卓越實證概述 Best Evidence in Brief

Are low-performing schools curable?

As a result of federal educational policy, four common responses to address low-performing schools have included turnaround (drastically improving the schools), labeling (motivating schools through public performance-related labels), closure (shutting down the schools), and charter conversion (changing the school to a charter school). However, opponents of these school improvement efforts often argue that they are not very effective. A recent paper adopted meta-analysis to investigate the effects of these school improvement approaches in K-12 U.S. schools.

Researchers from University of Virginia, Harvard University, and Education Trust conducted a thorough literature search, which returned 67 eligible studies of a variety of school improvement policy efforts. Their results suggested that:

  • On average, investing in improving low-performing schools has a significant moderate effect on standardized measures of math (ES=+0.06) and a small, non-significant effect on standardized measures of English Language Arts (ES=+0.02). 
  • On low-stakes tests, moderate impacts were identified on STEM (ES=+0.07), and humanities (ES=+0.08).
  • Apart from those academic outcomes, researchers also looked at three non-examination outcomes, and found that turnaround policies had no significant impacts on attendance (ES=+0.11), disciplinary infractions (ES=-0.01), and graduation (ES=+0.04).

Researchers also found that, across different disciplines, turnaround policies have the greatest benefits for schools serving majority-Latinx student populations. Another intuitive conclusion is that longer interventions and more intervention types tend to produce larger effects. In particular, extended learning time and teacher replacement are the two most recommended intervention features.

 

Source (Open Access): Schueler, B. E., Asher, C. A., Larned, K. E., Mehrotra, S., Pollard, C. (2020). Improving low-performing schools: A meta-analysis of impact evaluation studies (EdWorkingPaper: 20-274). Annenberg Institute: Brown University.

Leave a Comment

發表評論