Students with or at risk of reading difficulties (RD) benefit from accurate early identification and intervention. Previous research has employed various decoding measures to screen students for RD, but the criteria for identification have been inconsistent. Assessing students with RD is especially challenging in English Language Learners (ELLs), as vocabulary deficits can impact decoding. Additionally, few research syntheses have examined whether researchers use different measures to screen ELLs and EL1s for RD, and whether these differences result in distinct decoding profiles between ELLs with RD and EL1s with RD. To address these gaps, this study uses a meta-analysis to examine the decoding measures used in RD assessments and whether outcomes differ for ELLs and EL1s. The findings show that real word reading assessments identify students with more pronounced decoding deficits than nonword reading assessments. Despite the use of different RD screening measures for ELLs and EL1s, the gap between ELLs with and without RD was similar to that between EL1s with and without RD. These results suggest that real word-reliant measures, which are influenced by word knowledge, provide a more comprehensive assessment of RD than nonword-reliant measures for both ELLs and EL1s. We encourage future researchers to use consistent decoding measures when screening RD in both populations, to maximize comparability of findings.