In many countries, overall animal experimentation is not significantly decreasing or becoming less severe. Does this show that these countries’ programs to promote alternatives and the “three Rs” of “replace, reduce, refine” are failing? Scholars and activists sometimes take this for granted, but representatives of “three Rs” programs have disagreed. This article makes two contributions to the debate: one conceptual and one normative. First, it draws attention to the distinction between evaluating impact (whether a program makes a difference) and evaluating sufficiency (whether a program makes enough of a difference to achieve its goals). Total numbers are typically unhelpful in assessing impact, but depending on goals, they can be relevant in assessing sufficiency. Second, this article argues that an overall decrease in harm to animals in experimentation is a sensible policy goal. This article concludes with suggestions for how to go beyond the “three Rs” to effect overall change.

Animal experimentation raises value conflicts between animal protection and other goods, such as freedom of inquiry or health and safety. If governments can phase out the practice by non-prohibitive incentive-setting, the pro tanto moral rationale for doing so is obvious. So why should they not? This article first sketches a fictional scenario in which a government adopts a phase-out plan for animal experimentation. It then considers two moral objections to this plan: First, the plan unduly restricts freedom of inquiry, and second, it merely displaces animal experimentation across borders and thus fails to reduce animal suffering. Both arguments are refined premise by premise to articulate their strongest versions. The two objections can help to narrow down desiderata for good phase-out plans. However, they do not provide a compelling case against phase-out planning as such because they miss its incremental and constructive nature. Unless better arguments can be provided, it appears that government inaction on phasing out animal experimentation lacks moral justification.