When I was watching the serier Wire there was a depiction of school circumnstances and one of the points seemed to be that the teacher was frustrated with the conditions. It seemed odd that is was supposed to be commenting on real world conditions.
The problem (depicted and what I understand) is not that the supervising examinaations woudl be added paperwork and prepartion angst for the students. Rather it is that the teacher is supposed to teach so much in so little time that there is only room for the most route skim on everything. It is teaching to the test, every student barely passes the test (out of those that do). Minimized time budjet and maximised content expectation from school toward the teacher. No slack at all, constantly teetering on the edge of it being possible at all.
I guess the argument is that the current state is that we care so little about the effect of teaching that no effect is a acceptable outcome. And therefore caring to test that there is more effect than no effect would be an improvement. I feel like the essential part of that is the lack of care.
If you have the expectation that the thing wil not be done if you do not check for it, that is a very low trust attitude. In case you have trust you only need to start monitoring when you lose that trust. If you have to tease and pressure the agent to do the principals bidding you are only going to get exactly what you ask for. Empowering the agent you might get stuff that was not previously tested for. You can’t get Goodharted so bad if you do not micromanage while throwing more resources at it will get you more.
It is quite easy to think of a doctor that is tired and hurries up the patient in order to get enough patients served for that day, looking at X-rays while not listening to pain descriptions. Difference between 10 and 15 patients served is easy to verify. Misdiagnoses or missed depression diagnosis are hard to verify and to pin the causal pathway.
I am also sure that (some) hedge fund managers can appriciate not killing their gold egg laying geese. Or that in data analysis working smart instead of hard might be quite essential. Or that spending some networking time with billoinares is quite an acceptable excuse to be making only 50% volume of trades that day.
When I was watching the serier Wire there was a depiction of school circumnstances and one of the points seemed to be that the teacher was frustrated with the conditions. It seemed odd that is was supposed to be commenting on real world conditions.
The problem (depicted and what I understand) is not that the supervising examinaations woudl be added paperwork and prepartion angst for the students. Rather it is that the teacher is supposed to teach so much in so little time that there is only room for the most route skim on everything. It is teaching to the test, every student barely passes the test (out of those that do). Minimized time budjet and maximised content expectation from school toward the teacher. No slack at all, constantly teetering on the edge of it being possible at all.
I guess the argument is that the current state is that we care so little about the effect of teaching that no effect is a acceptable outcome. And therefore caring to test that there is more effect than no effect would be an improvement. I feel like the essential part of that is the lack of care.
If you have the expectation that the thing wil not be done if you do not check for it, that is a very low trust attitude. In case you have trust you only need to start monitoring when you lose that trust. If you have to tease and pressure the agent to do the principals bidding you are only going to get exactly what you ask for. Empowering the agent you might get stuff that was not previously tested for. You can’t get Goodharted so bad if you do not micromanage while throwing more resources at it will get you more.
It is quite easy to think of a doctor that is tired and hurries up the patient in order to get enough patients served for that day, looking at X-rays while not listening to pain descriptions. Difference between 10 and 15 patients served is easy to verify. Misdiagnoses or missed depression diagnosis are hard to verify and to pin the causal pathway.
I am also sure that (some) hedge fund managers can appriciate not killing their gold egg laying geese. Or that in data analysis working smart instead of hard might be quite essential. Or that spending some networking time with billoinares is quite an acceptable excuse to be making only 50% volume of trades that day.