Results of Milestone 4 will lay the groundwork for a field study to explore translation of our novel paradigm to real criminal cases. If our hypothesis regarding the superiority of the perceptual scaling protocol is supported by the proposed laboratory tests, we must then consider the problem of how to put this method in practice in the criminal justice system. To address potential implementation issues and to gain some understanding of their difficulty, we have made arrangements to collaborate on this second phase of the project with Prof. John Wixted of UCSD. Dr. Wixted has been working with Prof. William Wells of Sam Houston State University to conduct field studies with the Houston Police Department. Building on Dr. Wixted’s extensive experience with studies of this type, and tapping into his connections with law enforcement, we have developed a strategy to test our hypothesis under realistic conditions. This strategy is outlined below, along with a brief summary of some practical considerations related to this phase of our proposed research program. Our initial field tests of the perceptual scaling protocol must be performed in conjunction with a standard (e.g., simultaneous) lineup protocol, since the novelty of the former will surely provide fodder for defense attorneys seeking to discredit an identification, *regardless of any demonstrated superiority of the method in the laboratory*. In practice, we propose to do this in a field study by first subjecting the witness to the scaling procedure, followed immediately by presenting the witness with a simultaneous lineup of the same six photographs. This approach, which is identical to that in our proposed laboratory studies, will yield a traditional identification (satisfying the practical concern about basing the prosecution on a novel technique), but it will also allow us to directly compare witness performance using the two lineup protocols and to make that comparison in both laboratory and field environments. This approach may also help us to detect circumstances that lead to superiority of one protocol over the other. (For example, the emotional context associated with a real-world lineup may interfere to a greater degree with the scaling protocol, relative to the simultaneous protocol.) We see this approach of pairing the two protocols in the field as a natural step in validation of the novel scaling protocol. The situation is analogous in many respects to the precedent of introducing the sequential lineup procedure. The hypothesis that a sequential lineup would yield a lower rate of false identification relative to the simultaneous lineup was first put to a test in a laboratory study published by Gary Wells and colleagues in 1985. While the results supported the hypothesis and generated much excitement and optimism in the field, many laboratory replications followed the original finding (as there would be in any empirical science), including field validation, before the consensus was sufficient to change public policy and practice. We anticipate that a similar process will be required for the new scaling protocol before it can be deemed valid to stand on its own, even though this transition should take less time since the eyewitness research community is larger today and many of the conceptual and technical issues with this type of research are now well understood. A related issue is that the dependent variable in the perceptual scaling paradigm is different from that in the paradigms employed in traditional simultaneous and sequential lineup procedures. In the case of perceptual scaling there is no explicit verbal identification of the suspect. Rather, the dependent variable is a quantity that captures the likelihood (the uncertainty) for each individual in the lineup that he or she is the suspect. In simultaneous and sequential lineups, by contrast, the witness makes an explicit identification (“he’s the one”). The latter paradigms always have some degree of uncertainty associated with the identification, but no measure of that uncertainty is obtained, which is a significant weakness. The proposed scaling method overcomes this weakness. The fact that there is no explicit verbal identification associated with the perceptual scaling paradigm may prompt some resistance to this approach by the criminal justice community, even if we are able to demonstrate that the scaling procedure yields better discriminability. The only way to overcome such resistance is through field validation of the sort described above.