Stopping yourself from yelling at your annoying relative can be much different from resisting the urge to drink. Emotional self-regulation is a complex function, and as we’ve long known in psychotherapy, trying to willfully manage your emotional states through brute force alone is bound to fail. Instead, regulating emotions also includes skills such as shifting attention (distracting yourself), modulating your physiological response (taking deep breaths), being able to tolerate and wait out the negative feelings, and reframing beliefs.
When offered $5 today versus $10 in a month, many people illogically choose immediate gratification. However, when the question is reframed to make the tradeoffs explicit—“Would you prefer $5 today and $0 in a month or $0 today and $10 in a month?”—more people choose the larger, delayed reward. Research suggests that reframing the question in this way nudges people toward delayed gratification because the different versions of the question employ entirely different cognitive processes.
I could be wrong but those cognitive processes you refer to might be "elaboration likelihood model" basically:
I just quoted the article, but it makes a lot of sense to me that people probably use "willpower" to refer to a few fundamentally different things, which is why we still have neither a solid theory of "willpower" nor ways to train it reliably.
There is probably the component of "does this person react immediately, or do they slow down and consider things first?" Which may be partially a habit, and partially biological.
Then there is a question of the mental model the person has. If one person believes that "X means Y", while other believes that "X means Z", of course they are going to react differently when X happens. This model may refer to external or internal world.
For example, if you make a decision to stop smoking, but then you forger and smoke one cigarette, does your model of the world suggest that "it's okay, mistakes happen at the beginning, the important thing is to throw the rest of the cigarettes away and persevere at your decision", or does it say "well, this means you failed completely, you might as well smoke the remaining 19 in the box, because you are a loser anyway, it doesn't make a difference anymore"?
But other example, perhaps more relevant for the marshmallow test is "when other people tell me that if I do X now, I will get a reward in the future, are they usually telling the truth, or are they usually lying?" Because if the child e.g. has a parent who is a habitual liar, with such experience it is rational to grab the one marshmallow while it is there, instead of trading it for a promise of two marshmallows in the future.
Which means that the ways to "increase willpower" would include, perhaps depending on the situation:
Thank you for eleaborating. And I agree, a person's schema does influence their thinking as you described in your examples.
There continues to be too large of a magnet for psychological studies to "prove something interesting" so the study can become newsworthy. If that motivation comes into the picture, then there will obviously be some effect on the researcher's work. Whether the magnitude is as serious as some studies have shown, the replication crisis in psychology is a clear problem. Of course there are many academics not focused on discovering a pop-psych finding, but the pop-psych urge can't be helping the field as a whole.
Aside from the idea of ego depletion, it is clear that some people are able to include more future-thinking calculation in their decision-making than others, but it's not clear that someone can improve this attribute.