Two tables comparing J E L code assignments for Journal of Economic Literature articles using several methods.Top table (a): It has 15 rows and 8 columns, with rows 1 to 3 representing column headers. In row 1, columns 2 to 6 are merged for the header “Assigned J E L codes”, and columns 1, 7, and 9 are blank. In row 2, columns 1 and 2 are blank, and columns 3 to 8 are merged for the header “Based on abstract only”. In row 3, the column headers from left to right are “Article published in American Economic Review Volume 114 Issue 4 (April 2024)”, “Peer-reviewed American Economic Review publication, website”, “Chat G P T 3.5”, “Chat G P T 4”, “Copilot”, “Gemini”, “Average overlap with the final J E L codes”, and “Average difference in the number of J E L codes”. The data from the rows are as follows: Row 4: Column 1: Angelucci, Charles, and Andrea Prat. “Is Journalistic Truth Dead? Measuring How Informed Voters Are about Political News”; Column 2: D 72, D 83, L 82; Column 3: D 83, D 72 (66.67 percent, negative 1); Column 4: D 83, D 72, Z 13 (66.67 percent, 0); Column 5: D 83, D 72, D 91, Z 13, C 83, D 81 (66.67 percent, positive 3); Column 6: Primary: D 83, Secondary: A 12, D 72 (66.67 percent, 0); Column 7: 66.67 percent; Column 8: 0.50. Row 5: Column 1: Engelmann, Jan B., Maël Lebreton, Nahuel A. Salem-Garcia, Peter Schwardmann, and Joël J. van der Weele. “Anticipatory Anxiety and Wishful Thinking”; Column 2: C 91, D 12, D 83, D 91; Column 3: D 81, D 91, C 91 (50 percent, negative 1); Column 4: C 91, D 81, D 91 (50 percent, negative 1); Column 5: C 91, D 83, D 12, D 91 (100 percent, 0); Column 6: Primary: D 81, Secondary: C 91, D 03 (25 percent, negative 1); Column 7: 56.25 percent; Column 8: negative 0.75. Row 6: Column 1: Exley, Christine L., and Judd B. Kessler. “Motivated Errors”; Column 2: C 91, D 12, D 64, D 81, D 83, D 91; Column 3: D 03, D 83, C 91, C 91 (33.33 percent, negative 2); Column 4: C 91, D 81, D 91 (50 percent, negative 3); Column 5: C 91, D 12, D 64, D 81, D 83 (100 percent, 0); Column 6: Primary: D 83, Secondary: C 72, D 02 (16.67 percent, negative 3); Column 7: 50.00 percent; Column 8: negative 2.00. Row 7: Column 1: Grossman, Gene M., Elhanan Helpman, and Stephen J. Redding. “When Tariffs Disrupt Global Supply Chains”; Column 2: D 72, F 13, F 14, L 14, O 19, P 33; Column 3: F 13, F 14, L 14, D 83 (50 percent, negative 2); Column 4: F 13, F 23, L 14, C 63 (33.33 percent, negative 2); Column 5: F 12, F 13 (16.67 percent, negative 4); Column 6: Primary: F 12, Secondary: D 22, L 11, O 11 (0 percent, negative 2); Column 7: 25.00 percent; Column 8: negative 2.50. Row 8: Column 1: Baum-Snow, Nathaniel, Nicolas Gendron-Carrier, and Ronni Pavan. “Local Productivity Spillovers”; Column 2: D 22, D 24, G 32, L 25, R 11, R 32; Column 3: D 24, R 11, R 12, R 32 (50 percent, negative 2); Column 4: R 12, L 25, D 24, R 23 (33.33 percent, negative 2); Column 5: D 22, D 24, G 32, L 25 (66.67 percent, negative 2); Column 6: Primary: R 11, Secondary: C 61, O 40 (16.67 percent, negative 3); Column 7: 41.67 percent; Column 8: negative 2.25. Row 9: Column 1: Greenberg, Kyle, Parag A. Pathak, and Tayfun Sönmez. “Redesigning the US Army’s Branching Process: A Case Study in Minimalist Market Design”; Column 2: D 47, H 56, J 45; Column 3: D 02, D 71, H 56 (33.33 percent, 0); Column 4: D 47, M 21, J 45, C 78 (66.67 percent, positive 1); Column 5: D 47, D 02, D 71 (33.33 percent, 0); Column 6: Primary: D 72, Secondary: D 82, J 41 (0 percent, 0); Column 7: 16.42 percent; Column 8: 0.25. Row 10: Column 1: Cheremukhin, Anton, Mikhail Golosov, Sergei Guriev, and Aleh Tsyvinski. “The Political Development Cycle: The Right and the Left in People’s Republic of China from 1953”; Column 2: D 72, N 15, N 45, N 55, O 21, P 21, P 24; Column 3: O 11, O 47, P 26 (0 percent, negative 4); Column 4: O 11, P 16, Q 15, O 53 (0 percent, negative 3); Column 5: D 71, N 15, O 11, P 21 (42.86 percent, negative 3); Column 6: Primary: O 10, Secondary: N 16, P 16 (0 percent, 0); Column 7: 10.72 percent; Column 8: negative 2.25. Row 11: Column 1: Martínez-Marquina, Alejandro, and Mike Shi. “The Opportunity Cost of Debt Aversion”; Column 2: C 91, D 91, G 51; Column 3: D 14, D 91, G 11 (33.33 percent, 0); Column 4: D 14, D 81, D 91, G 11 (33.33 percent, positive 1); Column 5: C 91, D 91, G 51 (100 percent, 0); Column 6: Primary: G 21, Secondary: D 81, D 03 (0 percent, 0); Column 7: 41.67 percent; Column 8: 0.25. Row 12: Column 1: Alan, Sule, and Ipek Mumcu. “Nurturing Childhood Curiosity to Enhance Learning: Evidence from a Randomized Pedagogical Intervention”; Column 2: D 83, I 21, I 26, J 13, O 15; Column 3: I 21, I 28, O 15 (40 percent, negative 2); Column 4: I 21, I 28, D 91, C 93 (20 percent, negative 1); Column 5: D 83, I 21, I 26 (60 percent, negative 2); Column 6: Primary: I 21, Secondary: O 15, C 9 (40 percent, negative 2); Column 7: 40.00 percent; Column 8: negative 1.75. Row 13: Column 1: Average number of J E L codes; Column 2: 4.778; Column 3: 3.222; Column 4: 3.667; Column 5: 3.778; Column 6: 3.111; Column 7: Blank; Column 8: Blank. Row 14: Column 1: Average overlap with the final J E L codes; Column 2: Blank; Column 3: 39.63 percent; Column 4: 39.26 percent; Column 5: 65.13 percent; Column 6: 18.33 percent; Column 7: Blank; Column 8: Blank. Row 15: Column 1: Average difference in the number of J E L codes; Column 2: Blank; Column 3: negative 1.556; Column 4: negative 1.111; Column 5: negative 0.889; Column 6: negative 1.222; Column 7: Blank; Column 8: Blank. Bottom table (b): It has 15 rows and 5 columns, with rows 1 to 3 representing column headers. In row 1, columns 2 to 5 are merged for the header “Assigned J E L codes”, and column 1 is blank. In row 2, columns 1 to 3 are blank, and columns 4 and 5 are merged for the header “Based on whole article”. In row 3, the column headers from left to right are “Article published in American Economic Review Volume 114 Issue 4 (April 2024)”, “Working paper (presumably chosen by authors)”. “Peer-reviewed American Economic Review publication, website”, “Chat G P T 4, working paper”, and “Chat G P T 4, final article”. The data from the rows are as follows: Row 4: Column 1: Angelucci, Charles, and Andrea Prat. “Is Journalistic Truth Dead? Measuring How Informed Voters Are about Political News”; Column 2: L 82, D 72, D 83, D 90 (100 percent, positive 1); Column 3: D 72, D 83, L 82; Column 4: D 72, D 83, L 82 (100 percent, 0); Column 5: D 72, D 83, L 82 (100 percent, 0). Row 5: Column 1: Engelmann, Jan B., Maël Lebreton, Nahuel A. Salem-Garcia, Peter Schwardmann, and Joël J. van der Weele. “Anticipatory Anxiety and Wishful Thinking”; Column 2: C 91, D 83 (50 percent, negative 2); Column 3: C 91, D 12, D 83, D 91; Column 4: D 83, D 91, C 91 (75 percent, negative 1); Column 5: C 91, D 12, D 83 (100 percent, 0). Row 6: Column 1: Exley, Christine L., and Judd B. Kessler. “Motivated Errors”; Column 2: C 91, D 64, D 91 (50 percent, negative 3); Column 3: C 91, D 12, D 64, D 81, D 83, D 91; Column 4: D 03, D 91, C 91 (33.33 percent, negative 3); Column 5: D 12, D 81, D 83 (50 percent, negative 3). Row 7: Column 1: Grossman, Gene M., Elhanan Helpman, and Stephen J. Redding. “When Tariffs Disrupt Global Supply Chains”; Column 2: F 1, F 13 (16.67 percent, negative 2); Column 3: D 72, F 13, F 14, L 14, O 19, P 33; Column 4: F 13, F 23, D 21 (16.67 percent, negative 3); Column 5: F 13, F 23, D 23 (16.67 percent, negative 3). Row 8: Column 1: Baum-Snow, Nathaniel, Nicolas Gendron-Carrier, and Ronni Pavan. “Local Productivity Spillovers”; Column 2: Blank; Column 3: D 22, D 24, G 32, L 25, R 11, R 32; Column 4: R 12, D 24, C 23 (16.67 percent, negative 3); Column 5: D 22, D 24, R 12 (33.33 percent, negative 3). Row 9: Column 1: Greenberg, Kyle, Parag A. Pathak, and Tayfun Sönmez. “Redesigning the US Army’s Branching Process: A Case Study in Minimalist Market Design”; Column 2: D 47 (33.33 percent, negative 2); Column 3: D 47, H 56, J 45; Column 4: D 47, H 56, J 45 (66.67 percent, positive 1); Column 5: D 47, H 56, J 45 (100 percent, 0). Row 10: Column 1: Cheremukhin, Anton, Mikhail Golosov, Sergei Guriev, and Aleh Tsyvinski. “The Political Development Cycle: The Right and the Left in People’s Republic of China from 1953”; Column 2: Blank; Column 3: D 72, N 15, N 45, N 55, O 21, P 21, P 24; Column 4: O 11, P 16, P 26, N 55 (14.29 percent, negative 3); Column 5: D 72, N 15, N 45, N 55, O 21, P 21, P 24 (100 percent, 0). Row 11: Column 1: Martínez-Marquina, Alejandro, and Mike Shi. “The Opportunity Cost of Debt Aversion”; Column 2: C 91, D 14, D 91 (66.67 percent, 0); Column 3: C 91, D 91, G 15; Column 4: C 91, D 14, D 91 (66.67 percent, 0); Column 5: C 91, D 14, D 91 (66.67 percent, 0). Row 12: Column 1: Alan, Sule, and Ipek Mumcu. “Nurturing Childhood Curiosity to Enhance Learning: Evidence from a Randomized Pedagogical Intervention”; Column 2: N A, N A; Column 3: D 83, I 21, I 26, J 13, O 15; Column 4: I 21, I 28, C 93 (20 percent, negative 2); Column 5: I 21, I 28, D 83, J 13 (60 percent, negative 1). Row 13: Column 1: Average number of J E L codes; Column 2: 2.500; Column 3: 4.778; Column 4: 3.222; Column 5: 3.667. Row 14: Column 1: Average overlap with the final J E L codes; Column 2: 52.78 percent; Column 3: Blank; Column 4: 45.48 percent; Column 5: 69.63 percent. Row 15: Column 1: Average difference in the number of J E L codes; Column 2: negative 1.667; Column 3: Blank; Column 4: negative 1.556; Column 5: negative 1.111. Below the tables, a note reads, “The listed articles appeared in the April 2024 issue (Vol. 114 No. 4) of the American Economic Review which was the current issue as of April 2024, available at: h t t p s colon: slash slash w w w dot a e a w e b dot o r g slash issues slash 775. The first number in parentheses under assigned J E L codes is the overlap with the final J E L codes and the second number is the difference in the number of assigned J E L codes compared to the final ones. Used A I systems: Chat G P T, h t t p s colon: slash slash chat dot open a i dot com slash; Copilot, copilot dot microsoft dot com; Gemini, h t t p s colon: slash slash gemini dot google dot com slash. Data (incl. information about selected working papers) and full responses of A I systems are available in the Appendix slash Supplementary Material”. |
