The primary advantage of the Latin square design, compared to the factorial design, is that

"

 In most circumstances, it requires less data"

"

 It eliminates the need for interaction analysis."

"

 It allows higher significance levels."

"

 It does not require homogeneity of variance."