Superheting to the test
Thinking about the delusional British attempt
to transform Brit kids into Chinese kids by applying Chinese teaching methods.
A sloppy analogy struck me. (Splap!)
This is superhet teaching, while the alternate method (experience and discovery, and to some extent Common Core) is TRF teaching.
I'll let Danbo demonstrate both. (Assuming AM, amplitude modulation)
Radio signals are a sneaky way of carrying an audio frequency superimposed on a MUCH higher radio frequency. You could imagine the radio frequency as being like a color, and the audio frequency as being like a slow vibration imposed on the color. The left band in this picture represents an available spectrum of radio signals.
In a simple 'tuned radio frequency' (TRF) receiver, you use a filter to pick out the desired color, and then use a detector to read or hear the audio. When you tune across the band, you are picking one signal and bringing it directly into the detector, complete with its original color (radio freq). The detector catches the audio.
= = = = =
In a superheterodyne (superhet) receiver, the original signal is MIXED or modulated with a sine-wave generated
by a Local Oscillator. The OSC frequency is always above the tuned input signal, and the OSC slides its frequency in parallel with the tuning to keep the DIFFERENCE between the two CONSTANT. For instance, you might be tuning from 500 to 1500 kilocycles, and the OSC would slide its output from 600 to 1600 to maintain a constant DIFFERENCE at 100. The MIXER or modulator outputs the 100 kc difference, which still has the slow wiggles of audio riding on it. Note that one variable aspect of the original signal has been completely LOST. Color is constantly gray, but the slow wiggles of audio are still present and detectable.
The advantage of the superhet's complexity is not
intuitively obvious. When you have a result that's always on the same radio frequency, you can filter and process the result in consistent ways, without having to vary the parameters of your processors. The TRF result still carries a wide range of radio freqs (colors) along with the audio, so it's harder to filter and process it consistently.
'Teaching to the test' is superhet. (Sloppy part of the analogy!) Low-frequency students require a high-frequency Local Oscillator (tutoring and preparation) to bring their signal up into the gray range where the test can process it tightly and consistently. High-freq students require some damping to bring their signal down to gray.
TRF teaching (discovery, discussion, experience) leaves more of the original color or flavor on the student, and focuses strictly on DETECTING and amplifying the useful or meaningful music or information. You try to find which parts of this student's talent set can become job skills, without trying to reshape his PURPOSE into a dull uniform gray, and without trying to perform the more refined filtering that constant color enables.
Rigidity makes statistics work better, just as rigidity makes filters work better. With radio this is a good thing; with teaching it's just dumb. The proper goal of education is adults who are more competent in work and life than they would have been without education. Reliable and consistent stats are NOT the proper goal of education. Or the proper goal of anything, for that matter.
= = = = =
Footnote for clarity: I'm not saying we shouldn't do testing or evaluation. Obviously we need quality control on both
the student and the teacher. I'm saying we need to break predictability, stop mixing everything into constant gray, so neither
the student nor the teacher can form the horrible habit
of asking "Will this be on the test?"
Labels: Blinded by Stats, Danbo, Experiential education