DOJ Cares About Accurate Eyewitness Id’s as Well (Unless “Making America Great Again” and “You’re Fired” Changes That)

January 31, 2017
By Hanging Out with Carl Gunn


  • A recent DOJ memo on eyewitness identification contains recommended procedures to help prevent erroneous identifications, including either blacking out unique identifying characteristics or adding them to the “filler” photos; administration by an officer who doesn’t know who the suspect is or doesn’t know the order in which photos are presented; and videotaping or at least audiotaping the identification.
  • A lengthy National Academy of Sciences publication cited in the memo makes similar recommendations for procedure and also recommends more frequent use of pretrial hearings and eyewitness identification experts.
  • Think about using these materials in either cross-examination or motions to suppress identifications and/or present eyewitness identification expert testimony.



A panel attorney on one of the e-mail lists I’m on — Maria Davila in Tucson, Arizona — sent around an e-mail recently about a January 6, 2017 DOJ memo on eyewitness identification procedures, which is linked here.  The memo explains that “[o]ver the past year, a team of Department experts – including prosecutors, law enforcement personnel, and social scientists [no defense attorneys, I guess?] – have worked together to study the research and identify best practices,” and that their work “culminated” in a list of recommended procedures which are described in an attachment to the memo.  Interestingly, it was authored by Acting Attorney General Sally Yates, whom our (or at least some of ours) President fired yesterday in his latest effort to make America great again.  Hopefully, the ideas in the memo will survive, but I guess we’ll see.  And maybe there will be other ways to use the ideas in the memo.

Some of the recommended procedures aren’t necessarily new, but others are.  Some seem to be an improvement, though others I’m not so sure of.  The full list of procedures is too long for me to reproduce in full here, so I’d urge you to read the memo and attachment for yourself.  But I thought I’d list some of the ones I found more noteworthy.

1. It’s recommended that the witness view the photo array out of earshot and view of others.

2. It’s recommended that the administrator of the photo array avoid using a photo that’s several years old and “select a photograph of the suspect that resembles the witness’s description of the perpetrator or the perpetrator’s appearance at the time of the incident.”  Depending on what this means, this seems potentially problematic, since deliberately choosing a similar photo could shade the procedure toward identification.

3. It’s recommended that the filler photographs be “sufficiently similar so that a suspect’s photograph does not stand out, but not so similar that a person who knew the suspect would find it difficult to distinguish him or her.”  This also struck me as potentially problematic, because I’ve always thought the idea was to get people who looked as similar as possible.

4. It’s recommended that “a unique feature, such as a scar, tattoo, or mole” either be added to the filler photographs, so that all the photographs have the feature, or blacked out with similar black marks placed on the filler photographs so all the photos look the same.  This struck me as a good new procedure that I hadn’t heard of before.

5. It’s recommended that all of the photos in the photo array be of similar size, background, format, and color.

6. The memo allows administrators to use either sequential presentation, where photos are shown to the witness one at a time, or simultaneous presentation, where all the photos are shown at once (as in the classic photospread).  I haven’t ever seen the sequential method used in federal court, and I thought studies showed it was better, but the appendix in the attachment says some recent research “has raised questions about the superiority of sequential methods.”  It cites some of the studies if you want to read them, as well as a National Academy of Sciences publication – “Identifying the Culprit: Assessing Eyewitness Identification.”

7. The memo notes the importance of avoiding suggestiveness and suggests the possibility of a “blind” administrator who’s not involved in the investigation.  The appendix notes the administrator in this procedure wouldn’t even necessarily know who the suspect was, quoting the National Academy of Sciences publication.  If an administrator who’s not involved in the investigation can’t be used, it’s suggested an administrator who is involved in the investigation be “blinded” by sitting in a place where he or she can’t see the order or arrangement of the photographs, after having randomized them in some way.  I’d never heard of either of these procedures being used before, and both seem preferable to the usual procedure of the agent sitting down with the witness and just putting the photospread down in front of both of them.  The memo recognizes that sometimes neither of the “blind” procedures will be possible, but states that “[i]n those instances the administrator should document the reasons for the non-blind(ed) procedure and be prepared to explain the reasons for conducting such an alternative procedure.”

8. There’s also recommended instructions to be given to the witness, which may be a little more than what we’re used to, though it’s been a while since I’ve heard typical instructions.  Three instructions I’m not sure I remember are that the witness should tell the administrator how confident he or she is if he or she does recognize someone; that it’s “okay” if the witness doesn’t recognize anyone and officers will continue to investigate the case; and that the witness shouldn’t assume the administrator knows who committed the crime.

9. It’s recommended that if the witness makes an identification, the administrator ask the witness “to state in his or her own words how confident he or she is in the identification.”  If the witness’s identification is “vague,” like saying something such as “I think it’s #4,” the administrator should ask, “What do you mean by that?”  While the memo doesn’t expressly say so, this suggests the importance of asking in an open-ended non-suggestive way.  The last research I’d heard about, admittedly years ago, suggested a witness’s confidence actually doesn’t correlate with accuracy, but the appendix to the memo claims recent research suggests the contrary.  In any event, having the witness’s statement at the time is probably better than just getting it at trial.

10. It’s recommended that the identification be documented by either videotaping or audiotaping it, or “immediately writing down as close to verbatim as possible the witness’s identification and statement of confidence, as well as any relevant gestures or non-verbal reactions,” and that the witness should then confirm the accuracy of the statement.  This idea of recording the identification seems like a great idea. I haven’t seen it used before, but the appendix to the memos says that approximately one-fifth of state and local law enforcement agencies have started videotaping identifications and the National Academy of Sciences recommends it “become standard practice.”

11. It’s recommended that the administrator document (a) how long it took the witness to make an identification; (b) the method and order of presentation of the photos; (c) the names of all persons present; and (d) “[a]ny other facts or circumstances that would help contextualize or explain the witness’s selection.”

I’m also attaching the much more lengthy National Academy of Sciences publication cited in the appendix to the memo – linked here.  It concludes with a “Findings and Recommendations” section that has 11 recommendations – some about identification procedure, some about the treatment of identification evidence in court, and some about further research.  The recommendations about procedure include recommendations of “double-blind” procedures, standardized instructions, documentation of witness confidence, and videotaping of identifications.  The recommendations about the treatment of identification evidence in court include holding pretrial hearings and the use of expert testimony.  A prior post urges pushing for such expert testimony (see “Last Week’s Post and Eyewitness Identification Testimony” in the March 2013 link at the right), and the recommendation in this study provides additional support for those efforts.

We obviously can’t force law enforcement agencies to follow these recommended procedures, though maybe we could try in the rare case where there’s a post-arrest identification procedure that we know about.  There may be other ways to use this memo and/or the National Academy of Sciences report, however.  It might be a stretch, but we could argue at least the memo and its appendix is an admission of a party opponent about what best practices are, under the case law and arguments set forth in the “Government Confessions!  Or at Least Admissions” post in the June 2014 link at the right.  (Arguing this applies to the National Academy of Sciences Report will be harder because the National Academy is, as described at the beginning of the report, “a private nonprofit, self-perpetuating society,” albeit under “a charter granted to it by the Congress in 1863” and with “a mandate that requires it to advise the federal government on scientific and technical matters.”) Alternatively, one might try to cross an agent about whether he used these procedures and impeach the agent if he or she claims to have never heard of these procedures or says something like, “That’s not our procedure.”  Testimony like that could open the door to impeachment by the existence of a report that in fact does recommend these procedures, especially if the agent is with a law enforcement agency within the Department of Justice.  (And hasn’t been “fired” by our (or some of ours) President.)  As another alternative, perhaps the memo or report could be used in a motion to suppress a pretrial identification.