I have been a cofounder twice in my career. I have also left paying jobs to work as a contract consultant three times. Every time I do this I hear from several different people (wistfully) that they "don't know how I can do it." Invariably there's a remark about the security that comes with working at large companies, and the conversation moves on.
To be clear, that "security" is a myth. Just ask anyone that currently works at Intel. The last summer job I had was during the Summer of '80 working a melt shop floor in a steel mill. We were in the depths of the "blue collar" recession in Canada, and the steel industry was taking it the teeth. All the summer students were laid off mid-summer. The first programming job after university I had was with a consulting firm that designed steel mills. I watched them shrink from 350 people to about 15 over the next three years before closing their doors. I got over the idea that the company was my friend early in my career. I had many friends at DEC as it collapsed. When you watch any company (big or small) go through that sort of turmoil you quickly learn that regardless of the message you get about your position being safe, you won't honestly know until the pink slips stop falling. Security is the myth we tell ourselves.
What a startup does require however is a different attitude towards responsibility. There is no safety net. Things need to be done. The founders and early employees need to do them. No one else is there to catch the ball. There is no "staff" to take up the slack. (Indeed this is why many executives from large companies make terrible startup execs. They have always had staff to execute on their vision. In a startup, you get to have your vision and deal with it too.)
Decisions need to be made. You need to be very comfortable with the idea that you will make mistakes and they will be plainly there for all to see. My father (an engineer's engineer) was one of those few people I know that had a genuine mentor. Early in my father's career his mentor took him aside and told him that there was a management study that demonstrated that one out of four decisions he made would be wrong. Any more than this and he wasn't taking enough time to consider the decision. Any less than this and he was taking too long to make decisions. The trick is to ensure the mistakes are the less expensive decisions relatively speaking. My father rapidly climbed the management chain moving from company to company at his own direction and was very successful. He faithfully passed along this advice to me when I started working.
It wasn't until I was well into my first startup that I realized the truth. I have never seen such a study. I'm pretty sure his mentor gamed him. He needed a young capable engineer to make decisions. So he told him it was okay to make mistakes, told him the ratio that was "accepted", and that there was a study to prove it. God bless him.
In a startup you need to be very comfortable constantly making decisions with less than perfect data that will fundamentally effect the course of a small company. Some of those decisions will be wrong. Everyone on the team will know it. How you behave about such mistakes will define your personal success more than the bad decision ever could. How you behave when others on the team make their mistakes will fundamentally define the success of the team. As the old saying goes, your successes belong to the team, but your mistakes are all your own. No where is this more true than in a startup.
There's another feeling with which you need to be comfortable. That's the feeling that comes from realizing not only all the decisions you're responsible for making (or things you need to do), but that there is a class of important (but less urgent) things for which NO ONE else is yet responsible. Some things just don't have an owner yet. These might be things like benefits. Or 401K programs. You suddenly need to be responsible for those decisions at a personal level rather than having "the company" make them for you. Many people aren't comfortable with that feeling either.
To a certain extent, starting a company is a very Zen like affair. You need compassion and equanimity. It nicely balances the passion for your customers and the energy to build things. That journey of a thousand miles does indeed start with one step — you just need to be comfortable knowing you're going to stumble in front of people with every fourth step — acknowledge it, be comfortable with it, and simply keep on walking.
 If anyone ever hears of such a study, please pass along the pointer. It would have been published somewhere in the early 1960s.