VPN connects the two networks so you can pass IP traffic from one to the other. period. You are not necessarily remote controlling a computer or server at the host site (like Logmein or gotomypc or terminal services). There are reasons why you may still want to do remote control or term server sessions through the VPN tunnel, but that's a discussion separate from VPN.
There are 2 ways you can set up the VPN, either site-to-site, or client/server.
With a site-to-site VPN, the VPN tunnel is established directly between the routers at each site, and therefore any network device at either site can communicate with any network device at the other, unless restricted by ACL. The tunnel is always up and IP routes through the VPN tunnel from site to site, as if you had two separate IP networks routing to each other through a single router. Internet access from either site is via the site's own internet connection.
With a VPN client/server setup, the router at the main site, i.e. the office acts as a VPN server, then you launch a VPN client on the individual computers at the school from which you want to access office resources. In this scenario, the VPN tunnel is established between an individual workstation at the school with the office site, and so only the workstation that establishes the VPN client connection has access to the other site, and the tunnel is only up when the user launches the VPN client.
There are pluses and minuses to each scenario. I don't think that either method is more or less difficult to set up; the site-site is generally more reliable, but at the same time, being in a school environment you're potentially exposing the office network to any of the student workstations, again, unless you'll be enforcing ACL's. The client/server VPN may be somewhat less reliable (or not), and you have to train the user to establish the connection before accessing the office site resources, but you can turn it on and off when needed, and you can install the VPN client on only the workstations that need to access the office resources.
Now once you've established your VPN tunnel, however you choose to do it, you can pass IP traffic. period. What you do with that IP is up to you. In other words, as I mentioned earlier, whether you remote control a resource at the office site versus using the remote site workstation as a client of the resources across the tunnel, has nothing else to do with VPN.
For this decision, you have to factor in
a) what is the bandwidth between the sites (you mentioned a T1 at each)
b) what is the application or resource being accessed across the tunnel?
c) does it lend itself to being run across a WAN, and will performance be adequate across the skinny tunnel? Remember, a T1 is only 1.544kbps compared to local network speeds of 100mbps or 1gbps, so even a full T1 is skinny compared to local network speeds; do the math.
d) what is the tolerance of the application or resource of unexpected disconnects?
and more.
The answers to these questions will help you determine if you want to run as a full client across the tunnel or remote control a resource at the office site, such that network connectivity of the app or resource is local and you're merely sending console info, i.e. keyboard/video/mouse information across the tunnel. Generally, remote control requires much less bandwidth and therefore performance is very good, especially with T1 bandwidth, but of course the drawback is having to dedicate and install terminal services.
I rambled for a bit; hope this is helpful. Post back specific questions if you want more.