Typing by behavior is often claimed to be unreliable. I have come to think that it might in the end be more reliable than comparing your subjective self-perceptions to theory or to other people's subjective self-perceptions. You can actually measure behavior (quite objectively even) which makes it somewhat "hard data". Of course the interpretation of this data can be unreliable.
Now I was thinking that money usage is the most objective thing you can measure from a person and thus perhaps a reliable test could be built around it. E.g. questions like "Where did you get your money?", "How did you get them?", "How much did you get?", "How did you use them?", "What did you do with the money you didn't use?" and so on can provide a good picture of person's functional preferences. And I'm talking about real actualized money usage not imaginary questions like "What would you do if you had money?". Something which can be very objectively measured. Before you say "No fcucking way!!" think about it a bit and explain why you disagree