MathBench >

Statistical Tests

 

Testing Differences with the T-Test ...

fishBigger, faster, better... how can you prove things like that? Introducing the t-test ... sounds scary, but you can figure it out. Actually, everything you do in a t-test has a reason, and once you understand those reasons, the whole process will start to make sense. Use a t-test to convince your boss that you're a genius. Bonus cute wizard's lookup table.

 

Testing Goodness of Fit with the Chi-Square ...

dilbertThe hard-working, under-appreciated engineer's evil pointy-haired boss is convinced his workers are making off with long weekends -- after all, 42% of sickdays fall on a Monday or Friday. Use a statistical test to help the engineer accept or reject that hypothesis. Learn to construct the chi-square table and take apart the famous chi-square formula (oh minus ee squared over ee) so you can see how the parts fit together. Yet another cute wizard's lookup table of p-values. Then leave the engineer behind and use your skills to tackle the double hybrid Punnett Square.

 

Simulating Goodness of Fit Tests ....

diceWould you wear shoes that don't fit you feet? Then why accept data that doesn't fit your model?! This could be a job for the chi-square test... but there are other methods as well. Flex your muscles with the brute force method, then decide whether you like brute force or statistical tests better. Then use your skills to tackle problems ranging from genetics to ecology.