In-class experiments are similar because they are abstractions (modeling convection currents with heated water and food coloring)
They are different because automation is not as easy, we usually start with a hypothesis rather than looking for patterns that lead to hypotheses
The use of computer models for experimentation is making experimentation easier, more repeatable, safer and more. Now we can take complicated experiments, derive computer models and then have the ability to quickly change the variables.
what an advancement for scientific research!
Computational experiments are a lot less messy than experiments done in the real world. It is also easier to model things such as plate tectonics. It is also easier to show how the experiment interconnects with the real world. With computations science you can have choices being made and can reflect upon those choices.