In a latin square experiment, with 4 different golf-ball designs (A,B,C and D); if we suppose that A is the standard golf-ball design and the other three designs are newly developed. Which statistical method must we use to compare the three new designs with the standard design?