Know-How Exchange : Research/Metrics

having done questionnaires/ surveys I would say that its best to evaluate the percentage of respondents responses for comparison. There is no absolute guarantee on responses b/c you have a situation of:
failed, poor, neutral, good, very good
poor neutral good very good
failed poor good very good
poor satisfactory met standards exceeded standards

You see, it depends on how you structured the responses - the language can sway the response - its not necessarily just a comparison of 4pts vs. 5pts. You need to figure out how to divide the data based on the meaning of each point. so in comparitive results you might say....X% = failed/poor, X%very good/exceeded standards.

For the fifth point - if there was a heavy choosing of neutral for example...you point out in your report that when respondents were given the choice of "neutral" they select that response significantly. This can show that the responses are highly emotional and easily swayed by the pitch of the question. This can actually indicate that there is a higher tendency toward the average but that the respondents are hesitant towards being outwardly critical.