Statistical Significance Research
A Brief Overview of Our Research
Using the Smith’s Bible Dictionary, we were able to compile a list of over 3000 keywords that were all between 3 and 8 characters in length. We then built specialized software that would allow us to automatically search for each keyword between an ELS range of 1 and 1500 and to calculate and record the p-value of each discovery, one book at a time.
This process ran 24 hours a day for 35 days straight and we were able to gather over 59000 samples. We then divided our samples into groups. We determined as many as 6 groups per book in most cases, because each book may consist of an array of groups and each group is determined by the character length of each sample we collected.
We then calculated the mean (average) value for each group by adding the p-vales together and dividing by the group’s total number of samples. We then calculated the significance level using the conventional .05 and dividing by the total number of samples per group.
We are now able to test for and determine statistical significance in Bible Code. Because, we have set our significance level at 5%, this means we now have only a 5% chance of producing inaccurate results, therefore not only have we achieved the ability to show evidence for true statistical significance in Bible Code, but we now have a 95% chance of being right.
Automatic Fine-Tuning Design
Our online software is responsible for fine-tuning the entire system. From the viewpoint of the system, this is an automatic process. When the world uses our online software, each result is saved to our database where they can be accessed and placed in a group.
As we continuously add more and more samples for testing, our mean values and significance levels will become more and more precise and our tests will become more accurate. Thanks to the support from thousands of people from all around the world, we have just recently added over 2000 additional samples and we are continuing to receive more each day.
Current Precision and N/A Results
Despite having gathered over 59,000 samples, we currently do not have a sufficient amount of samples to satisfy the requirements of our statistical significance test for all groups in each book. This means that, in some cases, our significance ratings may not be accurate. As a side effect you may notice a HIGH to MID rating for a keyword thought to be random.
Moreover, the same could happen in reverse, where we may rate a keyword as LOW but as our precision increases we may find that this keyword is actually MID to HIGH. So, never be too quick to accept or reject any result based on its significance level.
Furthermore, in rare cases, we may not be able to perform a significance test at all, because there were certain books where we could not locate any samples for certain groups. This was typical of keywords of 8 characters in length. So, by chance that you ever find yourself in this instance, our software will provide a significance rating of N/A to signify the event.
A Brief Overview of Our Analysis
When our initial results were compiled, we built a series of graphs based on the total number of samples in each group and when these graphs were compared, side-by-side, a pattern of normal distribution is revealed to exist throughout the entire series.
At first, these results were surprising to us, we did not expect to find any similarities in terms of what we would discover, especially a normal distribution pattern. We kept asking, ‘How is it possible for the same list of >3000 words to produce virtually the same frequency pattern in every book we searched?’ After all, we are not talking about verbatim text.
So, to find out what was causing this to happen, we processed our word list and all 39 books of the English TaNaK (JPS) to find out how many words were possible for each group and how many of these words were actually found, and it became clear why this happened.
In order to understand this concept, we must first understand what we mean when we refer to our “discoveries”. What we are referring to is a hidden message that was encoded using an equidistant letter sequence (ELS), which does not produce verbatim text.
Therefore, probabilities play a major role in what we discover.
It all goes back to what we have already known all along, which is that some words can and some words cannot produce statistical significance in Bible Code. However, we could never be sure where statistical significance actually began, until now. By graphing the mean values we were able to see the exact point where statistical significance universally begins.
Our results showed overwhelming evidence that statistical significance begins with keywords that are 7 characters in length and that significance grows rapidly beyond this point.