« Assigning videos for class? | Main | Last Call for Nominations for Annual Award for Best Untenured Article on the Law of Federal Jurisdiction »

Wednesday, September 07, 2016

Reality, Experience, Video and the Law

A recent article in the New York Times (here) caught my attention.  The article tells the story of John Lewis, who plead guilty to killing a police officer that interrupted Lewis’ armed robbery of a Dunkin’ Donuts.  A key issue in penalty portion of the case was whether Lewis acted with premeditation. He argued that he did not; he simply panicked when he saw the cop on the scene and shot him. The state, on the other hand, argued that the murder was premeditated. 

A key piece of evidence for both sides was the store's videotape of the crime.  The videotape showed Lewis shooting the officer within 2 seconds of seeing him—which allows (but does not require) the conclusion that the killing was not premeditated.  But would your view change if you watched the video in slow motion? That’s exactly what the state did in this case.  Lewis opposed the slow motion replay, arguing that it would prejudice him because it would make it look like he had more time to think through his actions.  Lewis argued the point all the way to the Pennsylvania Supreme Court, where he lost.  The court held that the video was more probative than prejudicial. A short news story is here.

My first thought on this issue was: if the state has the right to play the tape in slow motion, does Lewis have the right to play it in fast motion?  He wouldn’t actually want to do this, of course, because it would suggest that he was trying to obscure something. But why?  Why aren't we assuming that the prosecutor is trying to obscure something by playing it too slow?  I guess the answer here is comes down to our general appreciation for slow motion.  We tend to think that slow motion, generally speaking, helps us better assess what actually happened. Slow motion replay is a big part of many sports, for example.  In those contexts, we seem to think that slow motion helps us “get it right.” So why not use slow motion in the courtroom?

The authors of the New York Times article give us a reason.  In a study they conducted, they found that “juries who saw the slow-motion video were nearly four times more likely to return a unanimous first-degree murder verdict than juries who saw the regular-speed version.”  (The video they are referring to was not Lewis’ video, but a video from a different incident.  You can view the regular-speed and slow motion versions here.)  This makes sense, doesn’t it?  When you slow down the tape, the viewer’s sense of time is distorted.  And time is relevant to discerning premeditation. 

I’m on board with this, but it seems to complicate our assessment of video evidence.  If a court were to rely on this study to prohibit slow motion replay, the court's holding would be based not on the video itself, but rather the viewers’ experience of the video.  Put differently, the court would be holding that, even though the video is factually accurate, it is experientially inaccurate.  That may be a defensible holding in cases like Lewis (where we have some scientific data) but what do we do with the all other video evidence out there?

Consider Kelly v. California, a California Supreme Court case involving the admissibility of a victim impact video. After obtaining an conviction against Douglas Kelly for the murder of 19-year-old Sara Weir, prosecutors put on a 20 minute video that described Sara’s life mostly by using still photos set to music. Here. If you have the time, I encourage you to watch it; I’ve seen it several times, but watching it again today reminds me how we (or, at least I) too often forget the pain and suffering that crime inflicts on families.

The video illustrates the possible stakes in choosing between factual and experiential accuracy. Consider the last scene of the video (it starts at the 19:00 mark): it's a scene from southern Alberta with two horses riding across a river.  Sara was Native American and her ancestors apparently haled from this area, and she also loved to ride horses.  So the scene made a lot of sense.  But what’s odd here is that the scene is just stock video; the video is not of Sara or any of Sara’s family.  Nor, apparently, had Sara ever ridden horses in this place.  The final scene is thus factually inaccurate, but I think all of us can appreciate that it might nonetheless be experientially accurate.  The scene conveys much about who Sara was, what she loved, how she saw the world. 

But how far will we let this go?  What if the entire video were stock footage?  Or better yet, what if Sara’s uncle was Steven Spielberg and he created an entirely fictitious movie that, through skillful metaphor and allegory, powerfully conveyed the essence of Sara?  The question, in short, is this: when should factual accuracy give way to experiential accuracy, if ever?  I don’t have an answer here—I’m not an evidence guy and I don’t know metaphysics.  But it’s a question that I think will somehow have to be answered as video evidence becomes more and more ubiquitous. 

Posted by Jack Preis on September 7, 2016 at 09:00 AM | Permalink


Great post. It isn’t often you get an empirical study that tests the efficacy of an evidence ruling, so this is great for evidence profs. That said, I am not sure the underlying study supports a finding of unfair prejudice and thus exclusion under Rule 403. The study appears to find no statistically significant distinction between votes to convict for 1st degree (71.9% vs. 70.7%) in the condition that replicated Lewis, where mock jurors saw the video at BOTH regular speed and slow motion. As I read it, the authors get to the more compelling contrary conclusion they report at the end of the Times article by running a "bootstrap analysis (based on the meta-analytic effect size)" that turns the non-significant finding into 1.55-fold higher unanimous conviction odds. A court would probably need to understand that statistical move before citing this study to prohibit slow motion.

On your question about fast-motion, the evidence rules would not allow it unless there was some relevance. The prosecutor in Lewis argued that slow-motion allowed the jury to get valuable information from a probably-grainy convenience store video (a glance over the shoulder, etc). It is unclear what additional info the jury gets from fast motion.

Posted by: J Bellin | Sep 7, 2016 10:22:13 AM

J. Bellin- Thanks for that clarification on the statistical findings. And on the relevance of a fast-motion playback, that makes sense. But what if Lewis were to argue that the fast-motion playback would help the jury appreciate (or using the word from my post, "experience") how fast the whole thing went down from his perspective. We've all heard someone say after a car crash something like "it all happened so fast; one second I'm driving my car and the next second upside down in the ditch." Of course, we've all also heard somebody say something like this after a crash: "it's was like the entire thing was happening in slow motion; I couldn't believe it was happening right before my eyes." I have no answer here (and am out of my depth when it comes to evidence law) but it's an idea that keeps nagging me for some reason.

Posted by: Jack Preis | Sep 7, 2016 10:37:32 AM

Yes, the trick is to come up with a theory of relevance for the fast motion clip. Maybe if the defendant testified that he experienced the events sped up that way, it could work, as an illustration of his experience. Might need an expert. On the stats, I should add that I might be wrong in how I am interpreting the study results. The study authors definitely seem to believe that their ultimate findings support exclusion - which is what makes the Times article so compelling, I just couldn't find support in the actual numbers they report.

Posted by: J Bellin | Sep 7, 2016 1:08:29 PM

The lawyers for Stacey Koon and the police charged in the Rodney King beating slowed the video down to the granular level to show that King was still moving and that, therefore, the officers were justified in continuing to use force. Since it was getting into movements that were invisible to the naked eye, it clearly left the realm of factual and experiential accuracy.

I wrote about it (in the sports context) here: http://prawfsblawg.blogs.com/prawfsblawg/2016/03/what-does-it-mean-to-get-a-call-right.html

Posted by: Howard Wasserman | Sep 7, 2016 2:25:46 PM


Great post. I wonder how the "bias" introduced by watching a video in slow motion is qualitatively different than, for example, reading and analyzing in painstaking detail an email or text originally written in haste? While the mode of presentation of video evidence can affect how it is interpreted, isn't that true of any kind of evidence? Lewis's lawyers made a clever argument, but I'm not necessarily convinced that the judge didn't weigh the tradeoff correctly. Or at least defensibly.

I'm glad you mentioned Kelly, which hasn't gotten much attention. The argument around the admissibility of victim impact videos is odd, because it relies on the claim that they shouldn't be admissible if they are "too emotional." But of course, conveying emotion is precisely the point, as you recognize. While I am sympathetic to the argument that they shouldn't be admissible at all (& its extension to all victim impact evidence), I have a hard time wrapping my head around the argument that they should be admissible only if ineffective.

Incidentally, I thought you might be interested in these videos I made, based on the video evidence submitted in Scott v. Harris & State v. Kelly, combined with excerpts from the oral arguments in those cases.


Posted by: Brian L. Frye | Sep 7, 2016 7:03:56 PM

Hey Brian, thanks for your thoughts (and the links). I totally agree with your point that all forms of evidence are bound to be inaccurate to some degree. Witness testimony is very frequently inaccurate, right? That is solved (at least theoretically) through cross examination. So why not just open the doors to any and all type of video manipulation? If one side manipulates the video too much, the other side will presumably expose the exaggeration and the jury will be won over. Or will it? Maybe the best way to think about the issue is to look at the qualification of experts. We don't let experts testify until the court decides that they are qualified. Should we not allow video evidence until it has been "qualified" in some sense: Minimum pixel count? Sufficient lighting? Too-slow motion? I guess the underlying question would be: can a jury be seduced by a bad video to the same extent as they can be seduced by a faux expert.

And with regard to victim impact videos, I also agree that it's tough to figure out the principle that underlies the law in these cases. We want them to convey emotion, but not too much. While watching the video of Sara Weir, who seemed to be a wonderful person, I couldn't help but to wonder what a victim impact video would look like if Sara was a bum who lived under a bridge or even Bernie Madoff. The victim impact video seems to suggest that the penalty should depend, to some degree, on the type of person you killed. But why? To be clear, though, I'm talking out of turn here, since I haven't really studied this issue beyond just thinking of the power of video generally.

Posted by: Jack Preis | Sep 7, 2016 8:22:09 PM

Post a comment