Arguing from the premise that a language test is a special case of a second language acquisition (SLA) elicitation device, I suggest that SLA and language testing share much common ground in terms of research methods, which have similar properties in that they are both used to make systematic observations of language performances from which inferences can be made about the state of a learner’s interlanguage ability underlying the performance. However, I also argue that whereas the concept of demonstrating validity and reliability has been integrated into how language testing research is conducted, SLA researchers have generally failed to recognize the need to demonstrate these qualities. I compare examples of SLA and language testing research articles in terms of their treatment of validity and reliability and argue: • that it is important for SLA researchers to provide evidence that the methods they employ to elicit data are appropriate for the purposes intended; • that the procedures provide stable and consistent data; and, consequently • that the interpretations they make of the results are justified.