Replying to @garjoh_canuck and @rickbruner, @jason_kint
I fail to see how comparing several bad choice architectures somehow shows anything about revealed preferences. It’s only revealing bad design all around, and possibly a tendency for learned helplessness.