ARCCA Thoughts

The Karen Read case heated up this past week as the Commonwealth accused Read’s defense team of violating a protective order with respect to their interactions with the experts from ARCCA – who were originally retained by the feds – and misrepresenting to Judge Cannone (and potentially the jury) the nature of the defense’s relationship and interactions with those experts. While it’s clear that Judge Cannone is upset and wants answers, I’d rather not speculate as to the seriousness of the issue or how Judge Cannon is likely to react. We’ll know the answers to that soon enough.

What I would like to address is the overall significance of the ARCCA experts’ testimony, which I think commentators on both sides of this case have either overstated or ignored.

First, among those who believe Read is guilty, many have suggested, even before this week’s controversy, that the ARCCA experts are biased and their analysis should be ignored. I think that’s a mistake if you’re interested in the truth. While I don’t mean to minimize the potential misrepresentations made by defense counsel, the reality is that the ARCCA team was hired by the U.S. Attorney’s Office and had already completed their report before interacting with Read’s attorneys. The feds also appear to have given them a fair amount of latitude to determine what theories to test. That differentiates them from typical defense experts, who are hired to try to justify a particular conclusion (see, e.g., Richard Green).

Moreover, even if they had been traditional defense experts, hired and paid exclusively by the defense team, that wouldn’t mean their analysis could just be waved away as biased. The prosecution would still need to substantively address and, if necessary, rebut their analysis (again, as they did quite well with the defense’s IT expert, Richard Green). And the reality is that the prosecution did a really poor job of doing this in the first trial, in large part because they did not have any comparably credentialed experts of their own.

Those of us who are inclined to believe Read is guilty really do need to substantively engage with the testimony of the ARCCA team and assess if/how our beliefs can be reconciled with that testimony.

On the flip side, those who believe in Read’s innocence routinely misrepresent what these experts actually said in court. Here’s a good recent example from Aidan “Turtleboy” Kearney:

And here’s a tweet I remember quite well from podcaster Brandi Churchwell, who prides herself on being much more of an objective arbiter of the facts. This was right after the ARCCA experts testified in the first trial.

With all due respect to Brandi, this is a comically absurd over-reading of what these experts actually said and betrays a real lack of understanding as to how to interpret expert testimony in the context of a trial.

The first key to understanding and contextualizing the ARCCA testimony is to distinguish between what they actually tested and what were merely off-the-cuff opinions offered at trial.

As their testimony makes clear, they only tested two narrow theories. First, they tested a theory of their own creation, specifically that the taillight on Read’s SUV could have been damaged by John O’Keefe throwing the cocktail glass he was holding at the SUV from close range. As I explained in a prior post (and will discuss further below), I think this is a possibility that both the defense and prosecution should take more seriously. Dr. Wolfe testified that he actually built a small cannon in his lab and was able to replicate the SUV’s taillight damage by firing a cocktail glass at the taillight at 37 mph (which he estimated to be a reasonable speed for an adult male to throw such an object).

Regardless of what you make of this finding, it quite obviously does not rule out the possibility that O’Keefe was subsequently hit by the SUV. Indeed, as I’ll discuss shortly, if the taillight was already broken at the time of the collision, it could very well explain some of the other issues these same witnesses identified in their testimony.

The second theory the ARCCA team tested was whether O’Keefe’s head could have collided with the SUV’s taillight. To test this, they dropped a dummy “head” onto a taillight fixture at a speed of 15 mph. Dr. Wolfe testified that they chose this speed because they needed a speed fast enough to break the taillight but not so fast that the head would be obliterated.

According to Dr. Wolfe, this test resulted in more damage to the test taillight than what was observed on Read’s SUV. The ARCCA team also concluded that the force exerted on O’Keefe’s head in this scenario would likely have caused more damage and bruising than what was actually observed on his body, as well as likely causing cervical spinal injury.

That’s it, though. Those are the only two scenarios the ARCCA experts actually tested, neither of which were the particular theory presented by the prosecution.

But that wasn’t the extent of their testimony. Judge Cannone granted the defense team quite a bit of leeway to ask the ARCCA witnesses questions that went beyond the scope of their testing.

For example, on the stand, both ARCCA witnesses were asked whether the SUV’s taillight could have been damaged by a collision with O’Keefe’s outstretched arm. Despite not having tested this scenario, Dr. Wolfe opined that such an impact would likely have caused more damage to the SUV’s liftgate and other areas beyond just the taillight fixture. Again, this appears to have just been his gut opinion, not the result of any actual testing. Moreover, this entire exchange with Alan Jackson was incredibly imprecise. There was no discussion of what speed the vehicle was moving or clarity as to the specific position of O’Keefe’s arm in this hypothetical.

Dr. Rentschler’s testimony on this point was equally imprecise. He testified that if the vehicle collided with O’Keefe’s arm at a speed sufficient to break the taillight, he would have expected to see more bruising and possibly fracturing, as well as less uniformity in the depth of the abrasions on his arm. Though not supported by any specific testing, this does at least strike me as a reasonable objection to the Commonwealth’s theory of the accident. It does seem, intuitively, that if the SUV made contact with O’Keefe’s elbow at nearly 24 mph, the damage would be more severe that what was noted in the autopsy.

Dr. Rentschler also testified (again without testing) that he did not believe that impact with a grass surface could have caused O’Keefe’s fatal skull fracture. He believed impact with concrete, asphalt, or potentially hard frozen ground (not grass) could have caused the injury.

There was also a long, very imprecise exchange in which Dr. Rentschler expressed skepticism that any kind of “side swipe” scenario could have “projected” O’Keefe’s body any significant distance. But in addition to going beyond any testing they conducted, this testimony was prompted by a series of vague and leading questions by Alan Jackson premised on the idea that O’Keefe was projected 30 feet through the air after the collision. As I’ve noted previously, the distance from where O’Keefe’s missing shoe was found (the logical point of impact) to where his body was found, is nowhere near 30 feet. The total distance between the shoe and where his feet appear to have come to rest is more like 5-10 feet.

Here’s what I think is a fair summary of the actual testimony of the ARCCA experts:

• Based on testing, they concluded that the taillight could have been damaged by O’Keefe throwing the cocktail glass at it.

• Based on testing, they concluded that it was unlikely that O’Keefe’s head injury was caused by direct impact with an intact taillight fixture at 15 mph or greater.

• Based on experience, they believed that O’Keefe’s arm injuries were inconsistent with impacting an intact taillight fixture at high speed (>15 mph).

• Based on experience, they did not believe O’Keefe’s fatal head injury could have been caused by hitting his head on grass.

• While they did not do any testing of side swipe collision scenarios, they were generally skeptical of any high speed impact scenario due to the lack of bruising and fracturing observed on O’Keefe’s body and the lack of significant injuries anywhere other than the back of his head.

Notably, this comes nowhere close to a conclusion that “Karen Read’s SUV did not hit John O’Keefe” or that “the laws of physics dictate that it couldn’t have happened.”

It is easy to come up with collision scenarios that their testimony does not rule out. Indeed, under cross-examination, Dr. Rentschler readily conceded that even a collision at negligible speed could have caused O’Keefe’s head injury if it knocked him off his feet and his head landed on a hard surface.

What’s perplexing to me is that the ARCCA experts’ own analysis and testing suggests a fairly obvious and straightforward way of connecting their own dots. Specifically, if you posit that the taillight was initially broken by the cocktail glass, which was clearly their preferred theory, then you no longer need to assume a high speed collision with O’Keefe. The only reason they assumed that the collision took place at high speed, for testing purposes, is because they believed a low speed collision could not have shattered the taillight.

But if the taillight was already broken, even a very low speed collision could have resulted in O’Keefe’s injuries. There would just need to be sufficient force to knock him over onto a hard surface (like the frozen dirt area right next to where he was found).

As for his arm injuries, they could have been caused by contact (at lower speed) with the already broken taillight fixture or with the frozen bushes or dirt on the ground. Nothing about the ARCCA analysis or testimony is inconsistent with the scenario I just posited.

Moreover, even higher speed collision scenarios should not be ruled based on their analysis. The fact is that they did not test any collision permutations involving the parts of O’Keefe’s body most likely to have made contact with the SUV (specifically his arm and his right temple). As Dr. Wolfe noted in his testimony, a glancing blow would not result in the full force of the vehicle transferring to O’Keefe. It would cause him to spin. But depending on the angle of impact, some of that force could also have propelled him (while spinning) a few feet back into the yard. Without having done any actual testing of this scenario – which happens to be the scenario actually posited by the Commonwealth – it is simply not reasonable to conclude that the ARCCA team’s testimony rules it out.

Which brings me to my final point. It never makes sense to form an opinion about a case by looking at expert testimony in isolation, no matter how credible the expert is. Experts are wrong all the time, and all they are even purporting to do is assess the likelihood of a certain thing happening, based on a set of assumed premises. But the reality is that unlikely things happen all the time. You have to view their analysis within the larger context of known facts and evidence.

In this case, we know for a fact that O’Keefe’s phone stopped moving right about the time Read left 34 Fairview. His phone was found with his body, just a few feet away from the place where we know Read dropped him off. His shoe was found right at that dropoff point, next to pieces of Read’s shattered taillight. This is the context in which we need to place the testimony of the ARCCA experts. It may well be the case that the exact physics of the collision theorized to have resulted in O’Keefe’s death are unlikely and would be difficult to replicate. But that doesn’t mean it didn’t happen.

Whatever you assess the likelihood of that scenario to be, viewed in isolation, you need to compare it to the extremely limited universe of possible alternatives. In this case, the only real conceivable alternative is that, after O’Keefe breaks the taillight with the glass (thereby explaining the presence of taillight pieces at the scene) and Read drives away, someone else suddenly appears on the scene and gets into a near instantaneous altercation with O’Keefe, somehow knocking him out of his shoe and fracturing his skull in the process.

Needless to say, that is also an incredibly unlikely scenario and one that is very difficult to reconcile with all of the other evidence in the case, including the testimony of all of the people who were in the house. It doesn’t make any sense to rule out one possibility because it is deemed to be unlikely when the only possible alternative explanation is far more unlikely. For whatever reason, this fundamental logical error pervades almost all FKR thinking about this case.

Final note:

I think one of the ways the prosecution went astray in the first Read trial (and why the ARCCA testimony was particularly damaging) was in tying themselves to a very particular theory of the collision. Based on the Tech Stream data, Trooper Paul pointed to a slight change in speed (from 24 mph to 23 mph) and suggested that this was the moment the SUV made contact with O’Keefe. While it’s certainly possible that’s what happened, it was clear that Trooper Paul did not have a particularly well-informed understanding of that data and had not done any kind of testing to see if a collision at that speed was consistent with the physical evidence or injuries observed.

This decision opened the door to some reasonable skepticism from the ARCCA experts. While I am certainly no crash reconstruction or biomechanical expert, it does seem strange to me that a collision at that speed would not have resulted in more damage to O’Keefe’s body, whether in the form of bruising or bone fractures or both. And other than the tiny speed blip in the Tech Stream data, there’s no reason whatsoever to assume that the collision happened at full speed. Obviously the SUV eventually slowed down to zero, before shifting back into drive. A collision at any point along the spectrum of 0-24 mph could have been enough to knock O’Keefe over. And even if we disregard the cocktail glass and assume the taillight was still intact prior to the collision, the damage to the taillight could have happened at a much lower speed (per ARCCA’s own analysis). And if you posit that the collision happened a lower speed, most of ARCCA’s objections carry less weight.

So if I were Hank Brennan, unless new data or new expert witnesses make it crystal clear that the vehicle was moving at 24 mph at the exact moment of impact, I would stay away from over-interpreting that isolated data point and leave myself room to argue that the collision could have happened at a much slower speed.

Addendum: While the actual report that ARCCA provided to the USAO in connection with the federal investigation is impounded and has never been shared publicly, Hank Brennan shared some more details about it in the Feb. 25 hearing. Intriguingly, he suggested that there may have been four findings in the report, while only two were discussed at trial. It will be interesting to see if these other findings are discussed at the second trial (or in a pre-trial voir dire session).