以配對比較方式探討學系的面試日期選擇

Exploring the selection of interview dates for the department by pairwise comparison

朱榕屏(Lon-Ping Zu)

《管理資訊計算》 10卷1期 (2021/03) Pp. 27-38

https://doi.org/10.6285/MIC.202103_10(1).0003

摘要

本文針對某選定大學的大部分學系，以配對比較方式（配對比較一個較早面試學系與另一個較晚面試學系）探究配對學系在相同學年度的正確篩選率、備取率、以及分發率的差異表現。由於使用相同學年度資料進行比較，考生總人數不會太過干擾比較結果。但是，該結果仍無法避免會受到配對學系先天差異的影響。由於任一個配對比較只牽涉兩個學系，本文執行多個配對比較，然後以次數統計方式判定究竟選擇較早或選擇較晚面試何者比較有利？如此作法的好處是，任何配對學系的先天差異對本文最終判定結果的影響，將隨著配對比較數目的增加而等比例的降低。本文的發現如下：有84.6%的配對比較結果證實較晚面試學系會有較高的正確篩選率；38.5%的配對比較結果證實較晚面試學系會有較低的備取率；以及8%的配對比較結果證實較晚面試學系有較高的分發率。這些發現支持學系選擇較晚面試可以得到有用的資訊協助分辨哪些考生屬於高就讀意願，但是無法確認選擇較晚面試會降低參與學系面試的考生人數。本文的發現亦預期，在考生總人數較多時期，學系無論選擇較早或較晚面試都會有較高的個人申請分發率；而在考生總人數較低的時期，選擇較晚面試學系相對於選擇較早面試學系會有較高的個人申請分發率。

關鍵字

成對比較 ； 先天差異 ； 分發率 ； 正確篩選率 ； 備取率

並列摘要

For most departments of the selected university, this article uses a pairwise comparison method (matching an earlier interview department and another later interview department) to explore the differences between the two departments in the correct selection rate, the reserve rate, and the distribution rate, under the same academic year. Since the comparison uses the same academic year's data, the total number of candidates will not interfere with the comparison result. However, the results of the comparison may be affected by the inherent differences between the two departments (not related to the selection of interview dates but caused by differences in the other factors between the two departments). As a pairwise comparison only involves two departments, this article has to perform multiple pairwise comparisons for the departments (the comparison objects selected for each pairwise comparison will not be the same), and then determine by frequency statistics which is more beneficial to the department to choose an earlier or later interview? The advantage of this approach is that the influence of the inherent differences between any two departments on the final judgment result of this article will decrease proportionally as the number of comparisons increases. The findings of this article are as follows: 84.6% of the pairwise comparison results confirm that the department that interviews later will have a higher correct selection rate; 38.5% of the pairwise comparison results confirm that the department that interviews later will have a lower reserve rate. ; 8% pairwise comparison results confirm that the late interview department has a higher distribution rate. These findings support that the department's choice of a later interview can obtain useful information to help distinguish which candidates are highly willing to study, but it is not certain that choosing a later interview will reduce the number of candidates participating in the department's interview. The findings in this article also predict that when the total number of candidates is large, the department will have a higher distribution rate of individual applications regardless of whether it chooses to interview earlier or later, and that when the total number of candidates is low, choose a later interview departments will have a higher distribution rate of individual applications than those that choose to interview earlier.

並列關鍵字

Pairwise comparison ； inherent difference ； distribution rate ； correct screening rate ； reserve rate

參考文獻

王瑞榮、朱榕屏 (2019)，面試日期改變能否影響個人申請入學分發率-以 X 學系為例，管理資訊計算，8(2)，48-55 頁。 doi: 10.6285/MIC.201909_8(2).0005

高一誠、林健暉 (2017)，招生賽局: 研究所入學考試的撞期分析，經濟論文叢刊，45(3)，495-513 頁。 doi: 10.6277/TER.2017.453.4

Avery, Christopher and Jonathan Levin (2010), Early Admissions at Selective Colleges, American Economic Review, 100, 2125-2156. doi: 10.1257/aer.100.5.2125

Chen, Wei-Cheng and Yi-Cheng Kao (2014a), Simultaneous Screening and College Admissions, Economics Letters, 122, 296-298. doi: 10.1016/j.econlet.2013.12.016

Chen, Wei-Cheng, Yi-Yi Chen, and Yi-Cheng Kao (2018), Limited Choice in College Admissions: An Experimental Study, Games and Economic Behavior, 108, 295-316. doi: 10.1016/j.geb.2017.12.002