Introduction

The 911 call from a homicide scene runs for two minutes and forty seconds. A jury will hear it played back. The question is whether they will also hear a detective tell them what the caller's tone of voice means about guilt.

That is the terrain of 911 call analysis, a technique with no robust empirical support that has nonetheless surfaced in more than 100 documented cases across 26 states. Defense attorneys ignore it at their peril.

The method traces back to a single architect, Tracy Harpster, whose original study examined just 100 calls and was intended as exploratory research. A co-author told reporters the findings were "not ready for real-world application." At least eight subsequent studies have failed to replicate the supposed guilt indicators. Yet police departments and prosecutors' offices continue to pay for training, and the technique keeps arriving in courtrooms.

The trick is how it gets there. Prosecutors rarely call the method's founders as experts, because that would trigger a Daubert or Frye challenge. Instead, a detective or dispatcher who attended an eight-hour training session testifies about the caller's demeanor, word choice, or emotional flatness, and jurors hear guilt inferences without ever learning the observations come from a disputed methodology. A Cornell Law Review article calls this "shadow forensics" because the underlying methods stay undisclosed.

What the technique actually "detects" is low emotional display, not guilt. Dr. Jessica Salerno's research found that both civilians and police officers became significantly more suspicious when callers expressed low emotion, regardless of whether the caller was later convicted or exonerated. Trauma, culture, personality, neurodivergence, and shock all shape how someone sounds in crisis. Robert Roberson, later diagnosed with autism spectrum disorder, spent more than two decades on Texas' death row after his flat affect on a 911 call was read as evidence of guilt.

The institutional landscape is shifting in 2026:

• North Carolina's Office of Indigent Defense Services issued a statewide warning to attorneys about the technique.
• The Illinois Supreme Court agreed to review Jessica Logan's conviction after investigative reporting exposed the 911 analysis at its core.
• Fair and Just Prosecution called on member prosecutors to review past cases where the method was used.
• The NACDL has labeled the technique "another example of forensic junk."

For defenders, the practical work starts before any motion is filed. 911 call transcript analysis that anchors every claim to a timestamp lets you compare what the caller actually said against the officer's characterization on the stand. Pairing that with jail call transcription and bodycam review builds the kind of grounded record that makes a Daubert challenge possible.

Faulty or misleading forensic testimony is the second most common factor in wrongful convictions nationwide. The calls that convict people are often the same calls that can free them, if someone listens closely enough.

Key Concepts

The phrase sounds scientific in a courtroom. In practice, 911 call analysis is a method with no robust empirical support that prosecutors have used in more than 100 cases across 26 states, and knowing how it enters a trial is the first step toward keeping it out.

The technique's chief architect, Tracy Harpster, trains officers and dispatchers to infer guilt from a caller's word choice, phrasing, cadence, and emotional display. The original study behind the method examined only 100 calls and was meant as exploratory research; a co-author later told reporters its findings were "not ready for real-world application." At least eight subsequent studies tried to replicate the supposed guilt indicators, and none held up under scrutiny, according to the Innocence Project.

Here's the mechanic that matters most in practice. Prosecutors rarely proffer Harpster or his co-authors as experts. Instead, a detective or dispatcher who sat through the training testifies about a caller's culpability. Because the underlying methods stay undisclosed and the founders never take the stand, the technique "largely escapes Daubert or Frye evidentiary challenge," as Cornell Law Review scholarship on shadow forensics puts it. Your cross-examination target is the officer's opinion, not a peer-reviewed method.

Watch for these markers that an officer's "gut feeling" testimony is Harpster-method analysis in disguise:

• Testimony about a caller's "flat affect," tone, or lack of emotion
• Opinions about a delay before calling 911, or about who called
• Guilt or credibility conclusions drawn from phrasing or word choice
• Training mentioned on voir dire or in a police report, but no disclosed methodology
• A detective describing "indicators" without naming a source or study

Dr. Jessica Salerno's research found that civilians and police officers alike grow more suspicious when callers show low emotion, and in studies using real 911 calls, that pattern held whether the caller was later convicted or exonerated. Trauma, culture, personality, neurodivergence, and shock all shape how people sound in a crisis. Robert Roberson's autism-spectrum flat affect was read as guilt; he spent more than two decades on Texas' death row. A University of Michigan policy brief notes the method "lack[s] robust empirical support" yet still reinforces investigative confirmation bias.

Momentum is shifting. North Carolina's Office of Indigent Defense Services issued a statewide warning to attorneys, Fair and Just Prosecution called on member prosecutors to review past cases, and the Illinois Supreme Court agreed to review Jessica Logan's conviction after ProPublica's reporting. NACDL president Nellie King has called the business "dangerous and insidious." Expect Harpster's rebuttal: he denies any innocent person was convicted through the method and says it has been used to exonerate.

Arguing that an officer's impression lacks scientific footing gets stronger when you can quote the caller's actual words against the officer's characterization. A 911 call transcript analysis with word-level timestamps lets you show exactly what was said, and when, rather than accepting a summary. Pair it with jail call transcription to build a timeline of the client's own statements across the case. Evidence Copilot's workspace keeps every AI answer cited to the exact moment it came from, always marked for your review, never a legal conclusion.

Best Practices

The most important habit you can build in 2026 is simple: listen to every 911 call in your case before the State builds a theory around it. Because 911 call analysis reaches jurors through a detective's "observations" rather than a named methodology, you will never see a report labeled as such. You have to recognize it in the transcript.

Start by mapping the testimony to the technique. ProPublica documented more than 100 cases across 26 states where law enforcement used the methods of chief architect Tracy Harpster, and prosecutors typically present the analysis through detectives or dispatchers who attended the training. When an officer starts describing a caller's "low emotional display" or "guilt indicators" as though it were common sense, you are likely looking at Harpster-method analysis in disguise. That framing matters, because the technique's founders are never proffered as experts, which is exactly how it evades Daubert or Frye scrutiny.

Next, build your own grounded record of the call. A word-level, timestamped transcript lets you put the caller's actual words next to the officer's characterization of them, and diarization keeps the dispatcher's prompts separate from the caller's responses. That distinction matters: Salerno's research found that both civilians and police officers grew significantly more suspicious when callers showed low emotion, regardless of whether the caller was later convicted or exonerated. Trauma, shock, culture, and neurodivergence all shape how people sound in crisis. Tools built for defense review, like Evidence Copilot, let you anchor every quote to the exact second of audio so your cross-examination questions point at the recording itself, not a paraphrase. The same workflow applies to post-arrest statements, which is where jail call transcription earns its keep.

Then attack the empirical foundation directly. The original study examined only 100 calls as exploratory research, and a co-author told ProPublica its findings were "not ready for real-world application." At least eight subsequent studies failed to replicate the supposed guilt indicators. Those are your cross-examination backbone.

Your motion practice should follow a consistent sequence:

• Demand the training materials and the analyst's notes, since the methods are routinely undisclosed.
• Object to lay testimony that smuggles in methodology-based credibility opinions.
• Move to exclude under Daubert or Frye if the opinion requires scientific validity.
• Offer Salerno's emotional-display research to rebut the "too calm" narrative.
• Cite the 2026 institutional backlash, including the Illinois Supreme Court's review of Jessica Logan's case.

Finally, remember that mistaken suspicion shapes the investigation that follows. The University of Michigan policy brief found these techniques "lack robust empirical support" yet reinforce investigative confirmation bias, diverting attention from more reliable evidence. Establishing the caller's actual words and timeline early gives you leverage before that bias hardens into a charging theory, and a grounded 911 call transcript analysis workflow makes that record defensible at every stage.

Implementation

The fastest way to lose this fight is to wait until trial to hear the words "guilt indicators" from a detective on the stand. Build your challenge into the case file from day one, when the 911 audio first lands in discovery.

Start with the transcript. Get a clean, timestamped, diarized transcript of the call itself, because every downstream argument depends on what the caller actually said versus what the officer later characterized. Word-level timestamps let you compare the state's narrative against the recording line by line. When you're ready to run a structured review, dedicated 911 call transcript analysis tools keep every quote tied to the exact moment in the audio.

Then hunt for the method's fingerprints. The Cornell Law Review's analysis of the technique notes that it "largely escapes Daubert or Frye evidentiary challenge" because the founders are never proffered and the methods go undisclosed. Your job is to unmask it early:

• Pull the testifying officer's training records. Eight hours of instruction, often costing up to $3,500 in taxpayer money, is all the "training" behind these opinions.
• Flag lay-testimony smuggling. If a detective or dispatcher opines that the caller's word choice or flat tone showed "guilty knowledge," that's the Harpster method arriving without an expert witness attached.
• Demand the foundation. The original study examined only 100 calls as exploratory research, and a co-author told ProPublica the findings were "not ready for real-world application."
• Cite the replication failure. At least eight scientific studies have since tested the method's supposed guilt indicators, and none held up.

Next, attack the emotional-display premise directly. Dr. Jessica Salerno's research found that both civilians and police officers became significantly more suspicious when callers showed low emotion, and the pattern held in real 911 calls regardless of whether the caller was later convicted or exonerated. Trauma, culture, personality, neurodivergence, and shock all shape crisis responses. The Robert Roberson case, where an autism-spectrum flat affect was misread on Texas' death row for more than two decades, makes the point for any jury.

Finally, use the 2026 institutional momentum. North Carolina's Office of Indigent Defense Services issued a statewide warning to attorneys, and the Illinois Supreme Court agreed to review Jessica Logan's conviction after ProPublica's reporting documented the technique in more than 100 cases across 26 states. Those are citable, on-the-record anchors for your motion practice.

One practical note on workflow: pair the 911 call with any jail call transcription from your client's detention, so you can establish the caller's actual words and timeline against the officer's characterization in one place. Grounded review keeps every claim anchored to a citable source moment, with AI output always marked for human review, which is exactly the posture you want when the state's theory rests on an undisclosed method.

Conclusion

911 call analysis has been used in more than 100 documented cases across 26 states, and the record behind it could not be thinner: the original study examined just 100 calls as exploratory research, and at least eight follow-up studies have failed to replicate its supposed guilt indicators. If you take one thing from this guide, take this: when a detective starts describing a caller's word choice or tone as evidence of culpability, that is not lay observation. That is a disputed methodology arriving without its founder on the stand.

The institutional ground is shifting in your favor. North Carolina's Office of Indigent Defense Services issued a statewide warning to attorneys. Fair and Just Prosecution asked its member prosecutors to review past cases. The Illinois Supreme Court agreed to review Jessica Logan's conviction after ProPublica's investigation documented how the technique spread. Prosecutors, judges, and defense attorneys are now calling for bans and post-conviction review, and NACDL has framed the technique as forensic junk. That momentum gives you citation-ready precedent when you object to a "guilt indicator" opinion at trial.

Your practical next steps come down to three habits:

• Listen to the 911 call yourself, from the first ring, before anyone characterizes it for you. Build a transcript with timestamps so you can pin every quoted phrase to the actual audio.
Spot the smuggling. If a testifying officer or dispatcher offers opinions about culpability from word choice, phrasing, or emotional display, ask on cross whether they attended a Harpster-method training. Eight hours of instruction is what the technique claims is enough to assess guilt from a call.Prepare the counter-frame. Dr. Jessica Salerno's research found suspicion tracks low emotional display regardless of whether the caller was later convicted or exonerated, and that trauma, culture, personality, neurodivergence, and shock all shape how people respond in crisis. The Roberson case shows what flat affect misread as guilt costs a defendant.Anticipate the rebuttal. Harpster denies any innocent person was convicted using the analysis and claims to have examined more than 2,500 equivocal death cases. Know his position before the State puts it in front of a jury.Document how the suspicion shaped the investigation. The Innocence Project puts it plainly: mistaken suspicion is rarely corrected by the investigation that follows; it shapes the investigation itself.

None of this replaces your judgment about whether the technique appears in your case. It starts with getting the caller's actual words on the record, which is why grounded 911 call transcript analysis belongs in your workflow: timestamped, diarized transcripts that let you check every officer characterization against the audio itself, with every AI answer citing the exact moment it came from and marked for your review. The same discipline applies to jail calls, where jail call transcription gives you the verbatim record prosecutors' summaries tend to smooth over.

The technique survived for decades because it operated in the shadows, introduced through lay testimony and never tested like science. Drag it into the light, one transcript and one cross-examination at a time.