The primary objective of this study is to evaluate empirically the ability of two cross-sectional models, the Cross-Sectional Jones Model and the Cross-Sectional Modified Jones Model, to detect earnings management vis-a-vis their time-series counterparts. The motivation follows because these two cross-sectional models have not been formally evaluated by prior research, and because their use offers substantial advantages to investors and researchers over their time-series counterparts. A secondary objective is to assess the robustness of findings of prior studies assessing discretionary-accruals models using our new sample and research method, which controls for potential research confounds. The evaluation involves examining the association between discretionary accruals and audit qualifications, using a sample of 166 distinct firms with qualified audit reports and a matched-pair control sample with clean audit reports. An association between large discretionary accruals generated by a model and an audit qualification provides evidence on the ability of the model to detect earnings management. Results from univariate tests that do not control for potential research confounds show that all models, except the DeAngelo Model, are consistently successful in discriminating between firms that manage earnings. Once potential research confounds are controlled, however, only the two cross-sectional models are able to detect earnings management. This last result, which highlights the importance of controlling for research confounds in earnings management studies using carefully selected samples, implies that the cross-sectional models are superior to their time-series counterparts. This finding is particularly important for future earnings management research because using a cross-sectional model rather than its time-series counterpart should result in a larger sample size that is less subject to a survivorship bias, and will also allow examining samples of firms with short history.