招生選才過程中不同甄審項目的實質影響力

Real vs. Claimed Weights of Different Assessment Procedures in the College Admission or Personnel Selection Process

招生選才的甄審項目經常包含客觀筆試及主觀評審，但招生簡章中公布的占分比例其實往往只是假象。本研究以電腦模擬方式實驗前述現象，並透過一些實際案例的招生資料作為佐證。實驗操弄的自變數包括：（1）主觀評審項目的不同評分範圍；（2）評審間的共識度；（3）評審項目占最後總分的比例，共形成3×4×3＝36種實驗情境。模擬的客觀筆試成績範圍為0～100的常態分配隨機變數，主觀評審成績則再依上述操弄情境加以轉換。實驗結果發現：主觀評審項目對總成績的實質影響力，大部分低於預設的占分比例，並隨著評審評分範圍的縮小及評審共識程度的降低而更趨嚴重。此外，本研究針對2009年台灣北區一所大學十五個學系的大學甄選分數及十五個碩士班的招生入學分數，進行兩個實際案例分析，結果發現「主觀或客觀」不是關鍵，甄審項目的分數範圍才是；成績分散範圍較廣／區辨力較強的甄審項目之實際影響力較大，且可能導致招生或選才簡章中公布的占分比例成為假象。本研究的結果可提供招生選才機構思考評分機制、計分程序及選才效果等問題，並提供應試者參考。

College Admission or Personnel Selection Process usually adopts both subjective and objective assessment procedures. The weight of each procedure might be claimed publicly， but it often did not reflect its real influence on the final grade. The current study simulated this phenomenon in the computer and employed data of real cases to investigate some factors behind the phenomenon. In the Monte Carlo study， the following variables were manipulated: 1. ranges of subjective rating scores; 2. levels of inter-rater consensus; 3. weights of different assessment procedures. A total of 36 experimental conditions were formed， and 20 replicated samples were generated in each condition. Normally distributed random numbers between 0-100 were generated as the scores of the objective assessment procedure， and they were transformed according to the experimental conditions to form the scores of subjective assessment procedures. Results showed that real weights of subjective procedures were lower than claimed weights of them in most conditions. The phenomenon became acute when narrowing the rating range and reducing the degree of inter-rater consensus. Some analyses were also performed on admission scores of year 2009 from real cases， including 15 departments and 15 graduate schools from one university at north Taiwan. Results indicated that the property of objectivity or subjectivity of assessment did not play an important role. The key factor was the discrimination power of assessment procedures. Those results have some implications for school admission or personnel screening process as well as for applicants in the process.