Document Type



Doctor of Philosophy



First Adviser

Chou, Shin-Yi

Other advisers/committee members

Yang, Muzhe; Deily, Mary E.; Li, Suhui


In Chapter 1, we use low birth weight (LBW) and intrauterine growth restriction (IUGR) as proxies for a compromised intrauterine environment experienced by one generation, and examine its association with the LBW (or IUGR) status of the next generation. We create two three- generational samples using Taiwan birth certificates from 1978-2006 to study both maternal and paternal transmissions. The results show that the intergenerational transmission only occurs matrilineally and it is stronger among female offspring. We find weak evidence that females, but not males, born to areas with lower unemployment rate, higher average income, and higher parental education can be buffered from these effects.Chapter 2 uses the most recent vaccine scare in the U.S., the Measles-Mumps-Rubella (MMR)-autism controversy, to investigate how well-educated people respond to information differently when information is mixed. The controversy was first provoked by a paper linking autism to the childhood vaccine of MMR and was retracted years later due to scientific misconduct. We combine state-level information exposures with individual vaccination records from the National Immunization Survey, 1998-2011. Results show that the persistent increase in MMR non-uptake rate is driven by biased beliefs among well-educated mothers, which in turn leads to strong responses to only new information that confirms their beliefs. We find evidence that online search has a more influential impact on the high education group than mainstream media. Chapter 3 investigates the association between published hospital report card information and hospital relative attractiveness to commercial HMO insurers for CABG surgery in Pennsylvania during report card episodes 2006-2010. Relative bargaining position between hospitals and insurers are measured using aggregated changes in individual willingness-to-pay for a particular plan if a hospital is included in its network. Plan's hospital networks are implied using PHC4 inpatient discharge data. Our results suggest that high charge hospitals in the most recent report card episode are 53.4% less attractive to insurers and low charge hospitals are 76.1% more attractive to insurers, given the plan's network. Based on our calculation, low charge hospitals in the most recent report card episode are 20.5% more likely to have a new HMO contract.

Included in

Economics Commons