Wednesday, October 12, 2011

Did Courtroom Demonstration Create Impression of Infallible Arson Dog? Prejudice Argument Swept Aside by Appeals Court

William Halliday was convicted of first degree arson and insurance fraud for setting his house on fire and attempting to collect insurance money. We are given no description of the events behind the fire by the Alaskan appellate court to which Halliday appealed his conviction. Rather, the appellate court focuses on Halliday’s assignment of error to one crucial piece of evidence, the testimony of an arson dog handler and a demonstration conducted by that handler with her dog before the jury.

Deputy Fire Chief Bridget Bushue testified that she used Jodi, an accelerant detection dog, to find places from which samples were taken for laboratory testing. The trial judge, Philip R. Volland of the Superior Court in Anchorage, also allowed Burshue to conduct a courtroom demonstration with Jodi, showing how the dog alerted to areas from which samples were then taken for analysis. Halliday argued on appeal that Judge Volland erred in allowing this demonstration.

Prosecution and Trial

Prior to trial, the prosecution gave notice that it intended to call Bushue as an expert witness and asked permission for her to conduct a demonstration before the jury that would explain how the dog was used to collect samples to be sent for laboratory analysis. Judge Volland first required that the demonstration be given without the jury present. Bushue placed Q-tips around the courtroom, some with accelerant on them, some without. Bushue further explained that Jodi had been certified since 2004, that she had never failed a scent-detection test, and that she had never alerted to a false positive.

The defense counsel objected to the demonstration “on the ground that the demonstration was not sufficiently similar to the actual investigation that was conducted. He argued that the demonstration would be prejudicial because the dog was cute, and he asserted that there was no particular reason to conduct this demonstration for the jury.”

Judge Volland concluded that Jodi’s “training and certification and prior successful track record” showed that the dog was “sufficiently reliable as a tool in fire investigation.” It is not stated whether defense counsel attempted to obtain and review the canine team's training, certification, or field records, since the only mention of the dog’s qualifications appears to have been given by the handler in testimony. Nor is any cross-examination described. The objection that the dog was “cute” has been made by defense attorneys in cases where therapy dogs help children testify in abuse cases, but may have been worth making as part of the argument regarding the prejudicial nature of the demonstration.

As to the demonstration:

“[Judge Volland] acknowledged that the demonstration in the courtroom was not the same as what Jodi would actually do in a fire investigation, but he concluded that, in spite of the difference, the demonstration would help the jurors to understand this part of the fire investigation. Judge Volland determined that the demonstration would not be unduly prejudicial. And he thought that the demonstration would be useful to show how the fire investigators decided to send certain samples from the fire scene to the laboratory to test whether residues that indicated accelerants were present on these samples.”

On the stand, Bushue qualified as an expert without objection. Given the fact that Bushue was going to make a potentially damaging demonstration, defense counsel may have made a strategic mistake in not questioning her qualifications. Bushue did not remember if she was called the same day of the fire or the day after, a curious lapse given that she would presumably have been well prepared for the witness box. Inside the house, Jodi alerted to a pile of debris, to the couch, and to an area in the kitchen. Samples were taken from these locations to the Alaska State Crime Laboratory and tested for accelerants. The lab found accelerants in three of the seven samples provided.

It is not clear if the defense questioned why four of seven samples were apparently taken from places where Jodi alerted but where no accelerants were detected in the laboratory. Was Bushue asked if these were false positives? Was any lab technician asked about his knowledge of how the samples were obtained? Did the lab have the same infallible experience with samples selected by the dog’s alerts as the handler described? Unfortunately, the appellate court’s summary of the record is sparse on some crucial issues.

Bushue then conducted the Q-tip demonstration. Apparently Bushue herself placed the Q-tips around the courtroom. It is not clear if she knew which Q-tips were scented with accelerants and which were not. If she did know, the demonstration was practically useless and should have given rise to an objection based on the possibility of cueing. There is no indication that the defense argued that someone besides Bushue place—hopefully hide—the Q-tips, and out of her presence. To make the experiment truly double-blind, everyone should have been removed from the courtroom who was going to be present during the demonstration to avoid possible third-party cueing. (See Ensminger & Papet, Cueing and Probable Cause.)


On appeal, Halliday’s counsel made a general objection to the admission of the testimony regarding accelerant detection canines under Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). The appellate court upheld the trial court’s rulings on the canine testimony in a single paragraph, without addressing the scientific objections, though it is not clear how well stated those objections were. The court’s concluding paragraph reads in full:

“We conclude that Judge Volland did not abuse his discretion in admitting the testimony and in permitting the demonstration. All that the testimony and the demonstration showed was how the fire investigators decided where to select samples from Halliday's residence to send to the laboratory for testing. Ultimately, it was the laboratory testing that determined the presence of accelerants. Accordingly, the testimony and the demonstration with Jodi were not prejudicial.”


Judge Mannheimer of the appellate court concurred, emphasizing that in finding the dog sufficiently reliable as a tool in a fire investigation, Judge Volland had made an irrelevant statement because neither Bushue nor the prosecutor had asked the jurors to rely on the dog’s detection skills as proof that the physical samples taken from Halliday’s house contained accelerants. Instead, the state relied on the subsequent laboratory analysis.

Judge Mannheimer further argued that "it ultimately made no difference to the State's case whether Deputy Chief Bushue selected the samples by using her dog, or by throwing darts, or by closing her eyes and collecting samples at random." Yet despite the ultimate reliance on the laboratory analysis, the lack of randomness in the selection of the samples quite likely convinced the jury that the dog detected not only that accelerants were used but also exactly where they were placed. This could well have given an impression of infallibility to the whole forensic process as to which the laboratory analysis was only the crowning moment.

Judge Mannheimer stated that “the demonstration of the dog’s prowess had very little relevance to the issues litigated at Halliday’s trail, and a good argument might be made that the demonstration was a waste of time.” Still, Mannheimer concluded that the demonstration was apparently a harmless, though perhaps entertaining, digression at trial. It could also be argued that if the demonstration was a waste of time, the possible prejudice involved should have resulted in its exclusion.


It is difficult to criticize a case that gives such little support to its own conclusions, and indeed provides so little background concerning the crime and the trial. Nevertheless, the demonstration with the dog gave an impression of infallibility: the dog finds the samples, the samples go to the lab, the lab verifies the presence of accelerants, and justice is done.

The demonstration was vastly different from the way a dog works at a fire scene. The dog will smell all kinds of substances, some of which, like some types of carpeting, may contain chemicals used in accelerants. (See sources cited in Police and Military Dogs, page 243.) Was this investigated by the defense? Further, the handler does not know where the accelerants might be found in a fire scene, unless there is other evidence, such as a splatter or burn patterns. Were the Q-tips visible to the dog or the handler in the courtroom? Why could they not have been hidden by someone else? Did the defense make any such suggestions?

In the fire scene investigation, why were four samples sent to the lab that did not produce accelerants? Were samples taken from areas where the dog did not alert but was allowed to smell in order to see if the dog made any false negatives? The minimalism of the decision raises more questions than are answered.

Our comments are not a criticism of Bushue’s method of investigating the fire. Accelerant detection dogs, as discussed in the chapter on accelerant detection in Police and Military Dogs, should be used to find places from which samples can be taken. They should be, as Jodi was here, a preliminary step in the investigation. Their use, however, should not be glorified as was done at trial here.

Halliday v. Alaska, 2011 WL 4107684 (September 14, 2011)

This blog was written by John Ensminger and L.E. Papet.

No comments:

Post a Comment