I have seen two methods of designing basically the same data model but
they are functionally different. I was hoping to get some insight as
to the strengths of the second model I present below.
Consider a survey with N questions, that have integer values as
responses, and are identified by a unique id for each person.
each respondant has one record with a unique ID, and columns 1 through
N, representing their answers.
each respondent has N records, each identified by a unique id, where
each record is an answer to a particular question on the survey.