Self-esteem is another one of those null concepts like "fear of success". In my own work, for example, I've identified at least 2 (and maybe three) distinct mental processes by which behaviors described as "low self-esteem" can be produced.

One of the two could be thought of as "status-based", but the actual mechanism seems more like comparison of behaviors and traits to valued (or devalued) behavioral examples. (For instance, you get called a crybaby and laughed at -- and thus you learn that crying makes you a baby, and to be a "man" you must be "tough".)

The other mechanism is based on the ability to evoke positive responses from others, and the behaviors one learns in order to evoke those responses. Which I suppose can also be thought of as status-based, too, but it's very different in its operation. Response evocation motivates you to try different behaviors and imprint on ones that work, whereas role-judgment makes you try to conceal your less desirable behaviors and the negative identity associated with them. (Or, it motivates you to imitate and display admired traits and behaviors.)

Anyway, my main point was just to support your comments about evidence and falsifiability: rationalists should avoid throwing around high-level psychological terms like "procrastination" and "self-esteem" that don't define a mechanism -- they're usually far too overloaded and abstract to be useful, ala "phlogiston". If you want to be able to predict (or engineer!) esteem, you need to know more than that it contains a "status-ative principle". ;-)