E4M L09

Evaluating Discrepancies for Significance

URDU Lecture -- First 15 minutes is general talk on diverse topics.

Later lecture looks at DATA SET on scores of Girls and Boys on SAT and asks if there is a significant difference. This is done be creating a Normal Model for the scores, assuming that Girl Scores are Normally distributed with some mean GM and variance GV. Similarly Boys scores are normal with mean BM and variance BV. Then we test the hypothesis that GM=BM, that is mean score of boys and girls is equal.

TWO METHODS are used to assess whether or not the differences is significant. One methods is via standard T-statistic and its exact distribution. Second method is to use SIMULATIONS to come up with the same result -- that is, we simulate the T=statistic to find its usual and unusual values.

E4M L09 Evaluating Discrepancies for Signficance - Hypothesis testing based on evaluating discrepancy between null and observation, and assessing significance via p-value