# A significance test for the lasso

arXiv: https://arxiv.org/abs/1301.7161 · HTML: https://mlsanity.com/paper/1301.7161/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 Jan 2013.

## Built on by (most cited first, 20 of 20)

- [More Powerful Conditional Selective Inference for Generalized Lasso by Parametric Programming](https://mlsanity.com/paper/2105.04920/) (2021)
- [Conditional Selective Inference for Robust Regression and Outlier Detection using Piecewise-Linear Homotopy Continuation](https://mlsanity.com/paper/2104.10840/) (2021)
- [The Terminating-Random Experiments Selector: Fast High-Dimensional Variable Selection with False Discovery Rate Control](https://mlsanity.com/paper/2110.06048/) (2021)
- [Mapping the landscape of histomorphological cancer phenotypes using self-supervised learning on unlabeled, unannotated pathology slides](https://mlsanity.com/paper/2205.01931/) (2022)
- [Quantum Algorithms for the Pathwise Lasso](https://mlsanity.com/paper/2312.14141/) (2023)
- [Spatially relaxed inference on high-dimensional linear models](https://mlsanity.com/paper/2106.02590/) (2021)
- [Two-Stage Robust and Sparse Distributed Statistical Inference for Large-Scale Data](https://mlsanity.com/paper/2208.08230/) (2022)
- [Bounded P-values in Parametric Programming-based Selective Inference](https://mlsanity.com/paper/2307.11351/) (2023)
- [Weak Signals in the Mobility Landscape: Car Sharing in Ten European Cities](https://mlsanity.com/paper/2109.09832/) (2021)
- [Adaptive Noisy Data Augmentation for Regularized Estimation and Inference in Generalized Linear Models](https://mlsanity.com/paper/2204.08574/) (2022)
- [On confidence intervals for precision matrices and the eigendecomposition of covariance matrices](https://mlsanity.com/paper/2208.11977/) (2022)
- [Simple and Scalable Algorithms for Cluster-Aware Precision Medicine](https://mlsanity.com/paper/2211.16553/) (2022)
- [A parametric distribution for exact post-selection inference with data carving](https://mlsanity.com/paper/2305.12581/) (2023)
- [Sparsified Simultaneous Confidence Intervals for High-Dimensional Linear Models](https://mlsanity.com/paper/2307.07574/) (2023)
- [Predictive Performance Test based on the Exhaustive Nested Cross-Validation for High-dimensional data](https://mlsanity.com/paper/2408.03138/) (2024)
- [Statistically Testing Training Data for Unwanted Error Patterns using Rule-Oriented Regression](https://mlsanity.com/paper/2503.18497/) (2025)
- [Spacing Test for Fused Lasso](https://mlsanity.com/paper/2509.14229/) (2025)
- [φ-Table: A Statistical Explanation for Global SHAP](https://mlsanity.com/paper/2512.07578/) (2025)
- [New explanations and inference for least angle regression](https://mlsanity.com/paper/2602.02491/) (2026)
- [Post-Anomaly Detection Inference for Deep SVDD](https://mlsanity.com/paper/2609.37935/) (2026)

## Abstract (arXiv)

In the sparse linear regression setting, we consider testing the significance of the predictor variable that enters the current lasso model, in the sequence of models visited along the lasso solution path. We propose a simple test statistic based on lasso fitted values, called the covariance test statistic, and show that when the true model is linear, this statistic has an operatorname Exp(1) asymptotic distribution under the null hypothesis (the null being that all truly active variables are contained in the current lasso model). Our proof of this result for the special case of the first predictor to enter the model (i.e., testing for a single significant predictor variable against the global null) requires only weak assumptions on the predictor matrix X. On the other hand, our proof for a general step in the lasso path places further technical assumptions on X and the generative model, but still allows for the important high-dimensional case p>n, and does not necessarily require that the current lasso model achieves perfect recovery of the truly active variables. Of course, for testing the significance of an additional variable between two nested linear models, one typically uses the chi-squared test, comparing the drop in residual sum of squares (RSS) to a chi21 distribution. But when this additional variable is not fixed, and has been chosen adaptively or greedily, this test is no longer appropriate: adaptivity makes the drop in RSS stochastically much larger than chi21 under the null hypothesis. Our analysis explicitly accounts for adaptivity, as it must, since the lasso builds an adaptive sequence of linear models as the tuning parameter λ decreases. In this analysis, shrinkage plays a key role: though additional variables are chosen adaptively, the coefficients of lasso active variables are shrunken due to the ℓ1 penalty. Therefore, the test statistic (which is based on lasso fitted values) is in a sense balanced by these two opposing properties - adaptivity and shrinkage - and its null distribution is tractable and asymptotically operatorname Exp(1).