Is this determined or influenced by what the DB is doing.
No. If the connection drops, Access gives the "disk or network error" and has to be closed -- it is a non-trappable, non-recoverable error. It doesn't always corrupt the back end, since the connection that's dropped may not have been in the middle of an operation that could cause corruption (a read operation is not going to corrupt the back end).
The reason I ask is I have a telemarketing DB that is used across a wireless network and by a few groups. There is typically about 4 to 7 computers on each network in each group. The wireless does drop out but all that happens is "clicking" obviously does not work. Then normally the hub is turned off and back on and sometimes the computers need to be rebooted. But there has never been any corruption problems.
If Access does not detect the dropped connection, it won't give the "disk or network error." If you're not seeing that it would indicate that you've either got a completely unbound app, or are somehow extremely lucky.
I find your report, well, completely implausible. If the network connection is being lost, users would have to force quit Access (because of the "disk or network error") and restart it. If they aren't experiencing that, then the network connection is actually not being lost at all.
But with WiFi, it's very easy for it to be lost, and thus, way too dangerous to ever attempt.
I am assuming one thing, though, that you're not running the app on a Windows Terminal Server. Everything I'm discussing here is in regard to having a front end on the local workstation and a Jet back end on a file server. If you're using SQL Server or some other server database for the back end, or you are running the app on Windows Terminal Server, none of my comments apply, because you're not opening a Jet back end across a wireless connection.