我们应该信任R2和调整后的R2多少：来自顶级经济学期刊和蒙特卡洛模拟的回归证据

IF 1.4 4区经济学 Q3 ECONOMICS

Journal of Applied Economics Pub Date : 2023-05-02 DOI:10.1080/15140326.2023.2207326

Qiang Chen, Ji Qi

{"title":"我们应该信任R2和调整后的R2多少：来自顶级经济学期刊和蒙特卡洛模拟的回归证据","authors":"Qiang Chen, Ji Qi","doi":"10.1080/15140326.2023.2207326","DOIUrl":null,"url":null,"abstract":"ABSTRACT R2 and adjusted R2 may exaggerate a model’s true ability to predict the dependent variable in the presence of overfitting, whereas leave-one-out R2 (LOOR2) is robust to overfitting. We demonstrate this by replicating 279 regressions from 100 papers in top economics journals, where the median increases of R2 and adjusted R2 over LOOR2 reach 40.2% and 21.4% respectively. The inflation of test errors over training errors increases with the severity of overfitting as measured by the number of regressors and nonlinear terms, and the presence of outliers, but decreases with the sample size. These results are further validated by Monte Carlo simulations.","PeriodicalId":51747,"journal":{"name":"Journal of Applied Economics","volume":null,"pages":null},"PeriodicalIF":1.4000,"publicationDate":"2023-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"How much should we trust R2 and adjusted R2: evidence from regressions in top economics journals and Monte Carlo simulations\",\"authors\":\"Qiang Chen, Ji Qi\",\"doi\":\"10.1080/15140326.2023.2207326\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"ABSTRACT R2 and adjusted R2 may exaggerate a model’s true ability to predict the dependent variable in the presence of overfitting, whereas leave-one-out R2 (LOOR2) is robust to overfitting. We demonstrate this by replicating 279 regressions from 100 papers in top economics journals, where the median increases of R2 and adjusted R2 over LOOR2 reach 40.2% and 21.4% respectively. The inflation of test errors over training errors increases with the severity of overfitting as measured by the number of regressors and nonlinear terms, and the presence of outliers, but decreases with the sample size. These results are further validated by Monte Carlo simulations.\",\"PeriodicalId\":51747,\"journal\":{\"name\":\"Journal of Applied Economics\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":1.4000,\"publicationDate\":\"2023-05-02\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Journal of Applied Economics\",\"FirstCategoryId\":\"96\",\"ListUrlMain\":\"https://doi.org/10.1080/15140326.2023.2207326\",\"RegionNum\":4,\"RegionCategory\":\"经济学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"ECONOMICS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Applied Economics","FirstCategoryId":"96","ListUrlMain":"https://doi.org/10.1080/15140326.2023.2207326","RegionNum":4,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ECONOMICS","Score":null,"Total":0}

引用次数: 0

摘要

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

How much should we trust R2 and adjusted R2: evidence from regressions in top economics journals and Monte Carlo simulations

ABSTRACT R2 and adjusted R2 may exaggerate a model’s true ability to predict the dependent variable in the presence of overfitting, whereas leave-one-out R2 (LOOR2) is robust to overfitting. We demonstrate this by replicating 279 regressions from 100 papers in top economics journals, where the median increases of R2 and adjusted R2 over LOOR2 reach 40.2% and 21.4% respectively. The inflation of test errors over training errors increases with the severity of overfitting as measured by the number of regressors and nonlinear terms, and the presence of outliers, but decreases with the sample size. These results are further validated by Monte Carlo simulations.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

Journal of Applied Economics ECONOMICS-

CiteScore

3.00

自引率

0.00%

发文量

审稿时长

40 weeks

期刊介绍： The Journal of Applied Economics publishes papers which make a significant and original contribution to applied issues in micro and macroeconomics. The primary criteria for selecting papers are quality and importance for the field. Papers based on a meaningful and well-motivated research problem that make a concrete contribution to empirical economics or applied theory, in any of its fields, are especially encouraged. The wide variety of topics that are covered in the Journal of Applied Economics include: -Industrial Organization -International Economics -Labour Economics -Finance -Money and Banking -Growth -Public Finance -Political Economy -Law and Economics -Environmental Economics