兰德-RAND_美_教育家面板技术_Description_42页_578kb

RAND American Educator Panels Summary

Core Content

The RAND American Educator Panels (AEP) are composed of two panels: the American Teacher Panel (ATP) and the American School Leader Panel (ASLP). These panels are designed to provide a representative sample of U.S. K-12 public school educators for evidence-based research on education policies and practices. The panels include classroom teachers (excluding preschool, resource, and athletic coaches) and school principals (excluding assistant principals), from various types of public schools, including charter and magnet schools, but not private or special education schools.

The AEP was created with funding from the Bill & Melinda Gates Foundation and was developed by the RAND Corporation's Education and Labor Division. The report outlines the methodology used in the recruitment of panelists, including sampling, weighting, and variance estimation.

Main Phases of Recruitment

Phase 1: The Westat School Sample (April 2014–February 2015)

• A multistage sampling approach was used, with 2,300 public schools sampled randomly from the Common Core of Data (CCD).
• Wave 1 included 800 schools, with 2 teachers and 1 principal selected per school.
• Wave 2 involved 1,572 schools not used in Wave 1, and contact information was purchased from Market Data Retrieval (MDR).
• Recruitment rates were:
  • Teachers: 11.0% (from Wave 2)
  • Principals: 13.0% (from Wave 2)
• Incentives included a $10 Target gift card, with email follow-ups.
• Contact information accuracy was an issue, with 22% of teachers having incorrect details, while principals had more accurate contact information.

Phase 2: The Helmsley State-Level Teacher Oversample and Second Principal National Sample (July 2015–March 2016)

• Target states: California, Louisiana, New Mexico, and New York.
• Teacher recruitment:
  • Oversampled groups: elementary school teachers (math/ELA or unspecified subjects), non-elementary math/ELA teachers, and non-elementary science/social studies teachers.
  • Sampling method: Single-stage sampling, with 400 panelists targeted per state.
  • Total teachers sampled: 6,200.
  • Recruitment success rate: 20.1%.
  • Additional recruitment via email only (no package) yielded 3.7% response rate.
• Principal recruitment:
  • 3,000 schools sampled nationally (excluding previously contacted schools).
  • Recruitment success rate: 26.7%.
  • Total ASLP panelists by March 2016: 1,340.

Phase 3: The Gates MLI State-Level Oversample (October 2016–March 2017)

• Target states: 22 states and New York City.
• Sampling method: Single-stage sampling, with independent teacher and principal samples.
• Teacher recruitment:
  • Target: 800 panelists per state.
  • Stratified by experience level in six states and by age in the remaining.
  • Oversampling was done for inexperienced and younger teachers.
  • Total educators contacted: 103,729 (63,299 teachers and 40,430 principals).
• Principal recruitment:
  • Target: 1,100 panelists per state.
  • Stratified by experience level in some states.
  • Census sampling was used in states with insufficient principal contacts.
• Recruitment methods:
  • FedEx and USPS were used for mailings.
  • Incentives varied, including preincentives and promised incentives.
  • Endorsement letters from national educator unions and state departments of education were included.

Key Methods and Procedures

• Inverse Probability Weighting (IPW) and Design Weights were used to adjust for nonresponse.
• Calibration of Weights was applied to ensure representativeness.
• Variance Estimation methods included Taylor Series Linearization and Jackknife.
• Recruitment Experiment:
  • Tested different incentive strategies and contact methods.
  • Group 1 (Standard): $10 Target gift card, FedEx.
  • Group 4 ($40 gift card): 25.0% response rate.
  • Group 1 was found to be the most cost-effective with a response rate of 27.6% and a cost per panelist of $66.55.
  • Email-only recruitment (Group 10) had a response rate of 18.8% and a cost of $80.85 per panelist.

Conclusion

The RAND American Educator Panels were created to support evidence-based research on education policies and practices. The recruitment process involved three phases, with different sampling strategies, incentive structures, and contact methods. The most effective recruitment method was found to be the standard approach with a $10 Target gift card and FedEx mailing, which provided a 27.6% response rate and cost-effective results. The panels are weighted and variance estimated to ensure statistical accuracy and representativeness. The project was funded by the Bill & Melinda Gates Foundation, and the findings reflect the authors' views, not necessarily those of the funders.