Seeing the Same Word Over and Over?

Source: blickpixel/Pixabay

One of the primary difficulties in studying coincidences scientifically is figuring out how many striking coincidences we should expect to encounter just by chance. After all, with as many events as we all encounter every day, some of them are bound to eventually match up in uncanny ways! What we need is to determine whether the number of uncanny coincidences we’re experiencing is greater than what we ought to experience if nothing more than chance is involved.

While this is a complicated process for most coincidences (Rawlette 2019), there are simpler methods that work for particular cases. For instance, I recently posted an easy method for testing the significance of number coincidences. Many coincidences, however, involve concepts that are expressed in words. Fortunately, at least one researcher has developed a method for quantifying those as well.

The late experimental psychologist William Braud (1983) not only developed a method for quantifying verbal coincidences but actually put it to use, with statistically significant results.

Here’s his method, as well as what he found.

Braud’s approach is tailored for cases in which a person has a feeling that they are about to experience a coincidence related to a particular word. This feeling may be the result of having already experienced a string of coincidences involving this word, or it may be based on just a subjective hunch. Either way, to begin the experiment, it’s necessary for the person to identify the “key” word that they now expect to encounter at the above chance level.

The next step is to identify a control word that is known to be used with the same overall frequency as the keyword. Braud found his control words using a frequency table published in the Journal of Verbal Learning and Verbal Behavior (Howes 1966). Today, word frequency data can easily be procured online.

Once both the keyword and control word have been determined, the person who is expecting a coincidence involving the keyword begins to read a lengthy document. This document should be fairly representative of overall word usage in the language in question and not selected because of particular relevance to the keyword. The document should be read straight through, noting both the time at which the person begins reading and the times at which they first encounter the keyword and the control word.

In his own experiment, Braud used print editions of the newspaper San Antonio Express. He read straight through the paper, “Beginning with the leftmost column of a page and reading all articles or small items beginning in a column before going on to the next column.” After 10 such trials (each on a different day, with a new paper, using different key and control words), Braud compared how long it took him to come across each keyword with how long it took him to encounter the associated control word.

In 3 of Braud's 10 trials, he encountered both words at approximately the same time (that is, within the same minute). In 1 of the 10 trials, the control word was encountered first, and in the remaining 6 trials, the keyword was encountered first.

Braud calculated that, if only the order in which the words were encountered was taken into account, the probability that these results were due to chance was p = 0.062, or 6.2%. If the actual amount of time it took to encounter each word was also factored in, then because some of the control words took so much longer than the keywords to be found, the probability that the results were due to chance dropped to p = 0.036, or 3.6%.

This means that, based on Braud’s data, the likelihood that something besides chance accounted for Braud’s word coincidences was somewhere between 93.8 and 96.4%. That is, Braud’s word coincidences were very likely more than a mere figment of his imagination.

Perhaps yours are, too. But how to know unless you put them to the test?