911 Call Analysis: Junk Science that Imprisons Innocent People

This post contains a video, which you can also view here. To support more videos like this, head to!

Hey, let’s start with a content warning: in this video I’m going to be talking about some topics like murder, so if that’s too much for you, you may want to give this one a skip.

In fact, I’m going to dive right into the murder: several different “true crime” outlets like Dateline have covered the story of Jill Halliburton-Su, a woman who was brutally murdered in 2014. Her husband was at work when he checked their home’s security cameras and saw a masked man in the house. He called their 20-year old son to ask if it was him playing a prank. The son said he wasn’t at home, but that he would go check things out. When he got to the house, he apparently found his mother dead in the bathtub with the water running. He called 911…here’s his call, and just one more warning, it’s extremely upsetting.

As you heard, her son initially said that his mom had killed herself, before realizing that she had been bound with a belt and stabbed many times.

The cops pulled him into an interrogation room and kept him there for 13 hours, convinced that he was the murderer. Why? Well, why on earth would he have assumed his mom committed suicide? (3:23)

They were convinced it was him, but eventually, the cops checked his location data on his phone and found that he was nowhere near his home when his mom was murdered, and DNA at the scene was connected to a man who had been arrested multiple times for burglary in the past and who was eventually convicted of the crime.

Because I consume way too much true crime content, I’ve seen this case covered repeatedly. To be fair, it IS an interesting case (for reasons I won’t go into here), but every time I see it pop up it infuriates me even more: is it possible that the son was guilty? Sure. But how dare those cops assume that they KNOW someone is guilty based on the way they react in the literal seconds in which they are experiencing one of the worst tragedies that can befall a person. He found his mom, brutally murdered. When he asks a cop how HE would have felt, the cop has the audacity to say he would not have reacted the same way, with panic and confusion.

This kind of thing comes up a LOT in true crime cases: cops, prosecutors, and the general public judging someone because they don’t react the “right” way to the worst moment of their life. Not enough emotion, too much emotion. I knew it was a problem, and I’ve talked before about the fake “body language” experts who infect the true crime genre these days. But until last week I had no idea just how big of a problem this really was.

In that previous video from 2019 I talk about how “microexpressions” that betray your true emotions started as a reasonable hypothesis by a psychologist, but spun out of control when they started to be used in criminal justice, like a prosecutor admitting on Twitter that she can tell when someone is lying by the way their eyes move. Well, these days it’s even worse, because the supposed “expert” doesn’t even need to be looking at a person to know they’re lying. They can use the supposedly “scientific” new tool of 911 call analysis to identify when “the caller is the killer.”

Every month, I do a livestream where I answer any questions posed by my Patrons, and every month there’s at least one question that sends me down a rabbit hole. For March, it was this question(/”more of a comment”) from Philip: “If you haven’t already, would you consider doing a segment on Tracy Harpster, who make a bundle peddling a training program for police on, “how to identify guilt and deception from the word choice, cadence and grammar of those calling 911”.

I read that, and then the next thing I knew it was six hours later, the sun had set, and I was FULL OF RAGE. As always, the transcript that I link below this video will include links that I highly recommend you check out – in this case, the bulk of my information comes from an excellent series of investigative pieces from Brett Murphy at ProPublica. But it’s a lot of reading, so I’ll do my best to give you a summary.

In 2009, a cop named Tracy Harpster, from a small town in Ohio that rarely saw a murder, published a preliminary study in which he combed through 100 911 calls, half of which had been made by the person who was later convicted of the crime being reported. He did this to identify patterns, coming up with a list of features he noticed in the “guilty” calls, like not immediately pleading for help, not demonstrating sufficient urgency, being polite by using words like “sorry” and “thank you,” giving extraneous information, or insisting that the victim is dead when their condition isn’t 100% known to the caller.

Though Harpster had no scientific training, that kind of analysis IS normal and even necessary in science – it’s a first step that says “hey, here’s a pattern I noticed in THIS dataset.” But because the researcher at this point is specifically looking for ANY data points that stand out, literally looking for the anomalies, it’s impossible to say that those anomalies will be found in a larger dataset. For instance, if I have a bag of 100 different colored marbles, I can reach in, pull out a handful, and record what I notice: 7 blue marbles and 3 red. That doesn’t give me a definitive answer but a hypothesis: if you pull one more marble out of that bag, it has a 70% chance of being blue. The next step is to test that hypothesis by reaching in again and grabbing a new handful and seeing if the numbers line up. And then doing it again, and again, and again.

Harpster never bothered with that crucial second step of actually testing the hypothesis. Others did, with one study in 2020 and another from last year both finding no correlation between Harpster’s list and actual cases of deception. But those studies haven’t mattered, because Harpster’s preliminary analysis had already been shared by the FBI to law enforcement agencies across the country, who immediately started putting it to the test and finding great “success” at identifying criminals based on their 911 calls. These successes encouraged Harpster to start charging for 2-day training sessions, paid for by taxpayers, in which he trained investigators on how to use his magical checklist.

Investigators quickly realized that because there was no actual scientific backing for the checklist, they would have to sneak it into court cases without actually calling Harpster (or “trained” detectives) as an expert witness, since there are rules for what qualifies as expert testimony in court. ProPublica has reams of documents that catch prosecutors doing this red-handed, creating a playbook on how to use Harpster’s now-debunked pseudoscience in court without subjecting it to scrutiny:

“First, identify law enforcement witnesses who have taken Harpster’s course. Then tell them how to testify about the guilty indicators by broadly referencing training and experience. As Esteves, the prosecutor in Iowa, put it in an email: “Have them testify why this 911 call is inconsistent with an innocent caller, consistent with someone with a guilty mind.”

“Next, prime jurors during jury selection and opening arguments about how a normal person should and shouldn’t react in an emergency. Give them a transcript of the 911 call and then play the audio. “When they hear it,” a prosecutor in Louisiana once told Harpster, “it will be like a Dr. Phil ‘a-ha’ moment.” Finally, remind jurors about the indicators during closing arguments. “Reinforce all the incriminating sections of the call,” another prosecutor wrote, “omissions, lack of emotion, over emotion, failure to act appropriately.”

“Juries love it, it’s easy for them to understand,” Harpster once explained to a prosecutor, “unlike DNA which puts them to sleep.”

Murphy gives in-depth details on several cases of innocent (and potentially innocent) people whose lives have been ruined in part thanks to investigators and prosecutors who use 911 call analysis to determine guilt before they even consider other evidence. And in a follow-up article, Murphy describes how his initial investigation came as a genuine shock to judges and defense attorneys who had no idea they were encountering abject pseudoscience in the courtroom. While Harpster’s system is now well-known amongst police and prosecutors, Googling it prior to ProPublica’s investigation turned up nothing. They purposely hid the technique from the public, supposedly to prevent criminals from using it to their advantage but let’s be real, it’s also probably to prevent researchers from continuing to point out that the method just doesn’t hold up to scientific scrutiny. If this becomes more widely known, then 911 call analysis will go the way of polygraphs: something cops use to scare suspects, but which cannot be submitted as scientific evidence of guilt in a court of law.
And that’s why I’m making this video. Please go read the ProPublica pieces and share them far and wide, so more justices, defense attorneys, and innocent victims know that 911 call analysis is junk science.

Rebecca Watson

Rebecca is a writer, speaker, YouTube personality, and unrepentant science nerd. In addition to founding and continuing to run Skepchick, she hosts Quiz-o-Tron, a monthly science-themed quiz show and podcast that pits comedians against nerds. There is an asteroid named in her honor. Twitter @rebeccawatson Mastodon Instagram @actuallyrebeccawatson TikTok @actuallyrebeccawatson YouTube @rebeccawatson BlueSky

Related Articles

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to top button