Sociolinguistics 3, Exercise 3: reading the "zee" table

Three questions worth asking (these are my suggestions, not the source's):
  1. How many people were counted in each group? The figures are percentages
     with no sample sizes.
  2. How were the groups chosen, and are the two years comparable samples of
     the same community?
  3. Is the difference a change in progress (younger speakers differ because
     they will keep the form) or age-grading (people change as they age)?
     Two snapshots cannot tell these apart by themselves.

Possible cause given by the source: it hypothesises that an American
children's alphabet song influenced the pattern. That is a hypothesis in the
source, not an established cause.