본문 바로가기 주메뉴 바로가기
국회도서관 홈으로 정보검색 소장정보 검색

목차보기

Title page 1

Contents 1

Abstract 3

1. Introduction 4

2. Background Literature 7

2.1. Biases in Editorial and Peer Review Decisions 7

2.2. Artificial intelligence (AI) as a screener and reviewer 10

3. Methods 12

4. Results 18

5. Theoretical Implications 26

6. Discussion and Conclusion 33

References 40

Appendix A: List of the base research papers used in creating submissions 50

Appendix B: Prompts for generating Fake AI papers 69

Appendix C: Prompts for Submission Evaluation 69

Appendix D: GPT-4o mini API Call Structure 72

Tables 42

Table 1. Predicting AI's Recommendations for Review and Acceptance at a Top-5 Economics Journal: Ordinary Least Squares 42

Table 2. Predicting AI's Recommendations for Different Academic Success Outcomes: Ordinary Least Squares 48

Figures 44

Figure 1. Predicted "Top 5 Acceptance Score" scores by authors' characteristics across publication quality categories. 95% confidence intervals based... 44

Figure 2. Predicted probabilities for the outcomes 'Accept/Very Minor Revision,' 'Major/Minor Revision,' and 'Reject/Reject & Resubmit' in the Top-5 Review... 45

Figure 3. Predicted probabilities for the outcomes 'Reject/Reject & Resubmit' and 'Accept/Very Minor Revision' in the Top-5 Review Recommendation Score... 46

Appendix Tables 61

Table 1A. Description Statistics 61

Table 2A. Pairwise Correlations of Independent Variables 62

Table 3A. Predicting AI's Recommendations for Review and Acceptance at a Top-5 Economics Journal: Ordinary Least Squares with Submission Fixed Effects 63

Table 4A. Predicting Top 5 Recommend Accepting by Publication Quality: Ordinary Least Squares 64

Table 5A. Predicting Top 5 Recommendation Categories With and Without Added Criteria: Ordered Logit 65

Table 6A. Multiple Hypothesis Testing: standard and Young p-values 67