Couple of comments.

1) If a door is pulled on a newer airframe - the plane in grounded for a short period. If a door is pulled on an older airframe (hours) - the plane can be grounded for a LONG time. Movement in the frame over time can cause the reseat to be a difficult task (perfect example is the old shuttle 727s - whose high rate cycles and heavier landing gear tend to stress the airframe more). In any case - this is a loss of the aircraft for a period of time. Considering in its best year, DL made $210/flight profit, but would loose $2600/delayed or cancelled flight - its easiest to say they want to keep them in the air and ON SCHEDULE (DL being a poor example but it makes the point of the usage).

2) The 'lack' of knowledge of 'pulling' is actually 'planned'. Doors vary slightly in weight - but in almost all cases - when a pax pulls - they are surprised by the mass/size combination, but the simplified instructions and urgency tend to have good results. In many cases the smaller person (height) will have greater difficulty in unseating the door than a taller person. Dumb luck it worked that way.

3) I've had the opportunity to 'pull' several times on grounded or retired planes - usually in the maintanence yards (function of being an invitee to some tour or another - been on several) - and once for real (but only due to a pair of faulting doors on a 73 combined with an country reg requiring evacuation of all pax in a declared emergency and the fact the strip didn't have any way to get us down), which entailed hours of interviews afterwards. The airlines are VERY protective about the data from pulled doors - and the systems work - so the idea of training isn't likely - as those who receive training will be at a 'calmer' statute versus the planned or managed chaos - this chaos benefits the ideal that you 'simply' pull the door and get off the plane.