Speaking of the types of reactor, there is one crucial difference between the graphite core and pressurized water reactors: passive safety. That is, what does it do uncontrolled, that is, if cooling fails? The answer: a pressurized water reactor stops, a graphite core reactor goes Chernobyl. Only Russians, North Koreans and Americans still use these potentially dangerous reactors.

Two things make up the basis of security of the reactor: the cooling system and the control rods. A nuclear reaction can produce power only with a moderator - without it, the reaction stops. The moderator is either graphite or pressurized water. In a pressurized water reactor, the same water acts also as the coolant. The control rods are used to slow or speed up the reaction, and using them the reaction can be stopped.

Now, if the cooling system and the control rods fail simultaneously, the reaction is too fast, and core will start to heat up. A graphite core will eventually catch fire, but the moderating water will start to boil. Steam is not a moderator, so the water reactor stops and doesn't produce heat. If the coolant drains away, the graphite core burns, but having no moderator, the water reactor stops.

No one could not blow up a pressurized water reactor with an explosion like in Chernobyl. Graphite core reactors are a risk.

Whatever happens, nuclear explosion does not happen in a power plant. The uranium is far too impure (3-5% 235U) to explode. It can overheat, but it never explodes like a nuclear bomb. If something explodes, it's because of the good old pressure and fire. This happened in Chernobyl. In any case, the radiating matter will not poison severely anything except the immediate neighbourhood of the reactor. The radiation released is usually only harmful, not lethal like from a bomb. However, the fallout of an exploded reactor has a longer half-life, making it dangerous for a longer time.

It must be acknowledged that in a nuclear fuel processing plant a failure can lead to an uncontrolled chain reaction, which radiates a miscellany of ionizing radiation, which can be harmful to the workers. Processing plants are also not as well protected as power plants.

A little reply or addition to DejaMorgana below: I've noticed that most people assign these three things together: nuclear blast, radioactive contamination and breaking the reactor shielding. These are very different things. Crashing a jet plane to a reactor dome could break the shielding, but the contamination would be very small, and certainly there wouldn't be any nuclear blasts. There will be no bomb-like blasts in reactors, because reactor-grade uranium is too impure. It is, however, poisonous, but even this is not excessively dangerous.

Remember that no matter if the contents of the reactor are radioactive, they are still poisons, not explosives. Poisons, no matter how poisonous, are dangerous to a certain extent only. The concept of the degree of severity applies to radioactive poisons as much as it applies to simple chemical poisons. It can be measured and controlled. Radioactivity is not magic. Radioactive contamination is not a supernatural evil force that permeates all barriers, but just another type of contamination!

Drinking toilet cleaner is crazy, and sprinkling it to food is equally insane. However, these are the only ways to make toilet cleaner lethal. It is possible to burn yourself with it, but that isn't likely to be lethal. There are definite ways to clean the poison. This applies to all poisons: poisons are not evil. They have their hazards, but these are predictable and well experimented. Radioactive ones are not an exception.

So, a jet plane crashing to a reactor dome is far from a worst case scenario. The likely toxicological effect is a small local contamination, which is immediately poisonous only to the people who clean the site. The likely consequence in the worst case scenario is the limitation of agriculture within a 20-km radius. You should be more concerned about why the jet would crash in the first place.