It often takes some playing around with the interval widths to figure out which gives the best sense of the
shape of the distribution.
1. What might be a problem with using interval widths that are too large?
2. What might be a problem with using interval widths that are too small?
3. What other considerations might go into choosing the width of an interval?