OLS_Wage_Equations_Web_fromExample1_2perPage

OLS_Wage_Equations_Web_fromExample1_2perPage - Econ 1002:...

Info iconThis preview shows pages 1–4. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Econ 1002: Applied Economics Matthew Wakefield University College London and Institute for Fiscal Studies m.wakefield (at) ucl.ac.uk Jan - Mar 2009 Oce Hour: Tues 3-3.55, Room 209, Drayton www.ucl.ac.uk/economics/degree-courses/undergraduate/course-list/courses/econ1002 coursepage or link from: www.ucl.ac.uk/economics/degree-courses/undergraduate/course-list/courses/applied-economics Thanks to the previous two lecturers, Professors Ian Preston and James Banks, for much of the content of this course. Econ 1002 Matthew Wakefield Outline Stats Properties BLUE FIT R-Squared Errors Example 1 Tests Intervals Multiple Regression Example 1 Outline II: An Introduction to Ordinary Least Squares (OLS) Regression Using the example of Wage Equations 1 Statistical Properties & Procedures Properties of OLS BLUE 2 Fit, Intervals and Tests R-Squared Standard Errors Example 1 - again Tests Intervals 3 Multiple Regression Regression Example 1 - yet again Matthew Wakefield (UCL/IFS) Econ 1002 Jan - Mar 2009 2 / 22 Econ 1002 Matthew Wakefield Outline Stats Properties BLUE FIT R-Squared Errors Example 1 Tests Intervals Multiple Regression Example 1 Stats Properties From the example, note that: If we want make a strong case about the economic interpretation , we must be prepared to assume that our simple wage equation describes how wages are generated. Only if that is true can we say that and extra year of schooling causes a 6.6% increase in wages (on average). But, the assumption that population relationship underlying data really is linear in parameters and as in: y i = + x i + u i is also starting point for thinking about statistical properties: If this is the population relationship then it makes sense to ask whether regression will be good at pinning down (identifying) these population parameters Matthew Wakefield (UCL/IFS) Econ 1002 Jan - Mar 2009 3 / 22 Econ 1002 Matthew Wakefield Outline Stats Properties BLUE FIT R-Squared Errors Example 1 Tests Intervals Multiple Regression Example 1 Stats Properties Why cant we just measure the parameters? When performing regression we (almost) never have the whole population Use a sample Many different possible samples in same population Each possible sample would return somewhat different estimates Useful if these different estimates are (tightly) clustered around the population values, so when we draw a given sample we can be confident of getting accurate estimates of and . Now discuss: When OLS will deliver estimates that are clustered around population parameters How we can quantify our confidence in our estimates Matthew Wakefield (UCL/IFS) Econ 1002 Jan - Mar 2009 4 / 22 Econ 1002 Matthew Wakefield Outline Stats Properties BLUE FIT R-Squared Errors Example 1 Tests Intervals Multiple Regression Example 1 Stats Properties Belief that OLS estimates will be clustered around population parameters can only be sustained if we a confident we have a random sample Non-random samples over or under represent units...
View Full Document

Page1 / 11

OLS_Wage_Equations_Web_fromExample1_2perPage - Econ 1002:...

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online