Background: Linguistically motivated treatment protocols like the Treatment of Underlying Forms (TUF: Thompson & Shapiro, 2005) have shown significant success in remediating aphasic individuals' sentence production deficits. However, adults with aphasia are not uniform in their response to TUF: not all individuals trained with TUF successfully acquire the sentence types they are trained on or generalise to untrained sentence types. More research is therefore needed to determine which individuals are most likely to benefit from TUF treatment. Aims: The current study analysed existing TUF treatment studies in an effort to determine what measures may be predictive of TUF outcomes for different aphasic individuals. Three different measures were tested: aphasia severity, auditory comprehension ability, and complex sentence comprehension ability. Methods & Procedures: A meta-analysis was conducted based on existing TUF treatment studies drawn from the aphasiological literature. These studies included individual demographic, language-testing, and treatment data from 30 aphasic individuals. Regression analyses were conducted comparing these individuals' improvements on production of treated and untreated sentence types (treatment and generalisation effects) with the three predictor measures (severity, auditory comprehension, and complex sentence comprehension scores). Outcomes & Results: Only one of the measures tested, general auditory comprehension, was predictive of the size of individuals' gains on treated sentence types. None of the measures tested was predictive of these individuals' generalisation to untrained structures. Conclusions: The current results suggest that general auditory comprehension appears to be related to improved sentence production following TUF treatment. In contrast, neither overall aphasia severity nor performance with complex sentence stimuli is a strong predictor of TUF treatment outcomes. Interestingly, there were no strong relationships between any of the measures and the generalisation effect scores. These findings suggest that clinicians should consider a patient's general auditory comprehension when deciding whether TUF would be appropriate. They also suggest that partially different cognitive mechanisms may underlie treatment and generalisation effects following treatment, at least for TUF protocols.