1. For which one the following data sets, we can use correlation coefficient? Explain why. 2 10 12 8 10 12 a b 12 10 10 8 6 2 2. Why correlation is not causation? 3. Why do we minimize the e in order to find the best fitted line for bivariate data? Why not just minimizing e;? 4. Why the following assumptions must be met in order for us to use linear regression? . The errors E1,...,En are random and independent. In particular, the magnitude of any error e; does not influence the value of the next error E; + 1. . The errors E1,...,En all have mean 0. . The errors E1,...,En all have the same variance, which we denote by o2. . The errors E1,...,En are normally distributed