Working in groups to analyse separate sections of the text may be a useful whole class activity even if pupils then work in smaller groups to analyse the data.
By sorting the letters by time values you can get an approximation to Morse's findings.
A simple frequency count would give a first clue to differences. Are there any letters whose frequency is significantly different when the two distributions are matched?
Using the new code and translating is there a "significant" difference in the times the messages would take?
By weighting letter time values according to their frequency it should be possible to see whether the average time taken to send a message using each of the two systems, or between systems generated by groups of pupils, differ.