An an economics professor at Brown University, Emily Oster, joins Paul E. Peterson to discuss Oster’s latest research, which finds that pass rates on standardized tests declined in district schools during the Covid-19 pandemic, and that those rates were lower in schools with less in-person instruction. We have included the podcast highlights for your below:
- Parents reported a higher rate of learning loss for remote learners.
- 12 states on average 14% drop in math pass rates on state exam from 2020 to 2021.
- If district moved from fully in person to fully hybring loss 10%.
- Test score losses are greater for students did not have access to in person learning.
- Disadvantaged students had less acccess to in person learning, while advantaged students had more access.
- ELA greater disadvantage to virtual learning among students of color.
- Students of color were more likely to only have access to virtual learning and virtual learning was worse for them.
Measuring learning loss
- Combined two sets of data
- state level test data and multi-year pass rates
- SEA state data for who had access to in person learning
- Question: Is it schooling, pandemic environment, or other challenges?
- States with more positive COVID cases actually had more re-opening (political influences)
- More complicated question is it the economic impact affecting performance?
- Determined that schooling matters more than unemployment and other pandemic related issues.
- Variation used – what students had access to.
- Districts who had access to in person learning performed better than those who did not.
- Virginia, Minnesota, Massachusetts, Nevada, West Virginia, Colorado, Ohio, Wisconsin, Connecticut, Rhode Island, Wyoming, Florida
- Adding Mississippi and Arizona
- Other states may not have had re-opening data, did not test, or participation was too low to use.
- Across states show direction of affect is the same in all states – consistent in person learning vs. other issues.
- Data is generalizable.
- Seeing drops in participation.
- Question of who is not participating?
- Tend to be students who are expected to do worse, ie. special ed, low performers, which may mean we are underestimating decline
- Bounds – 2 extreme scenarios: 8% non-participatants = ALL failed or ALL passed
- Found even in most extreme assumptions, show test score loss and larger in places with more virtual learning.
- Places learning virtually experienced greater learning loss.
- Largest loss in math, than ELA – consistent with general research
- Math may be more specific and harder for parents to teach – speculative
- ELA may be easier for parents to scaffold and help with.
- California lowest amount of in person learning last 2 years.
- Florida closed very little and showing very little learning loss.
- A lot of work to be done over next several years to ameliorate this loss.
- It’ll take time to recover and will need to continue to thoughtful use of resources.
- Places with the most learning loss will need more resources.
- Moving to virtual learning can be costly and does not appear to effect covid rates.
- Hope for hybrid learning, but did not provide the stability students need.
- Close schools last, not first.
- Likely that some of this loss will be made up.
- Some of the loss will persist indefinitely.
- May start to see more adults without education because high school students left school during and due to the pandemic.
Listen to podcast HERE
Oster’s paper, “Pandemic Schooling Mode and Student Test Scores: Evidence from US States,” co-written with Clare Halloran, Rebecca Jack and James C. Okun, is available now.