Wednesday, January 19, 2011

Machines and human witnesses

Paul Vinegrad has posed a question to me as a comment to an ealrier post. It’s sufficiently interesting that I’ll answer it here. Paul’s question:

Prof. Friedman, your thoughts on the following hypothetical would be illuminating:

Cop arrests suspect for DUI. Cop takes suspect's blood sample to government run lab. Cop watches as analyst places blood sample in gas chromatograph. Gas chromatograph prints out result of .21 BAC. Cop takes print out and leaves.

At trial, Cop describes this sequence of events. Cop identifies print out. Print out is marked as an exhibit and introduced into evidence to prove defendant's BAC was .21. Cop does not offer any opinion on accuracy/reliability of print out.

Did any human "witness" make a "statement" (testimonial or otherwise) that defendant's BAC was .21?


My answer is that, with one caveat discussed below, there is no human witness that the BAC is .21. What we have is a witness to a process that generated the identified piece of paper.

If instead of this test, it were a simple litmus paper test, and the witness identifies a particular piece of red litmus paper as the one that was dipped in a particular solution, and says it is now the color that it was after it was dipped in, I don’t think there’s yet been an assertion by a human witness that the solution was acidic. One can infer that proposition by understanding the processes that led to the red color.

Now of course the case Paul presents is different in that humans presumably created and calibrated the machine. But presumably that all happened before the particular testing. Even if one could discern a human assertion there, it’s harder to discern an assertion about the particular sample, and I think Melendez-Diaz suggests pretty strongly, for better or worse, that an assertion at this point wouldn’t be considered testimonial.

The caveat is that it is theoretically possible that somebody purposely set the machine so that it would report a BAC of .21. If so, I’d say that’s an assertion, made by a human through a machine – just as one might use a machine to broadcast his voice. But I think the burden would be on the defendant to produce evidence of this kind of manipulation.

The most difficult question here is one of relevance, or probative value: Is there enough of a basis for the trier of fact to conclude that because the piece of paper says .21 that in fact the blood alcohol content of the sample tested was .21? Compare this to a witness who says that when a robbery occurred the digital clock on the top floor of the bank said it was 1:45. In a case like that, I don’t think we’d demand proof as a precondition to admissibility that the clock was accurate, at least if the evidence is significant even without pinpoint accuracy; the jury could probably infer that the clock was likely reasonably accurate or it would’ve been corrected. (Though in my town there is a prominent digital clock, right where cars pull off a main road into one of the malls, that has been slow by about four minutes for many years). Of course, the prosecution might choose to supply such proof on its own initiative, and the defense may challenge the accuracy of the clock. But in this case, assuming the cop isn’t familiar with the machine from other encounters, is there enough of a basis for the trier of fact to infer that the machine is reasonably accurate? I don’t know. Certainly the prosecution would be better off presenting proof of accuracy. But I don’t think failure to do so is a confrontation problem.

15 comments:

Justin J. McShane said...

Well, in my personal opinion, the hypothetical is still testimonial. We are talking Confrontation, right?

For better or worse as defined by Crawford, Hammond/Davis (maybe it will change with Bryant) and Melendez-Diaz: Confrontation is required when there is a an out of court statement made in lieu of testimony made for its primary purpose of later criminal litigation where the defendant has not had a meaningful opportunity to cross-examine the person live and under oath. The particular witness rule.

This hypo is a series of statements made by the analyst: I took the defendant's blood vial. I prepared the blood vial using a procedure and following instructions. I ran the blood vial on a particular machine. I arrived at a result based upon how I programmed the machine and set its reporting parameters. I report it as 0.21 g/mL EtOH in that particular blood.

This hypothetical is just another form of even worse surrogate witness than in Bullcoming. At least in Bullcoming, they provided for another analyst familiar with the way the process was SUPPOSED to have occurred. Here it is even worse! No technical person who is presented.

How is this hypothetical outside of the definition? It is out of court. The DUI test is only made for purposes of later criminal litigation. It is a statement. It is offered in lieu of testimony. There is no cross of the person who made the result under oath. Confrontation.

Richard D. Friedman said...

As I understood Paul's hypothetical, Cop observed everything and testified as to what she observed. If the prosecution doesn't offer a statement by the lab tech who performed the test, then it isn't necessary to produce that tech, so long as someone else can testify to how the test was performed. If the witness who does testify isn't well qualified to explain what happened, that presumably will hurt the prosecution. But even if not, the problem isn't failure to given an opportunity to confront a witness, because no testimonial statement by the lab tech has been used by the prosecution.

Justin J. McShane said...

(3 of 3)
The second problem with allowing any form of surrogate in lieu of the particular witness is the issue of the burden of production and persuasion. In a DUI case, the EtOH is an essential element of the crime. As such, it must be proven by the the Government with evidence and it must be evidence of sufficient caliber to be beyond a reasonable doubt.

By not honoring the particular witness rule in the Government's case in chief, you frustrate the discovery of truth as the surrogate will have no answers at all in the example you posit of the cop and little, if anything, in the case of the supervisor or professional witness or substitute analyst. How do you cross-examine a piece of paper offered by someone who has no personal knowledge of the process or the test?

In practical effect by allowing this, you have made these reports an irrebuttable presumption in the prosecution's case-in-chief. Remember, this is an essential element!

Why is this so complicated in the case of forensic science results? I just don't understand how it is so controversial.

Justin J. McShane said...

In response to Professor Friedman...

Interesting.

Accepting the hypothetical as you explain it to be (and I write it out so I can understand it)--
There is this "other person" who is not the analyst and would testify in court. This "other person" takes no active role in the analysis, but instead literally looks over the analyst's shoulder from beginning to end. In this case, as you note, the particular witness who is performing the acts (the analyst) is not in court, but only this "other witness" is. You suggest that Confrontation is not implicated, correct?

If so, in that case as the "other person" would be available and presented and as such the Government is not required to produce the particular analyst who did the physical acts and who has the presumptive training to perform the acts, right?

My question becomes: Why wouldn't they just install video and record the whole process from beginning to end and authenticate it through the supervisor so it is admissible? Is that violative of Confrontation?

Unknown said...

Justin: The original hypothetical, and your further hypo regarding video recording, probably wouldn't implicate the Confrontation Clause, as Professor Friedman argued. However, you're correct that the police officer wouldn't be able to answer any of the relevant questions, and the video recording wouldn't have enough information.

However, I don't think this situation is as dangerous as you're suggesting.

There would be a few issues for any prosecutor trying to pull something like this. First, the witness introducing the evidence would have to have enough knowledge about the process to testify about what the printout means, not for Confrontation purposes, but for a general relevancy foundation. The witness introducing the document would have to have enough background in the science to establish that it's relevant to the prosecution, which a non-analyst probably wouldn't have. And if there was an analyst who witnessed every step of the procedure, then there's clearly no problem with them introducing the evidence, even if someone else performed the physical actions of testing.

Second, the operation of the machine and science behind it (as well as proof of testing of the machine) is what these cases often focus on, from my experience reading these transcripts. A non-qualified observer might be able to testify about what he actually saw, but there's no way he can testify that all of the specific procedures were followed correctly. He can't explain the workings of the machine. He can't verify that it was in proper working order. Etc, etc, etc. Any competent defense attorney would get all of those failings of information out during cross examination, and then focus on that the entire time. And I have to believe that even the most jaded of juries or judges would realize that they're not being presented with enough information.

The best evidence that this hypothetical isn't really a threat in reality is the fact that prosecutors haven't tried it yet. They *know* that the bare printouts from the machines aren't enough, and that's why they don't try to introduce those. The prosecution always tries to introduce various forms of certificates or affidavits from the analyst, asserting that all procedures were followed and that the result means such and such. The prosecution knows that they want more information than a non-qualified observer can give them, absent assertions by the actual analyst (or a qualified eye-witness). Given the choice between introducing a videotape of testing or a non-qualified observer's testimony versus the testimony of the actual analyst, any halfway decent prosecutor is going to bring the analyst in. And it'll probably be cheaper for them anyways.

So, yeah. This hypothetical is interesting in that it explains the line between machine-generated results and human witnesses. But it doesn't create a real-life threat in criminal cases.

Unknown said...

May I post my own hypothetical? I'd love to hear some thoughts. The State alleges that a sex offender improperly created a profile on MySpace. MySpace returns its data (IP address logs, profile information, etc.) related to the profile as a result of a subpoena issued. Must someone from MySpace testify in court in order to avoid running afoul of the Confrontation Clause? This evidence is the crux of the case.

paul v said...

Prof. Friedman, in your opinion is the CC implicated if my original hypothetical is changed as follows:

The gas chromatograph doesn't create a printout of the .21 BAC result. Instead, an LED light on the machine indicates the result as .21.

At trial, the cop is asked by the prosecutor what the result on the machine was? The cop states it was .21.

In your opinion, would this raise a CC issue, or, like the print out scenario, would this situation also not involve a CC "witness"?

Richard D. Friedman said...

So, in response to the comments above: (1) Yes, I think Justin's last comment has restated my views. It seems to me that the presence of someone who can testify from first-hand knowledge to everything that occurred ordinarily would take care of the Confrontation Clause concern. Andrew, my former (I can't say old) research assistant, has expressed well many of the other concerns that would arise in this setting, and as he says these explain why prosecutors don't do this.

(2)On Carla's case (which perhaps is not a hypothetical?), it seems to me that absent stipulation someone has to testify live as to what this stuff is and how it came to be. That's the testimonial statement that MySpace is making.

(3) On Paul's revised hypothetical, I don't think it matters so far as the Confrontation Clause is concerned whether the output of the machine is a red light or a printout. The other problems remain. And of course Cop is still a witness for purposes of the Clause; it's his statement that is being introduced/

Unknown said...

It seems that the MySpace hypo is similar to the old public/business records exception to the hearsay rule. In general, you wouldn't need someone to testify about the method of gathering or creating the documents (for confrontation purposes because the creation of that document isn't testimony) but you do need someone to testify about the retrieval of those documents, since that retrieval was done for prosecutorial purposes.

Justin J McShane said...

I think the MySpace hypothetical (under Hammond/Davis) is an example of testimonial evidence, isn't it IF THIS IS A NEWLY GENERATED DOCUMENT MADE IN PREPARATION OF THE CRIMINAL TRIAL?

To serve as a historical reference, and had been held to be a violation of the principle of Confrontation that existed at the time of the Framing, as implicating the core issue is the classic 1603 trial of Lord Raleigh for treason. Lord Cobham had implicated Raleigh in his crime before the Privy Council and in a letter and Raleigh's continued demand for confrontation, but not at the criminal jury trial. As the letter was produced for the "primary purpose" (and in fact its sole purpose) was for criminal prosecution for treason, there was a violation.

If we are stuck with the "primary purpose rule" is it not the case that the record generated (this was on MySpace at this time) was made for the primary purpose of later criminal trial testimony?

However, using the same jurisprudence, but slightly changing the facts, if there was some sort of independently generated report that pre-existed with the same information made by MySpace for MySpace purposes (e.g., a site audit), then the result would be different, wouldn't it?

Indeed the Davis court pointed out that this goes back to the classic example of the common law English tradition that was found not to be violative of Confrontation and that definitely existed at the time of the Framing of the Constitution which was the "shop keeper's ledger" example. Where in that case, as the distinction was made that the "primary purpose" of the ledger was to keep the books, not to aid in criminal prosecution.

If Bryant removes the "primary purpose" rule (which I hope that it would to provide for less fungibility and subjectiveness in the practical application of CC rules) in favor of a pure "in lieu of testimony" rule, then we would have a different result, wouldn't we with this internally created document example?

Again, as the court weighs the arguments presented in Bryant v Michigan, that might change. Until then, I am afraid that the "primary purpose" of the assertion is the SCOTUS rule.

Confrontation and Hearsay exception are totally different, of course. Both objections need to be overcome to gain admissibility.

paul v said...

Prof. Friedman, let's take my hypothetical one step further:

Cop (who took printout or observed .21 red light on GC) writes down that machine-generated result/"statement" on a piece of paper. Cop dies before trial.

At trial, would introduction of the piece of paper with the .21 result be violative of the CC? (Assume a proper foundation for the paper can be laid by another person at trial.) If so, why would the cop's simple act of writing down the machine-generated result (which you concede is not a "statement" by a human "witness" under the CC) transform a non-statement by a non-witness into something the CC should be concerned with?

Richard D. Friedman said...

Aha! Paul is playing law prof with me -- "Do you take it this far? OK, a little bit further? Further yet?" I think Andrew and Justin have already suggested one explanation in answer to Paul's latest hypo. The now-dead cop's notation was created in anticipation of prosecutorial use. That is a human act of witnessing, in this case a condition (the readout of the machine) that assertedly suggests guilt. If Cop had written, "The traffic light was red," knowing that this information would be crucial in a criminal case, there's no doubt that the statement would be testimonial. I don't think this one is any different. Perhaps it's more likely to be accurate, but that really doesn't matter for confrontation purposes anymore.

There's another factor as well -- remember that in Paul's original hypo, the cop observed the whole test. I assume that fact remains in the hypo -- if not, we have no evidence as to how the machine came to be displaying certain numbers. So if the cop's statement is taken (rather generously to the prosecution) as implicitly asserting that the test was conducted on the proper sample in accordance with satisfactory procedures, it is testimony as to much more than the machine's readout. And if it isn't taken to be asserting that, I don't see that the prosecution has an adequate basis for introducing the statement.

paul v said...

What if the cop didn't make a notation of the .21 reading, but simply photographed it? Just kidding. No response needed. Thanks for letting me play "Law Professor For A Day."

Justin J. McShane said...

Professor Friedman and others,
Thank you for this discussion as it has helped me to view Confrontation Clause issues in a more robust manner. If it would be possible could I ask for an opinion of all as to what the Court may do with the Bryant case, if anything?

Unknown said...

Hello everyone. Professor, you are right; it is not a hypothetical. This is a case I am litigating. At trial, the prosecutor is going to admit the evidence two ways: either through a business records affidavit, or through testimony from the detective, who printed off the profile.

Why two different ways? Because MySpace never provided him with an official copy of my client's profile. So it can't come in through the business records exception. Without it, they can't link the profile to my client on his name alone (it's a very typical name).

But even if there is a business records affidavit, wouldn't this run afoul of the CC if the MySpace rep isn't actually there? I cannot cross-examine her on how they keep the IP address logs, how they maintain their records, or how they know that the person whose profile it purports to be actually set up the profile?

I look at it as being similar to lab reports.

But Mr. McShane's helpful comments lead me to believe I'm incorrect. I'd love to hear more to make sure I'm on the right track.