Research code has different values. If you throw away 90% of the code you used after a week you'd code differently too!

The tricky part is that 10% of your code will become the basis of your thesis / postdoc;)

That hasn't generally been my experience, unless you qualify researchers as 'university students'.