There was a famous marshmallow experiment out of Stanfard looking at self-control. It's been followed with many other related studies. One recently had two groups of students do a mental task (if my memory serves, one group had to add or memorize 2-digit numbers, the other group had to add or memorize 7-digit numbers). As the students left the cognitive activity area, they were offered a snack and could choose between fresh fruit or chocolate. The students with the easier cognitive task more often than not chose the fresh fruit, while the students with the harder cognitive task tended toward the chocolate. Again, if memory serves, it was a pretty strong correlation.

The points being, (a) at any given moment we have a limited amount of self-control, and (b) that limited amount of self-control extends beyond any single given task or situation.

We can increase our overall self-control (e.g., focusing for five minutes can be increased to focusing for five hours), but not significantly in a short period of time (e.g., it might take years to increase a persons ability to focus from a five minute period to a five hour period.