self-enhancement biases (Sedikides et al., 2002), makes people attend more seriously to hypothesis-disconfirming evidence (Kruglanski and Freund, 1983), and makes people more self-aware and accurate judges of covariation in their environment (Hagafors and Brehmer, 1983; Murphy, 1994). Other times, “accountability” has perverse effects, increasing defensiveness and escalating commitment to sunk costs (Simonson and Staw, 1992), increasing susceptibility to being distracted by low-diagnosticity cues (Tetlock and Boettger, 1989) and superficially plausible but specious reasoning (Barber et al., 2003), and inducing rather indiscriminant ambiguity aversion (Curley et al., 1986) and risk aversion (Tetlock and Boettger, 1994).
At other times, achieving consensus on whether “accountability” is helping, hurting, or having no effect is impossible because so much hinges on observers’ sympathies and perceptions of whose ox is about to be gored. Examples of such accountability controversies include debates over how to structure relations between auditors and their corporate clients (Moore et al., 2006) or between legislators and citizens (Kono, 2006), how to make teachers accountable for improving student performance (Chubb and Moe, 1990), and how to hold managers accountable for making decisions on race-and-gender neutral grounds (Kalev et al., 2006; Tetlock and Mitchell, 2009). Setting up rules and incentives to encourage desired—and discourage undesired—behavior is a much-discussed problem that is far from fully solved (Kerr, 1975).
Understanding the effects of accountability requires clear definitions for both the independent and dependent variables, as well as a conceptual scheme for characterizing the myriad organizational processes for implementing accountability. On the independent-variable side, accountability is an omnibus term for a complex bundle of variables captured by the multipronged question: Who must answer to whom for what—under what normative ground rules and with what consequences for passing or failing performance standards (Schlenker, 1985; Scott and Lyman, 1968; Tetlock, 1985, 1992)? This definition could apply to virtually any level of analysis, from the societal to the interpersonal. One could hold governments accountable for policy miscalculations; governments could hold intelligence agencies accountable for flawed guidance; agency heads could hold their managers accountable for failure to check errors; and managers could hold individual analysts accountable for making the initial errors. To make this chapter manageable, we focus on the accountability pressures operating on individual analysts in their immediate working environment.
Even within this restricted focus, our definition still allows for enormous parametric complexity. One could hold analysts accountable to colleagues of the same or higher status whose own views are either well known or unknown before analysts submit their work, whose interpersonal style is more collaborative or adversarial, who are focused more on the process