Laments have grown in recent years over the increasingly widespread belief that most policy analysis is now being driven by the service of specific, powerful interests. Three types of responses are apparent: On the right in particular, interest groups have supported the development of "think tanks"; serving particular ideological perspectives, which nurture the development of a cadre of "hired guns"; to develop analytic resources to support established views. A second response is typified by the recent report charted by the minority party in the House, criticizing the politicization of analysis and presumably seeking a return to the world of "neutral expertise". The third response is found in the academic literature, where critics urge the development of a broader discourse (including unrepresented and under-represented groups) and information supporting neglected perspectives and voices (e.g. Hajer, 2003). We can infer a larger critique from these efforts -- that policy analytic tools, and the substantial resources of information and understanding that they represent, are skewing policy deliberation. We argue that these laments offer an incompletely specified response to the defective political process that they have identified. In answer, we offer what we call "the analytic contract" model of policy analysis which we believe offers an approach to analysis that can resolve the concerns that these critics have identified. The model has two important strengths: it prescribes a process that both enhances and focuses policy deliberation; in ways that should enrich participation, and it offers analysts a clearly prescribed role that relies on the strengths of a professional community. Briefly, the model envisages a several stage process. At the outset, an interested party or parties identifies the need for an analysis, perhaps in response to exogenous triggering events, or perhaps not. These interests, together with other interested parties, can then be expected to advertise the need for analysis. This leads to the "contract phase -- in this phase, a "professional analyst" is invited into the process, perhaps with some funding or other support, or perhaps operating on internal resources. The analyst (which may be an individual or may be an organization) then seeks to formulate an "analytic contract". They begin the process by developing a stakeholder survey: this involves the identification of relevant stakeholders (others, e.g. Dunn, 2003, have explored some of these issues, albeit not conclusively), and the development of an understanding of their interests. In this phase, the "policy analyst", operating as a professional, is responsible for identifying relevant stakeholders and incorporating them into the process. We argue that this process offers some hope of overcoming "moblization biases", by charging analysts to operate as professionals, accountable not only to recognized interests, but also to a professional community for their responsibility to identify and incorporate the interests of those interested, but not effectively mobilized. Once this stakeholder community has been identified, the policy analyst then moves to the development of an analytic contract -- but in consultation with the broader stakeholder community that has now been identified. In keeping with Lowi’s insight of many years ago, we argue that the crystallization of this process into a March and Olsen-like "choice opportunity" (as opposed to the more typical evolving soup of policy windows and deeply ambiguous moments of opportunity) improves the chances for the incorporation of poorly mobilized and thinly supported interests. This process is related, in ways to be explored in greater length in the paper, to the literature on democratic discourse and deliberative democracy. The analytic contract that is ultimately negotiated through the facilitation of the professional analysts specifies the questions to be investigated and the data and information resources to be employed (including any relevant research design issues). The analytic contract may, quite plausibly, address multiple and not necessarily commensurate objectives -- it is intended, after all, to "improve the opportunity for the exercise of informed judgement" as E.S Quade said, not to determine policy. Once this contract has been developed, the analysts is then prepared to develop funding for the work envisaged. The analyst remains responsible to the professional community for an appropriate execution of this analytic contract, but this does not preclude the development of support for the work that is more partial and self-serving (so long as the entire contracted exercise is undertaken). This process recognizes that not all analysis will prove "successful", nor will it yield unambiguous results. It seeks to offer "objective analysis"; in an "accountability framework"; (as Landau observed years ago, "objectivity is accountability"; "offering fair controls over assertions, not pristine, unambiguous facts). Analysts are accountable to the community of stakeholders (for the deliberations over the identification and clarification of their sometimes distinctive interests) and to the community of professional analysis for the appropriate development of a fully realized map of the stakeholder universe; for the development of a well articulated analytic contract, in response to their interests; and for the appropriate execution of the terms of this contract. The professional community, then, provides the validation of these analysts -- those operating without this imprimatur, then, are appropriately understood to be engaging in a different enterprise. [ABSTRACT FROM AUTHOR]