Do Metrics for Counterfactual Explanations Align with User Perception?

Felix Liedeker; Basil Ell; Philipp Cimiano; Christoph Düsing

cs.AIMar 16, 2026

Do Metrics for Counterfactual Explanations Align with User Perception?

Authors: Felix Liedeker, Basil Ell, Philipp Cimiano, Christoph Düsing

Organizations: Bielefeld University, Bielefeld, Germany

Abstract

Explainability is widely regarded as essential for trustworthy artificial intelligence systems. However, the metrics commonly used to evaluate counterfactual explanations are algorithmic evaluation metrics that are rarely validated against human judgments of explanation quality. This raises the question of whether such metrics meaningfully reflect user perceptions. We address this question through an empirical study that directly compares algorithmic evaluation metrics with human judgments across three datasets. Participants rated counterfactual explanations along multiple dimensions of perceived quality, which we relate to a comprehensive set of standard counterfactual metrics. We analyze both individual relationships and the extent to which combinations of metrics can predict human assessments. Our results show that correlations between algorithmic metrics and human ratings are generally weak and strongly dataset-dependent. Moreover, increasing the number of metrics used in predictive models does not lead to reliable improvements, indicating structural limitations in how current metrics capture criteria relevant for humans. Overall, our findings suggest that widely used counterfactual evaluation metrics fail to reflect key aspects of explanation quality as perceived by users, underscoring the need for more human-centered approaches to evaluating explainable artificial intelligence.

Figures & tables

Dataset	Instances	Features	Classes	F1 Score	Test Inst.	Valid CFs
Mushroom (MUS)	8124	22	2	1.00	1625	755
Obesity Levels (OBE)	2111	16	7	0.95	423	211
Heart Disease (HRT)	303	13	2	0.85	60	25

Table 1: Overview of dataset characteristics, classification performance (using XGBoost), and number of successfully generated counterfactuals.

	Acc.	Und.	Plaus.	Suff.	Sat.	CQS
MUS	2.22 $\pm$ 0.23	1.83 $\pm$ 0.16	2.08 $\pm$ 0.26	2.08 $\pm$ 0.23	2.08 $\pm$ 0.20	2.06 $\pm$ 0.17
HRT	2.25 $\pm$ 0.27	1.95 $\pm$ 0.19	2.24 $\pm$ 0.23	2.20 $\pm$ 0.23	2.29 $\pm$ 0.24	2.19 $\pm$ 0.20
OBE	2.14 $\pm$ 0.30	1.85 $\pm$ 0.27	2.16 $\pm$ 0.33	2.10 $\pm$ 0.25	2.19 $\pm$ 0.33	2.09 $\pm$ 0.27
All	2.20 $\pm$ 0.27	1.87 $\pm$ 0.22	2.15 $\pm$ 0.28	2.12 $\pm$ 0.24	2.18 $\pm$ 0.28	2.10 $\pm$ 0.22

Table 2: Mean ratings per dataset across all dimensions and the CQS.

Figure 1: Metric–rating correlations including the CQS for (a) MUS, (b) OBE, (c) HRT datasets. Significant correlations ( $p<0.05$ ) are marked with an asterisk.

Figure 2: Histograms of

R^{2}

for (a) linear regression and (b) the best performing non-linear model, RF, on the HRT dataset predicting user satisfaction rating.

Figure 3:

R^{2}

by number of included metrics for (a) linear regression and (b) RF, the best performing non-linear model, on the HRT dataset predicting user satisfaction rating. Lines show the best and mean

R^{2}

across all metric combinations at each complexity level, with shaded bands indicating

\pm 1

SD.

Counterfactual Evaluation

Counterfactual evaluation tests how models, explanations, and policies behave under hypothetical changes to inputs, evidence, actions, or environments.

See all 96 papers

CEL: Comprehensive Counterfactual Explanations Library and Benchmark
Jul 24, 2026Oleksii Furman, Łukasz Lenkiewicz, Marcel Musiałek +1Counterfactual Evaluation Explainable Artificial Intelligence
Wrocław University of Science and Technology, Wrocław, Poland
Improving understanding and trust in AI: How users benefit from interval-based counterfactual explanations
Feb 22, 2026Tabea E. Röber, Paul Festor, Rob Goedhart +2Counterfactual Explanations Explainability Evaluation
Amsterdam Business School, University of Amsterdam, The Netherlands · Department of Bioengineering, Imperial College London, UK · Institute of Artificial and Human Intelligence, University of Bayreuth, Germany
Does Explanation Correctness Matter? Linking Computational XAI Evaluation to Human Understanding
Mar 26, 2026Gregor Baer, Chao Zhang, Isel Grau +1Explainable Artificial Intelligence Explainability Evaluation
Information Systems Group, Eindhoven University of Technology · Human-Technology Interaction Group, Eindhoven University of Technology

Jul 24, 2026J/K move · Enter open · S save

CEL: Comprehensive Counterfactual Explanations Library and Benchmark

Oleksii Furman, Łukasz Lenkiewicz, Marcel Musiałek, Maciej Zięba

Wrocław University of Science and Technology, Wrocław, Poland

Counterfactual Evaluation Explainable Artificial Intelligence Counterfactual Explanations Benchmark Design

Open paper AskPDF arXiv

Counterfactual explanations are a prominent approach in explainable artificial intelligence (xAI), providing actionable guidance on what input changes would alter a model's prediction to a desired outcome. While early methods primarily focused on minimal feature changes, recent work incorporates additional properties such as sparsity, actionability and plausibility. Despite this progress, fair and systematic evaluation remains challenging. Existing studies often rely on different data splits, predictive models, and evaluation metrics, which limits objective comparison across methods. To fill this gap, we introduce CEL (Counterfactual Explanations Library), a unified library and benchmark for counterfactual explanations designed to support consistent implementation and evaluation. CEL includes 18 datasets of varying size and complexity and provides implementations or reimplementations of 14 widely used counterfactual methods. Using this standardized setup, we conduct a comprehensive quantitative comparison across a variety of methods on datasets that differ in size, number, and types of attributes. The evaluation protocol incorporates multiple complementary metrics capturing validity, coverage, sparsity, proximity, and distributional plausibility, including density- and outlier-based measures to assess the realism of generated counterfactuals. To the best of our knowledge, this is the first comprehensive benchmark that systematically evaluates recent counterfactual explanation methods within a unified and reproducible framework. While prior libraries and benchmarking efforts exist in the literature, many are outdated, limited in scope, or lack consistent evaluation protocols. The proposed benchmark aims to improve reproducibility, enable fair comparison, and establish a workbench for the development of future counterfactual explanation methods.

Abstract

Counterfactual Evaluation

Explore similar work

CEL: Comprehensive Counterfactual Explanations Library and Benchmark

Improving understanding and trust in AI: How users benefit from interval-based counterfactual explanations

Does Explanation Correctness Matter? Linking Computational XAI Evaluation to Human Understanding