The constitutional issues presented by the use at religious institutions of government-financed vouchers for social services have been addressed repeatedly, but virtually all of the writing in the field focuses exclusively on educational services. Most of that writing, moreover, takes the overly simple view that voucher programs are linear relationships running from government to recipient to service provider. Proponents of such programs argue that the presence of recipients as intermediate decision-makers breaks the link between the state and religion, and therefore solves the constitutional problem of church-state connection. Opponents argue that use of intermediaries is nothing more than money-laundering, and that the Establishment Clause precludes transfers of funds that originate with the state and end up in the accounts of religious organizations. In this paper, we expand the angle of constitutional vision in three important ways. First, we clarify the theoretical positions on the Religion Clauses that lead commentators, including ourselves, in particular directions on the voucher question. Second, the paper expands beyond the context of education; although the paper concludes with an analysis of the Cleveland school voucher case, now pending before the Supreme Court, most of the paper discusses other examples of voucher programs, including those devoted to child care and treatment for substance abuse. Third, this wider look at voucher programs leads us to analyze voucher arrangements not as a straight line, but instead as a triangle of relationships. Every voucher program involves relationships between government and recipients, government and providers, and providers and recipients. Examining the relevant features of those relationships leads us to a typology of voucher programs, in which we systematically connect programmatic features - in particular, the degree of government compulsion imposed, and personal transformation sought, by the state - to relevant constitutional inquiries into 1) the required mix of religious and nonreligious service providers, and 2) the problem of government monitoring of religious entities. The typology addresses, among other things, the placement of various substantive obligations with respect to the crucial constitutional issues. We conclude the paper by using our proposed typology to analyze the Cleveland voucher case. Although monitoring presents no problem in the case, the degree of compulsion and child formation present in the scheme should require the state to demonstrate that it has made a good faith effort to ensure the adequacy of the mix between religious and nonreligious alternatives open to Cleveland parents. Because the state failed to take important and obvious steps to meet this obligation, and because the defenders of the program cannot persuasively demonstrate that the steps in fact taken by the state are sufficient, the Supreme Court should not uphold the existing Cleveland program. Our analysis suggests, however, that properly designed voucher programs, educational and otherwise, can satisfy the Constitution. [Although we completed work on the paper prior to the Court's decision in the Cleveland voucher case, we have added a brief prologue explaining the continued relevance of our analysis to future decisions involving vouchers and various government services.]