A programming firm had developed a more elaborate, more complex version of its spreadshe...

A programming firm had developed a more elaborate, more complex version of its spreadsheet program. A “beta test” copy of the program was sent to a sample of users of the current program. From information supplied by the users, the firm rated the sophistication of each user; 1 indicated standard, basic applications of the program and 3 indicated the most complex applications. Each user indicated a preference between the current version and the test version, with 1 indicating a strong preference for the current version, 3 indicating no particular preference between the two versions, and 5 indicating a strong preference for the new version. The data were analyzed using JMP IN. Partial output is shown here.

a. Do the ROW PERCENT entries suggest there is a relation between SOPHIST and PREFER? If the data showed no relation, what would be true of the ROW PERCENTS?

b. Does the (PEARSON) CHI-SQUARE computation indicate there is a statistically detectable (significant) relation at usual values of α ?