Suppose you believe that, in general,graduates who have majored in your subject are offered higher salaries upon graduating than are graduates of other programs. Describe a statistical experiment that could help test your belief.

Answer 1

The first thing you need is a random selection of graduates whose salaries can be compared... and that's probably the trickiest bit, because getting a genuinely random selection of graduates could more difficult than you might think. Suppose for example your subject is accountancy, and you want to compare salaries of graduate accountants with those of graduate chemists. Where can you find a lot of accountants? In an accountancy firm, of course - but you won't find many chemists there. So you look up an accountancy firm and a chemical company in the phone book, and start asking people at random what their salaries are... and you'll get absolutely nowhere, because they probably won't tell you! Also if your companies are in different cities, you could find that the cost of living is higher in one city than the other - and that could be reflected in the salaries. On top of that, you could find for example that the average ages of the two groups are different - in which case the older group is probably paid more than the younger group, quite simply because they've been working longer, and their careers are more advanced.

You'll need to control as many extraneous factors like these as you can when selecting your groups, so I'd suggest that you look at all the graduates from a set of specific classes (i.e. they graduated from the same schools/colleges/universities or whatever at the same time), and then went to work in the same areas (or at least comparable areas in terms of the cost of living, cost off housing, local tax rates etc). Now you need to start thinking about HOW you're going to get them to tell you what their salaries are - and making sure that you compare like with like there as far as possible there as well. Is it possible for example that highly-paid accountants are going to be more cagy about telling you what they earn than poorly-paid chemists? I don't have an answer to that one, but it might help you start thinking about what you'll need to consider when setting up your experiment.

Answer 2

Thanks for the response