This registered report tests the reliability of participants' picture-naming speed within a session (split-half reliability), over time (test-retest reliability), and across task manipulations (speeded vs. non-speeded picture naming). The results will inform us on whether picture-naming speed is a stable property of an individual, which will have theoretical implications for correlational studies on individual differences in picture-naming speed. The findings will also have methodological implications for correlational research with picture-naming measures because reliability is related to statistical power.