140 likes | 250 Views
E470: Intro. I. ECMT as Econ * Stats. Econ Theory Model of Cause/Effect Ecmt Model. I. ECMT as Econ * Stats. Econ Theory Model of Cause/Effect Ecmt Model Stat Facts Identifiable Variables/Data Refined Data. I. ECMT as Econ * Stats.
E N D
I. ECMT as Econ * Stats • Econ Theory Model of Cause/Effect Ecmt Model
I. ECMT as Econ * Stats • Econ Theory Model of Cause/Effect Ecmt Model • Stat Facts Identifiable Variables/Data Refined Data
I. ECMT as Econ * Stats • Econ Theory Model of Cause/Effect Ecmt Model • Stat Facts Identifiable Variables/Data Refined Data • Stat Theory Ecmt Technique
I. ECMT as Econ * Stats • Econ Theory Model of Cause/Effect Ecmt Model • Stat Facts Identifiable Variables/Data Refined Data • Stat Theory Ecmt Technique Stat Analysis, incl. Hypothesis Testing, Policy Evaluation, and Forecasting
I. ECMT as Econ * Stats -moving from qualitative to quantitative
I. ECMT as Econ * Stats -moving from qualitative to quantitative -positive vs. normative economics
I. ECMT as Econ * Stats -moving from qualitative to quantitative -positive vs. normative economics -models and stats/data as proxies
I. ECMT as Econ * Stats -moving from qualitative to quantitative -positive vs. normative economics -models and stats/data as proxies -role and importance of (good) theory
II. Intro to regression analysis • sample equation: Y = B0 + (B1 * X) + E (error term) • when estimated: Y-hat = B0-hat + (B1-hat * X) + e (residual)
II. Intro to regression analysis Y-hat = B0-hat + (B1-hat * X) + e (residual) • Dependent and Independent Variables (Y & X) -causality vs. correlation
II. Intro to regression analysis Y-hat = B0-hat + (B1-hat * X) + e (residual) • Dependent and Independent Variables (Y & X) -causality vs. correlation • Estimated Coefficients (B0 and B1) -intercept (interpreting is tricky at best...) -slopes (linear, quadratic)
II. Intro to regression analysis Y-hat = B0-hat + (B1-hat * X) + e (residual) • Dependent and Independent Variables (Y & X) -causality vs. correlation • Estimated Coefficients (B0 and B1) -intercept (interpreting is tricky at best...) -slopes (linear, quadratic) • Error Term (E); Residual (e) -goal: minimize residuals; find best fit -sources of residual: measurement error, mis-specified model --> points to import of good data, good theory --> points to the import of looking at and analyzing residuals
II. Intro to regression analysis Y-hat = B0-hat + (B1-hat * X) + e (residual) • Error Term (E); Residual (e) -goal: minimize residuals; find best fit -sources of residual: measurement error, mis-specified model --> points to import of good data, good theory --> points to the import of looking at and analyzing residuals -observed Y has error term (vs. true line) vs. estimated/fitted value E(Y) or Y-hat (difference as residual) -residual = observed - estimated regression E(Y), a number -error term = observed - "true" regression E(Y), which cannot be measured