Free-response research-based assessments, like the Colorado Upper-division Electrostatics Diagnostic (CUE), provide rich, fine-grained information about students' reasoning. However, because of the difficulties inherent in scoring these assessments, the majority of the large-scale conceptual assessments in physics are multiple choice. To increase the scalability and usability of the CUE, we set out to create a new version of the assessment that preserves the insights afforded by a free-response format while exploiting the logistical advantages of a multiple-choice assessment. We used our extensive database of responses to the free-response CUE to construct distractors for a new version where students can select multiple responses and receive partial credit based on the accuracy and consistency of their selections. Here, we describe the development of this modified CUE format, which we call coupled multiple response (CMR), and present data from direct comparisons of both versions. We find that the two formats have the same average score and perform similarly on multiple measures of validity and reliability, suggesting that the new version is a potentially viable alternative to the original CUE for the purpose of large-scale research-based assessment. We also compare the details of student responses on each of the two versions. While the CMR version does not capture the full scope of potential student responses, nearly three-quarters of our students' responses to the free-response version contained one or more elements that matched options provided on the CMR version.