prefer binary pass or fail evaluations instead of scaled ones

Hamel Husain's idea

Prefer binary pass or fail evaluation instead of instead of scaled evaluation

of 1-5 ratings (Likert scales) for applied evals?