Leaving aside the feasibility of providing indefinite basic income as part of the study, it doesn't necessarily follow that the results would take longer to ach

Leaving aside the feasibility of providing indefinite basic income as part of the study, it doesn't necessarily follow that the results would take longer to achieve. The author's point, I think, is that people would behave differently if they were confident that the income would last forever, compared to only lasting 5 years. It's entirely possible that you could still get useful data within those first five years though. (And almost certainly more useful than in the time-limited case, because you wouldn't have to account for behavioural differences.)