【正文】 ds are behaving effectively from a shareholder perspective, including studies of CEO pensation, the adoption of poison pills, adoption of greenmail and other antitakeover provisions (. Conyon and Peck, 1998。 Frankforter et al., 2000). However, we know of no previous studies which have been able to directly measure board effectiveness from inside the boardroom in a way that might be used to directly examine the theoretical ‘black box’ between board practices or attributes and performance. Given the number of previous studies that have tied many of the aforementioned board attributes to pany performance, we expect that board effectiveness will actually act as a mediator of the board attributes–performance relationships that have been extensively studied by governance scholars. Moreover, we suggest that examining these attributes together, based on theory derived from research on effective teams, has the advantage of potentially integrating this vast set of studies into a more coherent story of how the practices work together to support better corporate governance. For example, Roberts et al. (2005) argue that a prerequisite for effective boards is for directors to bee engaged in carrying out their various responsibilities to the pany. This requires a significant time mitment to learn and keep up with the various intricacies of the pany’s operations and time is a serious constraint for most directors (Carter and Lorsch, 2004。 Lorsch and MacIver, 1989). Shen (2005) supports this argument but suggests that without appropriate incentives in place the likelihood of directors willingly putting in the time to adhere to their duties is low. This mediating role of board effectiveness is implied in previous boards research although not empirically tested. For instance, Golden and Zajac (2001) show that board demography and processes significantly affect strategic change. This essentially implies that board attributes impact board effectiveness, leading to higherquality strategic decisions and higher levels of organizational performance. Similarly, Forbes and Milliken’s (1999) prehensive theoretical model shows how board task performance (. board effectiveness) acts as a mediator, or intervening construct, between board processes and firm performance. Further, they argue that board characteristics are antecedents to task performance. Following these arguments and the many previous studies on corporate governance demonstrating the importance of boards of directors to the successful operation of the corporation (Baysinger and Butler, 1985), we propose two more hypotheses: Hypothesis 1: Corporations with more effective boards will demonstrate higher levels of financial performance. Hypothesis 2(a–e): Board effectiveness mediates the relationship between board attributes – (a) knowledge, (b) information, (c) power, (d) incentives, (e) opportunity/ time – and the financial performance of the pany. Taken together, the seven hypotheses discussed above test the prehensive model that boards with sufficient knowledge, information, power, incentives, and opportunity/ time are more likely to be effective, and more effective boards will be associated with superior corporate performance. Figure 1 visually demonstrates the overall model.Sample and Data CollectionTo minimize the effects of method variance, each stage of the hypothesized model was operationalized using various types of survey and/or archival data. The data for some of the board attributes and the effectiveness measure were drawn from an annual written survey of the directors of all Fortune 1000 panies conducted by Korn/Ferry International, an international executive search firm, in conjunction with the Center for Effective Organizations at the University of Southern California. Data were taken from the survey of 1996 boards in which 1151 chief executives, inside directors, and outside directors responded. Survey data were then matched with archival data on boards taken from proxy statements filed with the Securities and Exchange Commission (SEC), the Institutional Brokers Estimate System (IBES), and the Investor Responsibility Research Center (IRRC), which publishes detailed listings of corporate governance provisions for individual firms. The IRRC data are derived from several sources, including corporate bylaws and charters, proxy statements, annual reports, and 10K and 10Q documents filed with the SEC. Finally, pany performance data were taken from the COMPUSTAT database for years 1995–98. In order to match the survey data with archival information, pany names were recorded for as many respondents as possible using two different methods. Directors were asked to give the name of the pany for the board that they described in the survey, but this question was strictly optional. Second, pany names were taken from the return address on the envelopes in which the director returned the survey. Due to specific instructions on the survey, this method was only used when the respondent selfidentified as an inside director of a firm.[2] This process originally identified 420 board members from 335 different publiclytraded panies.In cases where there were multiple respondents from the same pany, one was chosen at random to represent that particular board. There were 71 panies for which there were two or more respondents. Fiftyseven of these had only two respondents, with a maximum of five directors responding from a single board. An evaluation of the interrater reliability between respondents from the same pany was made with the 71 boards that could be identified as having multiple respondents. Paired sample ttests indicated that one group of 71 responses could not be distinguished from a second group. Overall, directors from the same pany agreed 64–84 per cent of the time, depending on the particular question. There were 39 panies for which plete data were available for all 11 items used to measure board effectiveness