A simple experiment would be streaking a sufficient number (about 10^8 should likely do the trick, but it depends a bit on the strain) of E. coli on a medium containing penicilin (or a derivative like ampicilin) at a concentration just a little above the MIC. That should give rise to a few resistant colonies. Adding a mutagen (e.g. short exposure to UV) might increase this number (the trick is fiding the right dose). Re-streaking of the resistant cells can then be used to demonstrate that the gained ability is now stable.

A creationist will likely not be swayed by that as they usually not dispute microevolution (hard to argue against something that you can easily demonstrate) but macroevolution. That the latter is just a series of microevolution events is appearantly beyond them. And if challenged, they will usually just make up something new. The nylonase case is a prime example.

https://en.wikipedia.org/wiki/Nylon-eating_bacteria_and_creationism

Have a look at John Endler's classic "Natural Selection in the Wild" in the Princeton series. 

It would be a bit more work, but you could have your class examine selection against a common mutational strain of Drosophila melanogaster. In my classes, we first create F1 hybrids between mutants (ebony or vestigial) and then follow the population through 5 or more generations in a large population cage. The frequencies of the ebony and vestigial alleles decline from 0.5 over time. Both traits are recessive. Allele frequencies can be estimated if one assumes Hardy-Weinberg equilibrium. One can even estimate selection intensities from a long series. This would be a semester-long experiment. Otherwise the suggestion to use bacteria is a good one.