Table of Contents

• Module 1

Module 1

• model output:
  • estimated variance of Betas is square of std error (std dev is the root of variance)
  • use p value to see if beta value is 0. if p value is low, reject null hypothesis
  • Std. Er = estimate / t-value
• Simple Linear Regression Assumptions:
  • Linearity / Mean Zero: expectation of error terms becomes 0
  • Constant Variance: the variance of the errors is = \(\sigma^{2}\) across all predictor values
  • Independence: the errors are independent variables and normally distributed
• ANOVA
  • you have several populations which you sample, then you use box plots to compare the means, variances, etc.
  • assumptions: constant variance amongst groups, independence, normality (linearity assumption not necessary!)