For a valid experiment, the term should be a control variable on subjects: - 0 years (ie no income, as per current) - 1 year - 5 years - 15 years - life F

For a valid experiment, the term should be a control variable on subjects: - 0 years (ie no income, as per current) - 1 year - 5 years - 15 years - life For example, it's possible subjects knowing they will get zero or one year of BI could have better outcomes than ones who know they have a lifetime of BI. Or not.