Discussion about this post

User's avatar
Bob Armstrong's avatar

I had the PSSC physics curriculum in the early 1960s at Culver Military Academy .

I don't know how many schools used it , but I credit it in large part for my getting an 800 on the Physics SAT .

Now , close to 70 years later all high school physics curricula should , obviously , be even better .

Expand full comment
William Lynch's avatar

Such a courageous counter, John.

You could be spending the rest of your life in expounding upon this. I most definitely do not envy you.

I had a large document that i submitted to the Nationsl Library of Congress on the topic of evaluating both real and relative performance gains among individuals and among cohorts. I had given up on statements/proofs/examples to the State school board. controlled newspapers, and congressional committees. I even had a full-fledged plan for replacing NCLB - a noble idea that was butchered by politicians and teacher unions. There are good definitions of "good" tests but no correct official definitions of relative knowledge or of genuine progress. (My metrics assume, however, that what is being taught is what should be being taught. I have always had a positive view of all the teachers I have had from K1 to a PhD.)

The metrics chosen by the politicians, however, are the differences in scores or passing percentages from year to year, but the test score scale is NOT a knowledge scale. It is a logarithmic scale, in which one standard deviation (SD) represents a knowledge ratio of about 2.2. A 99-percentile student (+3SD) knows eleven times as much (about the test material) as the median student. A 95-percentile student (+2 SD) knows five times as much.

If the year-to-year tests are made less difficult - which they automatically are if the teachers are teaching to the test - the administrators (and newspapers) will declare that knowledge "gaps" between any two chosen cohorts are decreasing. In fact, 30 years of testing across all States demonstrate that the gaps between the two major cohorts has NOT changed.

Furthermore, my chosen set of metrics give the same relative knowledge ratios for all tests on that subject matter. Other truths can also be established. Impoverished family backgrounds (Free Lunch) vs. No Free Lunch students have approximately the same gap as the gaps based on ethnic differences.

This will not change as long as a typical classroom has one teacher with an incoming set of students with knowledge spreads of four to six SD. (i.e., a four to 120 range of knowledge). Everyone is favored if from 4th grade forward., principals make class assignments (with generous overlap) of zero to 20-percentile, 20 to 80-percentile, and above 80-percentile. The lowest 20-percentile must be saved from permanent failure and the top 20-percentile must be given much greater challenges. Both the original 0-20 percentile and the original 80 to 100-percentile student will have shown great progress by K12. ... Bill Lynch, bandglynch@gmail.com

Expand full comment
9 more comments...

No posts