lecture8_140b_2011

lecture8_140b_2011 - Lecture 8: Regression Analysis of...

Info iconThis preview shows pages 1–8. Sign up to view the full content.

View Full Document Right Arrow Icon
Olivier Deschenes, UCSB, Econ 140B, Winter 2011 Lecture 8: Regression Analysis of Experimental and Quasi-Experimental Data Chapter 13 in S&W Outline: Analysis of experimental data Analysis of quasi-experimental data Impacts of heterogeneity (a bit more advanced)
Background image of page 1

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

View Full DocumentRight Arrow Icon
Motivation: Why study experiments? Ideal randomized controlled experiments provide a benchmark for assessing validity of observational studies Real-world experiments are rare (expensive, impractical) but influential In Economics, many application of ‘lab’ controlled experiments (some examples in class presentations) Experiments can solve the threats to internal validity of observational studies, but they have their own threats to internal validity
Background image of page 2
Terminology: experiments and quasi-experiments An experiment is designed and implemented by human researchers. An experiment features a protocol and the use of a treatment and control group with random assignment (e.g. clinical trials of a drug) A quasi-experiment or natural experiment has a source of randomization that is “as if” randomly assigned, but this variation was not part of a controlled randomized treatment and control design Program evaluation is the field of statistics aimed at evaluating the effect of a program or policy, for example, an ad campaign to cut smoking
Background image of page 3

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

View Full DocumentRight Arrow Icon
Examples of controlled random experiment Clinical drug trial: does a proposed drug lower cholesterol? Y = cholesterol level X = treatment or control group (dose of drug) Job training program (ex: Job Training Partnership Act) Y = employment status (or Y = earnings) X = went through experimental program, or not Class size effect (Tennessee class size experiment) Y = test score (Stanford Achievement Test) X = class size treatment group (regular, regular + aide, small)
Background image of page 4
Random Experiments and Causal Effects An ideal randomized controlled experiment randomly assigns subjects to treatment and control groups More generally, the treatment level X is randomly assigned: Y i = 0 + 1 X i + u i If X is randomly assigned (for example by computer) then u and X are independently distributed, so E ( u i | X i ) = 0, so OLS yields an unbiased estimator of 1 The causal effect is the population value of 1 in an ideal randomized controlled experiment
Background image of page 5

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

View Full DocumentRight Arrow Icon
Estimation of causal effects in an ideal randomized controlled experiment Random assignment of X implies that E ( u i | X i ) = 0 Thus the OLS estimator is consistent When the treatment is binary, is just the difference in mean outcome ( Y ) in the treatment vs. control group, i.e. This difference in means is sometimes called the differences estimator 1 ˆ 1 ˆ C T Y Y
Background image of page 6
Potential Problems with Experiments in Practice A. Threats to Internal Validity 1. Failure to randomize (or imperfect randomization) Ex: openings in job treatment program are filled on first-come, first-serve basis; latecomers are controls Randomization device fails (or is incorrectly used)
Background image of page 7

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

View Full DocumentRight Arrow Icon
Image of page 8
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 09/04/2011 for the course ECON 140b taught by Professor Staff during the Winter '08 term at UCSB.

Page1 / 47

lecture8_140b_2011 - Lecture 8: Regression Analysis of...

This preview shows document pages 1 - 8. Sign up to view the full document.

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