Demonstrating a normal curve with foam plastic pieces
At 8:00 I tossed 409 plastic pieces on the floor, and 9:00 only 351 pieces. Both distributed reasonably normally, although of late I have become concerned that the data is only normally distributed in a radial or circular sense, two-dimensionally, and that the collapsing of the counts to a single dimension leads to a leptokurtic distribution.
The distributions are heap-like and demonstrate some of the basic features of a normal distribution - a central peak, two tails, roughly symmetric.
The add-on produces a normal quantile plot for frequency, confirming that the distributions are not normal. I suspected that the plots ought to be leptokurtic and that does appear to be supported by a QQ plot for the frequency data.
The beads do not spread sufficiently into the tails to be a true normal distribution. As a way to introduce the normal distribution visually in the classroom, the beads suffice. A Galton board might be better, but I want the normal distribution to arise whole cloth from randomness of physical objects. The Galton board could be conceived by the students to be a special case. Plastic pieces tossed in the air really seems like a random activity.