Upload
trinhthien
View
214
Download
0
Embed Size (px)
Citation preview
1
Course Syllabus
PubH 8811: Seminar: Health Services Research Methods
Fall 2014
Credits: 3
Meeting Days: Monday and Wednesday, September 3 – December 15, 2012
Meeting Time: 10:10 am – 11:40 pm
Meeting Place: 325 Lind Hall
Instructors: Bryan Dowd
Office Address: 15-205 PWB
Office Phone: 612-624-5468
Fax: 612-624-2196
E-mail: [email protected]
Office Hours: By appointment
Required text:
Greene, William. Econometric Analysis (7th
edition). Prentice Hall: Edgewood Cliffs, New Jersey
(2011). Feel free to use earlier editions.
Highly recommended text:
Kennedy, Peter. Guide to Econometrics (6th edition). Wiley Blackwell (2008). Feel free to use the
5th
edition (MIT Press: Cambridge, MA.,1993).
Other good texts:
Baum, Christopher F. An Introduction to Modern Econometrics Using Stata. Stata
Press. (2006). Errata at http://www.stata-press.com/books/errata/imeus.html.
Cameron, A. Colin and Pravin K. Trivedi. Microeconometrics: Methods and Applications: 2nd
Edition. Cambridge University Press (2010).
Cameron, A. Colin and Pravin K. Trivedi. Microeconometrics Using Stata, Revised Edition /
Edition 2. StataCorp LP. (2008).
Wooldridge, Jeffrey M. Econometric Analysis of Cross-Section and Panel Data. MIT Press (2002).
2
Description and Course Objective:
The course covers problems encountered in empirical health services research. The course is
organized around different estimation problems which arise in health services research data and
covers, for each problem:
a. A description of the problem
b. The consequences of the problem,
c. The methods used to deal with the problem, and
d. A review of the health services research studies which have applied (or
should have applied) those methods
In class we cover (a) through (c), with references to (d). The applied readings primarily are left as
out-of-class assignments, so that we can cover more topics in class.
An important part of the course is the computer exercises that allow the student to analyze data from
health services research projects using the methods covered in class. I will distribute a handout for
reading our MEPS dataset into R, but our exercises will assume you’re using (any version of ) Stata.
Upon completion of the course, the student should be able to take descriptions of applied health
services research problems, model the problem, identify potential problems in estimating the
parameters of interest, and apply appropriate estimation procedures.
Grading:
Grading is based on three exams (which are weighted equally) and a set of shorter assignments
including computer exercises. To receive an A in the course, students must have an A average on the
exams, and turn in all the computer assignments. Shorter assignments typically are due one week
after they are assigned. Late assignments are marked down one letter grade. Shorter assignments
may not be turned in after the last class meeting. All computer assignments must be completed to
receive a passing grade in the course. Students may collaborate on shorter assignments, but
collaborating on exams is strictly forbidden and can result in dismissal from the program.
Assumed Level of Knowledge:
It is assumed that the student has completed, minimally, Biostatistics 7401-7402 with a B average.
Thus, the following sections of Greene's text are assumed to be familiar to the student: Chapters 1-4
and Appendices A-D.
3
Schedule of Topics and Readings:
Class
Meeting Topic
1,2 Causal modeling and properties of estimators
Greene: Chapter 8: Endogeneity and Instrumental Variable Estimation
Pages 251-254: Natural experiments and the search for causal effects, and Summary and
Conclusions.
Kennedy: Chapter 1; Chapter 2, Section 2.1 - 2.8
Dowd, Bryan E. “Separated at Birth: Statisticians, Social Scientists and Causality in Health
Services Research,” Health Services Research (April 2011). Also see accompanying
commentaries by Judea Pearl and James O’Malley.
Holland, Paul. “Statistics and Causal Inference.” Journal of the American Statistical Association
81:396 (December 1986) 945-960.
Whitehouse, Mark. “Is an Economist Qualified to Solve Puzzle of Autism,” Wall Street Journal
(February 27, 2007). http://bpp.wharton.upenn.edu/mawhite/ReadingsFor911/11%20-
%20Whitehouse%20-%20WSJ%202-27-07%20-%20IV%20Methods%20and%20Autism.pdf
Supplementary:
Pearl, Judea. “Statistics and Causal Inference: A Review,” Sociedad de Estadistica e
Investigacion Operativa Test 12:2 (2003) 101-165. ftp://ftp.cs.ucla.edu/pub/stat_ser/R282-test-
prelim.pdf
3 Violation of OLS assumptions: Overview and Non-normal errors
Overview
Greene: Chapter 2 – The Linear Regression Model
Non-normal errors
Greene: pp.127-131: Non-normal disturbances and large sample tests; p.277: Weighted least squares
Kennedy: Chapters 3 and 15
Dowd, Bryan E., Swenson, Tami, Kane, Robert, Parashurma, Shriram, and Robert Coulam.
“Can Data Envelopment Analysis Provide a Scalar Index of Value?” Published “early view”
4
online in Health Economics (2013).
Halvorsen, Robert and Raymond Palmquist. "The Interpretation of Dummy Variables in
Semilogarithmic Equations," American Economic Review 70:3 (June 1980) 474-475.
Kennedy, Peter. "Estimation with Correctly Interpreted Dummy Variables in Semilogarithmic
Equation, American Economic Review 71:4 (September 1981) 801.
Manning, W.G. and John Mullahy. “Estimating Log Models: To Transform or Not to Transform,”
Journal of Health Economics 20 (2001) 461-494.
4 Violation of OLS Assumptions: Variable Specification and Functional form
Greene: Chapters 5: Hypothesis Tests and Model Selection
Chapter 6: Functional Form and Structural Change
Kennedy: Chapters 4 to 6
5 Violation of OLS Assumptions: Heteroscedasticity and generalized linear regression
Greene: Chapter 9: The Generalized Regression Model and Heteroscedasticity
Kennedy: Chapters 7 and 8
6 Violation of OLS Assumptions: Autocorrelation
Greene: Chapter 20.7 – 20.9 Testing for autocorrelation; Efficient estimation when Ω is known;
Estimation when Ω is unknown.
Kennedy: Chapter 8
7 Analysis of panel data: Fixed versus random effect models, and difference-in-
differences
Greene: pp.156-158: Difference-in-Differences
Chapter 11: Models for panel data
Kennedy: Chapters 18 and 19
Bertrand, Marianne, Duflo Esther, and Sendhil Mullainathan. “How Much Should We Trust
Differences-in Differences Estimates?” Working Paper 8841 http://www.nber.org/papers/w8841
National Bureau of Economic Research, 1050 Massachusetts Avenue, Cambridge, MA 02138
(March 2002). Also in the Quarterly Journal of Economics,119:1 (February 2004) 249-275
8,9 Violation of OLS Assumptions: Missing Data, Measurement Error and
5
Stochastic Regressors
Missing Observations
Greene: Section 4.7.4: Missing values and data imputation
Stochastic Regressors: Measurement Error
Greene: Section 4.7.5: Measurement error
Kennedy: Chapter 10
Stochastic Regressors: Lagged Dependent Variables
Greene: Section 20.7.3-4 Testing in the presence of a lagged dependent variable; 20.9.3 Estimation
with a lagged dependent variable
Kennedy: Chapter 10
10-11 Estimation
Greene: Chapters 12: Estimation Frameworks in Econometrics
Chapter 13: Minimum Distance Estimation and the Generalized Method of Moments
Chapter 14: Maximum Likelihood Estimation
Kennedy: Section 2.9, Chapter 14
12 Non-linear optimization methods
Greene: Appendix E
13-14 Discrete dependent variables: Binary logit/probit and interaction terms in non-linear
models, standard errors of non-linear functions
Greene: Chapter 17: Discrete Choice
Kennedy: Chapter 16
Karaca-Mandic, Pinar, Norton, Edward C. and Bryan E. Dowd “Interaction terms in non-linear
models,” Health Services Research (2011) (posted online).
Greene, William. “Testing hypotheses about interaction terms in non-linear models,” Economic
Letters 107 (2010) 291-296.
Dowd, Bryan E., Greene, William H., and Edward C. Norton. “Computation of Standard
Errors,” Health Services Research 49:2 (April 2014)731-750.
Supplementary:
6
McFadden, Daniel. "A Comment on Discriminant Analysis "versus" Logit Analysis," Annals of
Economic and Social Measurement 5:4 (1976).
15-16 Polychotomous dependent variables: Multinomial/conditional Logit; Multinomial
probit; Ordered logit/probit.
Greene: Chapters 17: Discrete Choice
Kennedy: Chapter 16
Feldman, Roger, Finch, Michael, Dowd, Bryan and Steve Cassou. "The Elasticity of Demand for
Health Insurance Plans," Journal of Human Resources 24:1 (Winter, 1989) 115-142.
17-19 Limited dependent variables: Tobit and two-part Models
Greene: Chapter 18: Discrete Choices and Event Counts
Kennedy: Chapter 17:
Supplementary:
Manning, W.G. “The logged dependent variable, heteroskedasticity, and the retransformation
problem,” Journal of Health Economics 17:3 (1988) 283-296.
Mullahy, J. “Much Ado about Two: Reconsidering Retransformation and the Two-part Model in
Health Econometrtics,” Journal of Health Economics 17:3 (1988) 2826.
Buntin, Melinda Beeuwkes and Alan M. Zaslavsky. “Too Much Ado About Two-Part Models and
Transformation? Comparing Methods of Modeling Medicare Expenditures,” Journal of Health
Economics 23 (2004) 525-542.
20-24 Endogenous Explanatory Variables: Sample selection, Instrumental Variable Models,
Natural Experiments, Residual Inclusion, Propensity scores
Greene: Section 10.6: Simultaneous equation methods
19.5: Incidental truncation and sample selection
19.6: Evaluating treatment effects
Kennedy: Chapter 9
Dowd and Town, “Does X Really Cause Y?” (http://hcfo.net/pdf/xy.pdf)
Nichols, Austin. “Causal Inference with Observational Data,” Stata Journal 7:4 (2007) 507-541.
Bhattacharya J, Goldman D, McCaffrey D. 2006. Estimating probit models with self-selected
7
treatments. Statistics in Medicine 25: 389-413.
Basu, Anirban. “Estimating Person-Centered Treatment (PeT) Effects using Instrumental
Variables: An Application to Evaluate Prostate Cancer Treatments,” Journal of Applied
Econometrics (2013) (online at http://onlinelibrary.wiley.com/doi/10.1002/jae.2343/full)
Terza, Joseph V., Bradford, W. David, and Clara E. Dismuke. “The Use of Linear Instrumental
Variable Methods in Health Services Research and Health Economics: A Cautionary Note,” Health
Services Research 43:3 (June 2008) 1002-1120.
Stukel, Therese A., Fisher, Elliot S., Wennberg, David E., Alter, David A., Gottlieb, Daniel J. and
Marina J. Vermeulen. “Analysis of Observational Studies in the Presence of Treatment Selection
Bias: Effects of Invasive Cardiac Management on AMI Survival Using Propensity Score and
Instrumental Variable Methods,” Journal of the American Medical Association 297:3 (January 17,
2007) 278-285.
Baiocchi, Mike, Small, Dylan, Yang, Lin, Polsky, Daniel, and Peter W. Groeneveld. “Near/Far
Matching – A Study Design Approach to Instrumental Variables,” Stanford University:
Department of Statistics.
http://www.google.com/url?sa=t&rct=j&q=near%2Ffar%20matching&source=web&cd=2&ved=
0CCkQFjAB&url=http%3A%2F%2Fwww-
stat.wharton.upenn.edu%2F~dsmall%2Fnearfarpaper.docx&ei=f-
s3UN2uMcqFyQG3o4HIBw&usg=AFQjCNFRbhzm2gEeuyoD-gK9KpvjwC0BNQ&cad=rja
Harris, Katherine and D.K. Remler. “Who is the Marginal Patient?” Health Services Research 33:5
Part 1, (December 1998) 1337-1360.
Imbens, Guido W. and Donald B. Rubin. “Estimating Outcome Distributions for Compliers in
Instrumental Variables Models,” Review of Economic Studies 64 (1997) 555-574.
Optional:
Angrist, Joshua E. and Jorn-Steffen Pischke. “The Credibility Revolution in Empirical Economics:
How Better Research Design is Taking the Con out of Econometrics,” Journal of Economic
Perspectives 24:2 (Spring 2010) 3-30. Be sure to read the commentaries by other authors that follow
this article in the same journal.
Heckman, James J. “Building Bridges Between Structural and Program Evaluation Approaches to
Evaluating Policy,” Journal of Economic Literature 48:2 (June 2010) 356-398. And Heckman,
8
James J. and V. Joseph Hotz. “Choosing Among Alternative Nonexperimental Methods for
Estimating the Impact of Social Programs: The Case of Manpower Training,” Journal of the
American Statistical Association 84:408 (1989) 862-874.
Duan, Naihua, Manning, Willard G., Jr., Morris, Carl N. and Joseph P. Newhouse. "A Comparison
of Alternative Models for the Demand of Medical Care," Journal of Business and Economic
Statistics 1:2 (April 1983) 115-126.
Hay, Joel W. and Olsen, Randall J. "Let Them Eat Cake: A Note on Comparing Alternative Models
of the Demand for Medical Care," Journal of Business and Economic Statistics 2:3 (July 1984)
279-282.
Duan, Naihua, Manning, Willard G., Jr., Morris, Carl N. and Joseph P. Newhouse. "Choosing
Between the Sample-Selection Model and the Multi-Part Model," Journal of Business and Economic
Statistics 2:3 (July 1984) 283-289.
Nawata, Kazumitsu. “Estimation of Sample Selection Bias Models By the Maximum Likelihood
Estimator and Heckman’s Two-Step Estimator,” Economic Letters 45 (1994) 33-40.
Maddala G.S. "A Survey of the Literature on Selectivity Bias as it Pertains to Health Care Markets,"
in Advances in Health Economics and Health Services Research, Richard Scheffler and Louis
Rossiter, eds. JAI Press:Greenwich, CT (1984).
Manning, W.G., Duan, N. and W.H. Rogers. "Monte Carlo Evidence on the Choice Between
Sample Selection and Two-Part Models," Journal of Econometrics 35 (1987) 59-82.
Dowd, Bryan, Feldman, Roger, Cassou, Steve and Michael Finch. "Health Plan Choice and the
Utilization of Health Care Services," Review of Economics and Statistics 73:1 (February, 1991) 85-
93.
Leamer, Edward E. "Let's take the Con out of Econometrics," American Economic Review 73:1
(March 1983) 31-43.
Vella, Francis. “Estimating Models with Sample Selection Bias: A Survey,” The Journal of Human
Resources 33:1 127-169.
25-28 Endogenous Explanatory Variables: Simultaneous equations
Greene: Chapter 8: Endogeneity and Instrumental Variable Estimation
Kennedy: Chapter 11
9
Supplementary:
Luft, Harold S., Hunt, Sandra S. and Susan C. Maerki. "The Volume-Outcome Relationship:
Practice-Makes-Perfect or Selective-Referral Patterns?" Health Services Research 22:2 (June 1987)
157-182.
McClellan, M. B., McNeil, B. and J. Newhouse. "Does More Intensive Treatment of Acute
Myocardial Infarction in the Elderly Reduce Mortality," Journal of the American Medical
Association (September 1994) 859-893.
McLaughlin, Catherine G. "HMO Growth and Hospital Expenses and Use: A Simultaneous
Equation Approach," Health Services Research 22:2 (June 1987) 183-205.
Optional
Heckman, James L. “Causal Parameters and Policy Analysis in Economics: A Twentieth Century
Retrospective,” NBER Working Paper 7333 (September 1999).
29-30 Analysis of duration data
Greene: Section 19.4: Models for Duration
Kennedy: 17. 4
Supplementary:
Keifer, Nicholas M. “Economic Duration Data and Hazard Functions,” Journal of Economic
Literature 26:2 (June 1988) 646-679.
Amemiya, Takeshi. Advanced Econometrics. Section 11.2
Welch, W. P. "HMO Enrollment and Medicaid: Survival Analysis with a Weibull Function,"
Medical Care 26:1 (January 1988) 45-52.
Additional topics (time permitting)
Count data
Greene: 18.4: Models for counts of events
Regression discontinuity
10
Greene: 19.6.3: Regression discontinuity
Cook, T.D. “Waiting for Life to Arrive”: A history of the regression-discontinuity design in
Psychology, Statistics and Economics. Journal of Econometrics 142:2 (2008) 636-654.
Card, David, Dobkin, Carlos, and Nicole Maestas. “Does Medicare Save Lives?” Quarterly
Journal of Economics 124:2 (May 2009) 597-636.
Supplementary:
Lee, David S. and Thomas Lemieux. Egression Discontinuity Designs in Economics,” Journal
of Economic Literature 48:2 (281-355).
Imbens, Guido and Thomas Lemieux. Regression Discontinuity Designs: A Guide to Practice,”
Technical Working Paper 337 http://www.nber.org/papers/t0337 National Bureau of Economic
Research, 1050 Massachusetts Avenue, Cambridge, MA 02138, (April 2007).
Microsimulation
Atherly, Adam and Bryan E. Dowd. “Should Health Medicare Beneficiaries Postpone
Enrollment in Part D,” Health Economics 18 (2009) 921–931.
http://onlinelibrary.wiley.com/doi/10.1002/hec.1413/pdf
Joyce, G.F., Keeler, E.B., et al. (2005). “The Lifetime Burden of Chronic Disease Among the
Elderly.” Health Affairs Web Exclusive 5:R18-R29.*
Spielauer, M. (2011) “What is Social Science Microsimulation?” Social Science Computer
Review 29(1):9-20.*
Matching and Linking data
Disability/Accessibility Statement
1) It is University policy to provide, on a flexible and individualized basis, reasonable
accommodations to students who have documented disability conditions (e.g., physical, learning,
psychiatric, vision, hearing, or systemic) that may affect their ability to participate in course
activities or to meet course requirements. Students with disabilities are encouraged to contact
Disability Services for a confidential discussion of their individual needs for accommodations.
Disability Services is located in Suite 180 McNamara Alumni Center, 200 Oak Street. Staff can
11
be reached by calling 612/626-1333 voice or TTY. The website is
http://disserv3.stu.umn.edu/index2.html.
2) Letter grades will be determined by total effort as follows:
A = 95-100 points
A- = 90-94 points
B+ = 87-89 points
B = 83-86 points
B- = 80-82 points
C+ = 77-79 points
C = 73-76 points
C- = 70-72 points
F (or N) – Represents failure (or no credit) and signifies that the work was either (1)
completed but at a level of achievement that is not worthy of credit or (2) was not completed and
there was no agreement between the instructor and the student that the student would be awarded
an I.
S – Achievement that is satisfactory will be expected to complete all assignments and receive a
minimum of 70% to receive a passing scores (achievement required for an S is at the discretion
of the instructor but may be no lower than a 70%).
I – An incomplete grade ("I") is permitted only in cases of exceptional circumstances and
following consultation with the instructor. In such cases, an "I" grade will require a specific
written agreement between the instructor and student specifying the time and manner in which
the student will complete the course requirements. Extension for completion of the work will not
exceed one year. Additionally, some majors in the School of Public Health may place a hold on
a student’s registration until an “I” is cleared up, or a plan for completion of the work submitted.
3) Students may change grading options without written permission as specified by the
University and without penalty during the initial registration period or during the first two weeks
of the semester. The grading option may not be changed after the second week of the term.
4) School of Public Health students may withdraw from a course through the second week
of the semester without permission. No “W” will appear on the transcript.
After the second week students are required to do the following:
1. The student must contact and notify their advisor and course instructor informing them of
the decision to withdraw from the course.
2. An email must be sent to the SPH Student Services Center (SSC)
12
([email protected]) from the student. The email must provide name, ID#, course
number, section number, semester and year with instructions to withdraw from the
course, and acknowledgement that the instructor and advisor have been contacted.
3. The advisor and instructor must email the SSC acknowledging the student is canceling
the course. All parties must be notified of the student’s intent.
4. After the SSC receives all emails (student, advisor, instructor) the SSC will complete the
process by withdrawing the student from the course. A “W” will be placed and remain on
the student transcript.
After discussion with their advisor and notification to the instructor, students may withdraw up
until the eighth week of the semester. There is no appeal process. For more information, contact
the SPH Student Services Center at 612.626.3500.
A refund schedule for tuition and fees is listed in the University class schedule. Please refer to
these dates when withdrawing from courses.
5) Scholastic dishonesty is a violation of the student conduct code and is defined as “any act
that violates the rights of another student in academic work or that involves misrepresentation of
your own work. Scholastic dishonesty includes (but is not limited to): cheating on assignments
or examinations; plagiarizing, which means misrepresenting as your own work any part of work
done by another; submitting the same paper, or substantially similar papers, to meet the
requirements of more than one course without the approval and consent of all instructors
involved; depriving another student of necessary course materials; or interfering with another
student’s work.” Scholastic dishonesty in any portion of the academic work for a course shall be
grounds for awarding a grade of “F” or “N” for the entire course. Please consult the student
conduct code at: http://www1.umn.edu/regents/policies/academic/StudentConduct.html