I am not sure is it a good ideas to add coefficient of restitutions between the balls.

It will make the simulation more like real case, but it will make it more complicated.

Now, students can check that energy is conserved.

For elastic collision: The velocity after collision could be calculated as

V1'=(2*m2*V2)/(m1+m2)+(m1-m2)*V1/(m1+m2);

V2'=(2*m1*V1)/(m1+m2)+(m2-m1)*V2/(m1+m2);

Actually, for educational purpose, it will be better to ask:

how to modify the above equations if we want to add coefficient of restitutions between the balls. :D

I can add coefficient of restitutions between the balls easily, and if you give me information about the mass/size/... and the experiment data for the real case, may be we can try to find good parameter sets to fit with your data.

I would suggest you make a video of the experiment and convert the movie file to quicktime format, then, those data can be analyzed with tracker and compared with ejs simulation.