個案公司為台灣一家以光電產業為主的高科技公司，2004年曾經針對所採用的甄選工具「感知與偏好量表」進行效度的驗證。結果顯示各職業別高低績效的員工在人格特質上並沒有顯著的差異，也因此產生實務上的懷疑。所以本研究主要的目的有二：第一是是考驗感知與偏好量表與主管評核績效之關聯性；第二是比較研發工程人員與非研發工程人員高低績效在人格特質上的差異性。 本研究採用2006年工作績效作為效標，進行預測性效度的檢驗。研究發現感知與偏好量表解釋總變異量為5.5%，其解釋總變異增加了1.8％並達到顯著的水準(p=0.039<0.05)。其中生活步調達到顯著水準(標準化β係數為0.12，t值為2.44，p值<0.05)。 研究進一步探討全體員工、行銷研發人員、非行銷研發人員高低績效員工人格特質有哪些顯著的差異。結果發現全體員工績效特優組的成就需求得分顯著較高（平均差=1.09；t=2.53；p<0.05）。但是在細緻性上績效特優組顯著較低（平均差=-1.18；t=-1.96；p<0.05）。研發工程人員在活動性、成就需求、喜新性等構面上，績效特優組的得分顯著高於績效落後組（平均差=1.72；t=2.47；p<0.05；平均差=2.05；t=3.51；p<0.01；平均差=1.13；t=2.10；p<0.01）。但是在細緻性上績效特優組則低於績效落後祖（平均差=-2.19；t=3.20；p<0.01）。至於非研發工程人員高低績效在人格特質上則沒有顯著的差異。 Case Company is a High-tech Company mainly focused on the optoelectronic industry. The selection tool which Case Company is now adopting called PAPI (Perception and Preference Inventory). They had an investigation on in 2004 and the validity turns out to be low. No evident reveals that there is any significant difference between high performance employees and low performance in any career. As a result the study has two purposes: First, Using 2006 performance evaluated by supervisors as criterion to discuss the predictive validity of PAPI (Perception and Preference Inventory). Second, to investigate whether employees have significant difference in personality between high and low performance employees. The study proves that PAPI do have significant predictive validity(total variance 5.5%;incremental variance1.8％,p=0.039<0.05).The construct “Life pace” is a useful predictor (standardβ=0.12,t=2.44，p<0.05). At the whole employees’ part, high performance employees have significant higher scores in the following personality constructs: “Achievement Needs”（mean difference=1.09；t=2.53；p<0.05）.High performance employees also have significant lower scores in “Attention to detail” (mean difference=-1.18；t=-1.96；p<0.05) constructs. At the Research engineering employees’ part, high performance employees have significant higher scores in the following personality constructs: ” Need for change”,” Achievement Needs”,” Work pace”（mean difference =1.72；t=2.47；p<0.05；mean difference =2.05；t=3.51；p<0.01；mean difference =1.13；t=2.10；p<0.01）.High performance employees also have significant lower scores in “Attention to detail” (mean difference=-2.19；t=3.20；p<0.01) constructs. At the Non-Research engineering employees’ part, there is no significant difference between high and low performance employees.
