and discussion and probably a whole project. Here are just a few ideas...
  • Make hypotheses about the table before the results are shown. Maybe even survey students based on these hypotheses.
  • Use measures of central tendency to compare different geographical areas.
  • Plot box and whisker diagrams to show spread in those areas.
  • Plot scattergraphs of the results of two versions of the survey and do some linear regression.
  • Do the same for geographical areas - which areas show the most correlation? - why might some show more than others?
  • Ask questions about the data collection methods and samples. Look through the website for answers.

I am sure there is so much more, but I am just getting the ball rolling. Thank you Internet for continuing to supply such rich and relevant data for us to work with.