The Gateway EVP Project Phase 1 Methodologies

We’ve had a lot of questions and comments about Gateway since we posted the unofficial final results here on our page in July of 2013.   I say “unofficial” because Gateway is still on-going but there were many requests for periodic updates during phase 1 of the project – the study itself.  Phase 2, which is currently underway, is the EVP experiment.

I wanted to take a minute to give everyone a little breakdown of some of our methodologies.  I can only divulge certain things because we are still under non-disclosure.  I will provide as much information as possible without violating the contract.

Over the years, we heard many questions regarding the phenomenon of EVPs.  Our cynosure of the occurrences was the common hypotheses did not seem to match the observations accurately.  In fact, the hypotheses, when considered closely, seemed to be a case where fundamentally incorrect assumptions were based on simple observations.  One of the most commonly accepted ideas was EVPs were not heard at the time they were recorded because they are sound waves occurring at frequencies below the range of audible human hearing.  Hence, the members of IRG, set out to find out if this assumption was accurate.  Additionally, we questioned whether or not there was a specific range of frequencies applicable to EVPs like that of speech and hearing.  Was there a minimum and maximum range they could or did occur in?  There were additional questions we wanted to explore as well.  Thus, the Gateway EVP Project was born.

We began the study researching the phenomenon and all that entailed.  Based on our understanding of the characteristics, behaviors, and functions of various waves and their frequencies, we outlined the parameters of the study and set out to gather samples.

We spent months working on criteria and parameters.  The very first thing we had to establish was a definition for the phenomenon we were attempting to study.  For the purposes of the study, we defined EVP as anomalous recordings captured on electronic media that are not heard at the time of recording.  We also further defined EVP for this study as that which was recorded on either a digital or analog voice recording device.  After defining EVP for the purposes of the study, we next had to identify all of the criteria.  As a result, we generated the Guidelines For Inclusion (GFI):

1.  It had to be EVP only; Verbal or vocal in nature.  Examples of EVPs that were vocal or verbal in nature included, but were not limited to:

a.  Words, phrases, sentences;

b.  Screams, yells, screeches;

c.  Laughs or giggles;

d.  Coughs; and

e.  Crying.

2.  It had to be obvious, in every sense of the word, that those present during the recording did not hear it at the time.  If there was any indication that someone in the recording “heard something”, tags something, or says anything even remotely close to “did you hear that”, regardless of to what it may have been referring, the entire recording was excluded from the study.  Any acknowledgement was excluded – “did you hear that”, “what was that”, “that was loud”, “was that outside”, etc.  This was to avoid potential contamination of the results and ensure accuracy.

3.  IRG’s WIND principle was to be remembered at all times.  The WIND principle simply states When IN Doubt, throw it out.  It is the same principle many groups use when analyzing evidence.  If the operator doubted or questioned the authenticity, reliability, or validity of any sample at any time, the entire sample was excluded.

4.  Samples had to be original copies only with no alterations of any kind.  It had to be the raw recordings.  No enhancements, no pitch adjustments, etc.  Absolutely NO changes of any kind.  Any sample in doubt was completely excluded.

5.  Samples had to be audio files only from a digital or analog voice recorder.  No audio from any other type of equipment was included.

6.  We required specific information about the sample, which included, but was not limited to:

a.  Date & time of recording;

b.  Make & model of the equipment it was recorded on; and

c.  Names of those present during the recording.

7.  Samples had to be submitted in only one of two file formats: either mp3 or wav.

8.  Samples had to be submitted via email.  We did not accept any samples from other sources, such as a website.  It had to be an emailed sample to an email address we provided to them.

9.  The samples had to be emailed by the original owner of the files unless otherwise permitted.  Owners maintained all rights to their submissions and no information, other than that which was immediately relevant to the study, was to be released to third-parties without prior written consent of the owner of the file and the IRG director.

10.  For the purposes of the study, we were not concerned with content (“what was being said”) in the recording.

11.  At least three people had to agree that there was “something said” present in the recording.  Those three could only be from IRG or from one of our partners assisting with the project.  If we were not absolutely positive there was something anomalous present, the sample was excluded.

12.  Samples had to be analyzed under IRG’s Objectivity and Impartiality Principle.

13.  We only included samples where the potential EVP was absent any other type of noise or voices.  For instance, an EVP that came in on top of an investigator’s voice would be excluded.  If the voice occurred between investigators speaking, and no other investigator voices occurred at the time, it would have been included, assuming it met the other conditions.

If any sample did not meet all of these stipulations, it was excluded.

Prior to beginning the study, we designed an EVP Intake Form to document individual EVPs and their file information.  Additionally, we established an EVP Master Database to record the information obtained from the intake form.

When we completed the criteria and parameters, we next set out to collect samples.  A mass notice was sent out on various social media to request samples be sent in and provided the information on how to do that.  Soon, samples started pouring in.  We listened to the submitted samples and analyzed them for possible inclusion, making sure they satisfied all of the mentioned criteria.  Once it was determined the sample was, in fact, usable, we analyzed the waveform.  We then manually calculated the period and frequency of each sample.

** And before anyone says it – Yes, we are aware that audio recordings contain many different frequencies and we did take that into consideration.  Our goal was to determine a potential range of frequencies rather than trying to isolate individual ones.**

Each sample, regardless of usability, was assigned a file number and corresponding information was loaded onto the database.  Our goal was to have at least 1,000 viable samples included from which to base the results as we felt this would be a sufficient representation.  Of course, we are continuing to work on this study (refining as necessary) to make the results even more accurate.  The database was generated in an Excel spreadsheet with several cells formatted with various formulas for automatic generation of data when certain information was recorded.  This also allowed us to automatically generate results and graphs as needed to finalize the results and prevent, as much as possible, operator error.

I hope this answers many of your questions.  We welcome constructive criticism so please feel free to play devil’s advocate.  Afterall,  any information that helps us is greatly appreciated.

Please note:  While we certainly welcome and encourage constructive criticism, we will not tolerate disrespectful manners or language.  This is an attempt to scientifically scrutinize one aspect of EVP and expect responses, comments, questions, and feedback to remain respectful.