Statistical Case Study - MBA Salaries

News on Cases—MBA Salaries

Some folks had questions about the Pelosi Case 2, MBA salaries—a representative sample? We are comparing our sample to the entire Population. This tests two things: [1] do we know how to select a Random Sample from a Population? [2] If their salary data were skewed by a few outliers, we should see some big differences. However, in many ways, a median average for the entire Population might have been even more representative but that is another story!

So, you want to calculate a Random Sample of 20 salaries and you need to mention how you ensured they were randomly selected. We need to know your method. Then you needed to run the basic stats on your 20 and compare them to the text. Here is an example: The Mean (Average) salary for the entire population of 100 is $26,385 according to the text. My random sampling method was to pick every 5th salary figure. The mean of the sample I derived was $27190. (Mean of my 20 or Arithmetical Average). I then calculated a percentage of error or difference (3%) and decided that, at least for that first statistic, the sample I generated was representative (or close enough). My conclusion would be (just about Average at this point) that my random sampling method must have been working since the difference to the whole Population of 100 salaries was pretty small. The reason we use samples is that sometimes it is impractical, difficult or too expensive to measure the entire Population.

Definitions: Population—the whole group

Sample: a part of the group

Difference: in this case, my average sample salary minus the text average salary.

Percentage difference: difference $xx divided by text average salary.

I hope this helps!