ECON 104 Lecture Notes - Lecture 3: Normal Distribution, Statistical Parameter, Simple Random Sample
Document Summary
If x >/< mean(x), and z >/< mean(z) then the covariance > 0. If covariance < 0 -> one decreases/ other increases. If independently distributed, x direction doesn"t influence z direction of movement. Cov = 0, doesn"t mean completely independent - when x is very far away from mean, z is also likely to be very far away from the mean _____ Can compute a covariance between the 2 variables, and they are negatively correlated. Correlation is a simple concept derived from covariance. If you want to understand concept of height and weight, but way you measure it shouldn"t matter. Perfect positive linear association = x is a perfect linear function of z, slope is positive. Joint distributions are useful because they tell us about conditional distributions. We want to know what is the distribution of test scores given the str is equal to something. Conditional mean is the mean of the conditional distribution - v important.
