Improving the Interpretation of Electronic Voice Phenome

by Mark Leary, Ph.D.
Published in the Spring 2013 ATransC NewsJournal
Read Part 1 and Part 2

In the past two issues of the NewsJournal (Winter and Spring 2013), I described two research studies that examined the problem of EVP interpretation. The first study looked at experienced investigators’ interpretations of nearly 100 EVP, and the second one examined lay people’s interpretations of EVP that were recorded using “radio-sweep” techniques. Although no experienced EVP enthusiast will be surprised that listeners disagreed in their interpretations of the various EVP, many will find the exceptionally low level of agreement troubling.

In the first study, only 21% of the listeners agreed on the most common interpretation on average, and many of the EVP showed no agreement across listeners whatsoever. Agreement was even worse in Study 2. When people listened to EVP without knowing what the investigators who recorded them thought the EVP said, they agreed with only 6% of the words that the investigators heard. And, only 1 out of 360 interpretations perfectly matched the investigator’s interpretation. These findings are particularly troubling when we consider that the investigators presumably submitted these particular EVP because they thought that the sound clips were among the best they had recorded.

All EVP investigators know that particular EVP are often interpreted in different ways by different people, yet they often act as if they know what a sound clip actually says. The low rate of agreement in interpretations of EVP is obviously a concern for those who are interested in EVP, so in this article, I will tackle the thorny question of how EVP enthusiasts should deal with this issue.

Why Should We Care?

In many cases, whether an investigator’s interpretation of a particular EVP is “correct” may not matter very much. Those who record EVP as a personal hobby or do paranormal investigations for mere enjoyment don’t harm anyone when they confidently claim to hear something that most other people would not interpret similarly and that, in fact, might not actually be there.

But, in other cases, EVP interpretations have consequences. Most notably, when EVP are presented as the words of a deceased loved one, the messages they supposedly contain can affect people deeply. Simply believing that the words come from a loved one is sometimes reassuring to people, and if the message is positive, it can create great relief. But what about when the words seem to convey a dark or troubled message? How sure should an investigator be about the source and content of a message before delivering it to a deceased person’s loved ones? Similarly, homeowners who believe that their property is haunted sometimes invite paranormal investigators to examine the house, and such investigations often yield EVP. When should an investigator feel confident enough to relay a purported EVP message to the homeowners?

Even when the specific content of an EVP doesn’t matter much, confidently claiming to know what a particular clip says nonetheless seems dishonest and carries the risk of undermining an investigator’s credibility when other people do not hear the same thing. (Paranormal television shows are particularly bad about this, providing interpretations of EVP that are often inconsistent with what viewers themselves hear.)

What Should We Do?

As much as investigators would like all of their EVP recordings to be crystal clear to everyone, they very rarely are. Rather than sweeping this problem under the rug and pretending that EVP are clearer and less ambiguous than they really are, investigators need to address the issue head-on. Below I offer seven recommendations for improving the quality of EVP interpretations.

  1. Don’t Be So Certain. The resounding conclusion from our two studies is that investigators should not be as certain of their interpretations as they often are. Investigators sometimes feel that they have special insight into the EVP they record, but in our two studies, only a very small percentage of listeners agreed with the interpretations given by the people who recorded them. Furthermore, given the low rate of agreement and the tendency for people to overestimate the likelihood that they are correct, investigators should express their interpretations in a cautious, tentative manner that conveys that their interpretation might not be right. Too often, we hear people assert “This EVP says…” when a far more honest and defensible claim would be “I think I hear…” or “To me, it seems to sound like….”
  2. Don’t Share an Interpretation Until Others Listen. All EVP enthusiasts know that people’s interpretations of EVP are sometimes affected by what they think other people hear. In fact, it is often difficult not to hear what someone else said they heard. In our second study (NewsJournal, Spring 2013), we found that agreement with the individual words in an EVP jumped from 6% to 23% when listeners were told what the recording investigator thought the EVP said. The implication is clear: If we want to increase our chances of finding the best interpretation of an EVP, we must allow people to come to their own conclusions before hearing what other people think.
  3. Offer Alternative Interpretations. When listeners suggest different interpretations of an EVP, any interpretation that is independently offered by multiple people must be taken seriously. If the interpretation is being shared with a client – such as a grieving family or the owner of a property that has been investigated – all of the most common alternative interpretations should be presented. It’s okay to admit not knowing for certain what an EVP says and to offer several possibilities.
  4. Calculate an Index of Agreement. Every EVP enthusiast has had the experience of confidently arriving at an interpretation of a particular EVP only to find that no one else agreed with him or her. They have also had the experience of having another investigator claim “This is a Class A EVP” (which, by definition, would be interpreted similarly by everyone) when, in fact, no one else hears the same thing. Since we all assume that our own interpretations are reasonably correct (or, at least better than other people’s interpretations), the only way to find out whether our interpretation is plausible is to have several people – 10 at the minimum – independently listen to the EVP and privately record their interpretation. In this way, an investigator can see the percentage of listeners who agree with his or her interpretation (as well as possibly identify a better interpretation that more people agree with).
    But how much agreement should we require before claiming than an EVP says this or that? Each investigator must decide for him- or herself when to share interpretations with others, but let’s consider an analogy to put the problem in perspective. Imagine that your doctor detects symptoms that might or might not indicate that you have a serious illness. How sure would you like the doctor to be before he or she shares a specific diagnosis with you? And if the doctor conferred with other doctors, what percentage of other doctors would you want to agree with his diagnosis of your condition before he reported that you have a particular disease? More to the point, would you trust a doctor who said “I think that you have Disease X, but only about 20% of other doctors agree with me?” That’s roughly the average percentage of agreement that we found in our first study.
    Among behavioral researchers (such as research psychologists), the minimum agreement that is considered acceptable before data can be used is 70%. That is, if two independent researchers count, rate or interpret some aspect of people’s behavior, they must agree at least 70% of the time in their ratings or interpretations for the ratings to be reliable enough to use. That figure strikes me as a reasonable criterion. Investigators should not assert that an EVP conveys a particular message unless at least 70% of listeners independently agree.
  5. Interpret EVP Word-by-Word and Encourage Partial Interpretations. As would be expected, the results of our two studies showed that listeners agree on individual words more often than they agree on entire EVP. This suggests that EVP should be interpreted word-by-word (if not syllable-by-syllable), with listeners indicating uninterpretable syllables by an asterisk. In our first, study we had the impression that listeners who interpreted every word sometimes “heard” words that helped a phrase make sense. Listeners should not try to make sense of the entire phrase but rather should simply write down each word that can be interpreted and ignore those that are unclear.
    Having a group of people give partial interpretations of only the clearest syllables may collectively provide a good interpretation. Although this idea remains to be tested, I suspect that a group of people who each deciphered only the words (or syllables) that are clearest to them will generate a better interpretation of an EVP than any given person.
  6. Challenge Others’ Interpretations (Gently). Many investigators hesitate to question others’ interpretations when they disagree with them. Most of us do not want to provoke disagreement and conflict, particularly when we know that some people can become rather ego-involved in their interpretations. In addition, knowing how unclear most EVP are, many investigators may disagree with another interpretation yet have little confidence in their own interpretation of a particular sound clip. Yet, failing to indicate when one does not hear another person’s interpretation may give an impression of implicit agreement, leading an investigator to be more confident of his or her interpretation than is warranted.
    When challenging an interpretation, the approach should never be “You’re wrong, and I’m right,” because, on average, one’s own interpretation is no more likely to be correct than anyone else’s. Rather, the message should simply be “I’m not sure that I hear that. To me, it sounds more like ….” When such disagreements arise, as they inevitably will, the automatic and default recourse should be to get more independent interpretations, with no effort to pressure people into hearing any particular thing. The goal should be to find the best translation – not to prove that you are right.
  7. Leave Ambiguous EVP Uninterrupted. It’s okay to say “I have no idea what this EVP says.” Particularly when an analysis of agreement across several people shows little or no agreement (as occurred on many of the EVP we used in our studies), the most honest conclusion is that the EVP is uninterruptable. In some cases of uninterruptable EVP, the vocal characteristics may be so pronounced that an investigator will nonetheless conclude that the sound clip is a bona fide EVP, but that it is simply not possible to decipher it (just as one can hear voices through the wall of a hotel room but not understand what they are saying). However, in many cases, the failure to arrive at an interpretation that others independently agree with should lead an investigator to reconsider whether the sound clip is an EVP


Low agreement in EVP interpretations is the elephant in the room among those who are interested in EVP. All investigators know that low agreement is a problem, but they hate to confront it because it casts a pall on the entire enterprise of recording and interpreting EVP. Yet, failing to confront the issue simply creates more difficulties. Consistently acknowledging the agreement problem and encouraging investigators to be honest and cautious in how they assert their interpretations is an important first step. And following recommendations such as those offered here will help to restrain us from claiming more than we actually know.

Read Part 1 and Part 2

learythumbnailsmallWith a Ph. D. in social psychology, Dr. Leary is a research psychologist who studies topics related to self-awareness, motivation, and emotion. He has conducted research on topics such as reactions to social rejection, the effects of excessive self-attention, people’s concerns with their social images, and the relationship between personality and behavior. He is on the editorial boards of several scientific journals in social psychology and recently released a psychology course on DVD entitled “Understanding the Mysteries of Human Behavior.”

Editor’s Note

For an additional study of how people hear EVP, please refer to the article EVP Online Listening Trials in the ATransC online Journal.

“Radio-sweep” is a generic name for EVP thought to be formed using sound produced by sweeping a radio dial. In principle, it produces a form of EVP referred to as “opportunistic EVP.” Please review Locating EVP Formation and Detecting False Positives and Radio-Sweep: A Case Study. Also see the article on page 9: “A Two-Year Investigation of the Allegedly Anomalous Electronic Voices or EVP.”

Leave a Reply