News  General
Alternate Memories Data

How big is a "group"?

There's something at the heart of the Mandela Effect which has been troubling from the start.

Most people know it's when a group of people have an alternate memory of something for which there's no evidence today, but the definition of group is vague. Technically, and mathematically, it's just more than one. but that isn't very satisfactory for obvious reasons. At the other end too, it's hard to say just how many trigger it - is it more than 100? More than 1000? Who defines all this anyway? Whilst there's certainly no "Mandela Effect police" regulating this, it's also true a consensus has been obtained which is generally understood.

Real numbers

To address this, the online survey was started in Jan 2018. This randomly asked a simple "current" or "alternate" ME question, collected the anonymous data, and performed analytics. At the least this would collect some hard numbers on that group problem, and there might be a bonus as more trends/patterns might be spotted.

The analytics were published from January 2019. Already some interesting results came out, but the real excitement came in January 2020 when the 1 million question count was passed.

Recognising there are other ways and tools to perform analytics than we have at our disposal, it was decided to release this 1 million question data set into the public domain. The hope is that others can process it with a greater change of finding a pattern than we could alone. Conveniently, this data set cover the two years from 2018 to 2019.

Questions

Each test consists of 12 questions, so we can see here how the 1 million question point was reached:

tests over time 2018 2019 521x265

The decision to open source this data set was an easy one. 

Although the examples use standard SQL for the analytics, many other techniques exist including the pattern matching features associated with AI and Machine Learning. Whilst 1 million records might be considered small for those kind of features, it at least provides a start which could encourage further development of the data collection.

Since the data is anonymised, it's safe to be able to release it into the public domain.

Remember - if anyone does find anything in there not seen before, we'd love to hear about it and report your findings here!

You can access the data here, but you'll need to create an account first. You'll then also be able to participate in the forums.