Counterbalancing

subject-by-subject & across-subject counterbalancing

counterbalancing < (2&)

subject-by-subject counterbalancing

controls progressive error for each subject by presenting all treatment conditions more than once

subject-by-subject counterbalancing

disadvantage: longer time, expensive, or boring procedures

subject-by-subject counterbalancing

its disadvantages is compounded as the experimenter increases the number of treatments

reverse counterbalancing / block randomization

subject-by-subject counterbalancing < (2/)

reverse counterbalancing

subject-by-subject < administers treatments twice

reverse counterbalancing

subject-by-subject < ABBA

reverse counterbalancing

subject-by-subject < used for 2 levels only

block randomization

subject-by-subject < researchers assign each subject to several complete blocks of treatments

block randomization

subject-by-subject < used for 3+ levels

across-subject counterbalancing

presents each treatment ONCE and controls progressive error by distributing it across all subjects

complete & partial counterbalancing

across-subject counterbalancing < 2&

N factorial

complete counterbalancing AKA

complete counterbalancing

across-subject < uses all the sequences an equal number of times

complete counterbalancing

across-subject < determine how many possible sequences by using factorials (!)

(number of treatment conditions)!

partial counterbalancing

across-subject < we present only some of the (N!) orders

randomized partial & latin square counterbalancing

across-subject < partial counterbalancing < (2&)

randomized partial counterbalancing

across-subject < partial < used when there are 5+ treatment conditions

randomized partial counterbalancing

across-subject < partial < may not control for order quite effectively compared to complete counterbalancing

randomized partial counterbalancing

across-subject < partial < all the possible sequences are put in where the researcher could randomly pick from it and assign the picked sequence to a participant

randomized partial counterbalancing

across-subject < partial < subjects are assigned with randomly selected sequences

latin square counterbalancing

across-subject < partial < each treatment appears only once in any order in the sequences

latin square counterbalancing

across-subject < partial < done with 4 treatment conditions only

latin square counterbalancing

across-subject < partial < may have carryover effect