1 / 18

Agenda

Agenda. Admin: Midterm exam P olicy memos Quasi-experimental methods Education and the law. The challenge of methods. What’s the problem ? • We want to know if an educational policy causes an outcome:

vine
Download Presentation

Agenda

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Agenda • Admin: • Midterm exam • Policy memos • Quasi-experimental methods • Education and the law

  2. The challenge of methods • What’s the problem? • • We want to know if an educational policy causes an outcome: • - Does attending a private school make you more likely to go to college? Is phonics more effective than other instructional methods • - Does smaller class size/charter schools/tutoring increase test scores? • - Does level of centralization of the educational system matter?

  3. What’s the solution? • Experiments (randomization) • Quasi-experiments: • Instrumental variables • Regression discontinuity designs • Difference in differences • Propensity scores • Virtual twins

  4. Instrumental Variables • Best motivated by example • Suppose income and schooling are related by the following equation: • Interested in estimating β1– the return to schooling – but we don’t observe ability • Since ability is almost certainly correlated with schooling and income levels, OLS estimates will be biased

  5. IV, cont’d • In other words, it is unlikely that schooling levels are “as good as randomly distributed” • One possible solution: run an experiment and randomize how long people attend school • (never going to happen in the real world…) • More feasible solution: find something that creates random variation in schooling, without affecting income in any other way • We call something that meets these criteria an instrumental variable

  6. Angrist/Krueger (1991) • One Answer: Your birthday • Lots of state policies require students to enter school on the calendar year when they turn 6 • At the same time they require them to stay in school until their 16th birthday • So, someone who wants to leave school as soon as possible and is born on December 31st has to stay 364 days longer in school than somebody born on 1st January • One starts on Sept 1 of year N and finishes on December 31st N+10 and the other starts Sept 1 of year N+1 and finishes on January 1st of year N+11 • We can test whether your quarter of birth predicts how many years you spend in school…

  7. Does Quarter of Birth Influence Educational Attainment?

  8. Does Quarter of Birth Influence earnings later in life Max difference is about $14 a week

  9. Conditions for IV • Assume we want to estimate the causal impact of a variable X on an outcome Y (but we are worried there is OVB) • Assume we know of a variable Z for which the following two conditions hold: • Valid First Stage: Z must be correlated with X – i.e. Cov(Z,X) ≠ 0 • Exclusion Restriction: Z must influence Y only through its relationship to X • Then we call Z an instrument for X • Note that we can test the first criterion using data, but the second one is fundamentally un-testable

  10. Regression discontinuity and college decisions

  11. Regression discontinuity: Maimonides rule

  12. A current case • Given the apparently conflicting rulings of the Supreme Court's due process cases, do you think teachers and school administrators today have been given clear guidance on what searches are reasonable? Put yourself in their shoes by predicting how this very recent case ruled: • A school official in a Virginia high school searched a student's belongings after receiving information that a student fitting his description was seen smoking marijuana on a school bus. The school official patted down the student, searched his backpack, shoes, pockets, and cell phone. • Plaintiff claimed that HCSB was liable for Fourth Amendment violations based upon its failure to properly train school administrators on how to search students Aside from the court’s decision: • Was this a reasonable search? • What would you have done as a school administrator? Gallimore v. Henrico Cnty. Sch. Bd., No. 14-009 (E.D. Va. Aug. 5, 2014)

  13. (1) whether the search “was justified at its inception” • (2) whether the search as actually conducted “was reasonably related in scope to the circumstances which justified the interference in the first place”.

  14. The court found that Turpin and Saunders had justification for the search from the inception because W.S.G. generally fit the description that the parents had supplied • A male student with long hair. With regard to the reasonableness of the searches, it stated: “Turpin’s pat down of W.S.G. and the search of his backpack, shoes, and pockets were all reasonable in scope because W.S.G. could have hidden drugs in these places.” • It also found that Saunders’ searches of W.S.G’s Vaseline jar and sandwich wrapper were reasonable because they were potential hiding places for drugs. • “The search of the cell phone was not ‘reasonably related’ to the objective of the search—finding evidence of drug use on the school bus earlier that day.”

  15. Plaintiff claimed that HCSB was liable for Fourth Amendment violations based upon its failure to properly train school administrators on how to search students. • The court ruled that it was not, as the following were absent: • (1) the existence of a pattern of constitutional violations • (2) that the School Board failed to train school administrators in an area where there is an obvious need for training.” • Gallimore v. Henrico Cnty. Sch. Bd., No. 14-009 (E.D. Va. Aug. 5, 2014)

More Related