從推敲可能模式的觀點探討共同性指標之偏誤－－以平衡計分卡為例

Adopting the Perspective of Elaboration Likelihood Model to Discuss Common Measures Bias-Using Balanced Scorecard as an Example

本研究主要探討管理人員在以平衡計分卡作為評估部門績效標準時，是否可透過確信報告（如：協議程序執行報告）的提供，而降低共同性指標偏誤的問題或者是減少判斷不一致的問題。此外，我們將接受資訊者分成高或低涉入者作為探討。因此本研究主要研究問題為涉入程度高或低的管理人員是否會因為接受的資訊不同，而對於所做出的決策判斷有不同的結果？研究結果顯示當以分階段評估方式進行績效評估時，各組皆無偏誤產生。在獎酬分配部分，增加協議程序執行報告的提供將有助於降低偏誤。此外，魏克森符號檢定之下則發現（1）高涉入者不論是否有其他資訊的提供，都能夠做出無偏誤的績效評估，（2）低涉入者在獎酬分配部分若是沒有周邊資訊的提供，將會有共同性指標偏誤情況發生。

The research questions of this study include: (1) whether the superiors who receive different information about the BSC would make different decision or not. (2) Whether providing assurance report would reduce common measure bias or not. The results indicate that disaggregated strategy makes no common measure bias in all experimental groups. In the decision of bonus allocation, providing assurance report can decrease common measure bias. Besides, the study adopts Wilcoxon’s Signed-Rank Test to find that (1) high involvement subjects can make performance evaluation judgments in no bias without assurance reports, and (2) low involvement subjects will make common measure bias in allocating bonus when they don’t receive peripheral cues.