It is known that a two-factorial design is balanced iff each combination of the level of the two factors have the same replication. ie A has 2 levels, B has 2 levels then the replication for treatment A1B1,A2B2,A1B2,A2B1 are all the same. But, how is balanced considered inside an ANCOVA design? For example, if I have a model with a factor A of 2 levels and one continuous covariate,x. Does it mean that the design is balanced if A1 and A2 have the same replication(ignoring the continuous covariate)?