I am looking for some sort of reliable (i.e. I need numbers),
preferably scholarly, research on the reliabality of a web-based
survey in a proctored environment rather than at home.

For example:
We are preparing to administer a web-based survey to approximately
1000 students on a college campus.  We would prefer to have the
students just take the survey at home.  However, if necessary, we can
reserve a computer lab on campus, and administer the survey in a more
"chaperoned" environment.

What I am looking for is some sort of information regarding whether
one would be more reliable.  If studies show that the information
obtained from a user taking the survey at home isn't reliable, then we
will reserve the spot on campus.

One note on the surveys---they are not anonymous.  The users are
required to create an account using their email addresses and their
student ID's.

Thanks!!