Antecedent choice and consequence choice procedures are often used as interventions to increase on-task behaviour and reduce problem behaviour. This systematic literature review considers the conditions under which individuals show a preference for choice. Results suggest that preference for choice is variable, with some individuals preferring choice which allows access to more highly preferred stimuli and others preferring choice regardless of the outcome's preference value. It is recommended that choice be used as a non-invasive intervention which may positively affect some individuals' behaviour by either allowing them access to more highly preferred items or by accommodating their preference for choice. Trends across different samples and procedures are considered and recommendations are made for future research.