paper-with-me

Papers

Correcting Selection Bias in Standardized Test Comparisons

2023-09-19 · Onil Boussim

This paper tackles the critical issue of sample selection bias in cross-country comparisons using international assessments such as PISA (Program for International Student Assessment). While PISA is widely used to rank educational performance, its reliance on samples of students still in school at age 15 introduces survival bias, potentially distorting comparisons. To address this, I developed an econometric framework based on a quantile selection model. Under a stochastic dominance assumption, the selection-corrected quantile function is partially identified, with point identification achieved under parametric restrictions on the joint distribution of test scores and selection. Applying this method to PISA 2018 data, I demonstrate that correcting for selection bias leads to significant changes in country rankings based on mean performance. These findings underscore the importance of accounting for sample selection bias to ensure accurate and meaningful international educational comparisons.

📄 PDF Abstract BibTeX arXiv:2309.10642

Code (0)

등록된 구현이 없습니다.

Tasks

counterfactualSelection bias

Methods 이 논문이 사용한 방법론

PISA PrIme Sample Attention (PISA) directs the training of object detection frameworks towards prime samples. These are samples that play a key role in driving the detection…

Similar Papers 제목 키워드 기반

Non-Clicks Mean Irrelevant? Propensity Ratio Scoring As a Correction

2020-05-18 · Nan Wang, Zhen Qin, Xuanhui Wang, Hongning Wang

Recent advances in unbiased learning to rank (LTR) count on Inverse Propensity Scoring (IPS) to eliminate bias in implicit feedback. Though theoretically sound in correcting the bias introduced by treating clicked docume…

Learning-To-Rank

Correcting for Selection Bias in Learning-to-rank Systems

2020-01-29 · Zohreh Ovaisi, Ragib Ahsan, Yifan Zhang, Kathryn Vasilaky 외

Click data collected by modern recommendation systems are an important source of observational data that can be utilized to train learning-to-rank (LTR) systems. However, these data suffer from a number of biases that ca…

counterfactualLearning-To-RankPositionRecommendation Systems+1

Improving Evaluation of Debiasing in Image Classification

2022-06-08 · Jungsoo Lee, Juyoung Lee, Sanghun Jung, Jaegul Choo

Image classifiers often rely overly on peripheral attributes that have a strong correlation with the target class (i.e., dataset bias) when making predictions. Due to the dataset bias, the model correctly classifies data…

Classificationimage-classificationImage Classification

FastAMI -- a Monte Carlo Approach to the Adjustment for Chance in Clustering Comparison Metrics

2023-05-03 · Kai Klede, Leo Schwinn, Dario Zanca, Björn Eskofier

Clustering is at the very core of machine learning, and its applications proliferate with the increasing availability of data. However, as datasets grow, comparing clusterings with an adjustment for chance becomes comput…

Clustering

An experimental approach: The graph of graphs

2025-08-24 · Zsombor Szádoczki, Sándor Bozóki, László Sipos, Zsófia Galambosi arxiv

One of the essential issues in decision problems and preference modeling is the number of comparisons and their pattern to ask from the decision maker. We focus on the optimal patterns of pairwise comparisons and the seq…