Diagnosing language disorders in linguistically diverse populations is a clinical challenge. One source of the difficulty is that there is a lot of heterogeneity in typical bilingual development that is associated with the experiences of each child. Other sources of difficulty are the shortage of bilingual clinicians in English speaking countries, and a lack of appropriate tools in each language. Despite such obstacles, speech-language pathologists are tasked with the challenging job of accurately measuring language to make clinical diagnoses. Researchers have proposed that administering processing measures, like sentence repetition tasks, in one language could be a potential solution. Sentence repetition tasks are commonly used in clinical practice since they measure a variety of language skills in one task, and they have exhibited good diagnostic accuracy across languages. To evaluate the potential of this practice, Cowan and Miller (in review) conducted a meta-analysis and found that administering sentence repetition in one language does not yield adequate diagnostic accuracy for use in clinical practice. In this talk, I will discuss how incorporating individual differences into language evaluations could improve the accurate diagnosis of language disorders. First, I will discuss how currently used diagnostic tools were created and evaluated. Then, I will present preliminary data evaluating how individual differences in exposure and proficiency interact with semantic and lexical properties to influence sentence repetition performance. The results will help speech-language pathologists more accurately interpret sentence repetition performance in clinical practice.