Baked_In Bias: Ask Not Why
When you ask why, you are inherently baking in bias. ‘Why’ questions are answered with ‘what’ and ‘hows’. If you had total understanding then you would not need to ask such a limited question. Stop asking it and you will not be such a limited person.
Since we can not ask a question that we have never conceived of before, it is important to make our curiosity as efficient as possible.
Humans are fearful creatures that hate dying. We are wired to infer causation, its one beautiful subconscious trick we have. In the past it allowed us to see a negative incident and assume the causation in order to avoid future consequence. This process has gotten humans thus far and we had to use this process because we have had limited data. Sampling is a great way of compensating for understanding large datasets, but what if we could see everything?
Our brains normalize even the most absurd things, like crushing up rocks to paint on our eyes so that mates with think we are beautiful: make up. Maybe its Maybelline or maybe its because our brains feel safe when we think we understand. The problem with this preemptive understanding is that it bakes in bias.
Since you cant possibly know everything, stop trying.
Currently we are able to process more data than ever before. Even individually, each human processes in a day what humans used to process in their entire lifetime. The old way of processing for ‘why’ is holding us back. We think we know what we are thinking, but that knowledge is limited to what you already know. Using ‘big-data’ data sets, you do not have to use the scope of your knowledge to know the “why” or “causality” between variables, the data already knows and holds their true relationship. (ref: bias in collection)
There are dumb questions … and… dumb questions yield dumb answers.
Sampling experiments requires structuring data around specific assumptions to yield mostly-true results. Some of those assumptions lead to blind spots to obvious trends. For example: “why would you do that?” has the assumption “one normally would not do that” baked right in and might get the response “because I wanted to”, to which a person-who-asks-why might ask , “why” to which the individual might response, “felt like it”. “What do you get out of that,” might yield some personal insight about the individual (or event) “How did you do that?” might yield some mechanical insight about event. Both examples represent much more meaningful answers.
We are entering an age where that will soon be a problem of the past. Many tools exist to normalize outliers, when in many cases, the outliers are where the money is.
Consider we chop every word up and measure its correlation to positive or negative sentiment. This would negate context and isolate word-instance. Since words require context, this is an especially sensitive area for baked in bias. No matter if it generational, cultural, or demographical bias. Asking why a word is positive or negative has entirely to do with experience. Where would the word ‘fuck’ fall? If you ask my grandma: negative. If you ask my boyfriend: very positive and how soon. From this limited scope we can say that we can predict a word’s sentiment based on age, in Colton’s (my) family during years 2009–2020.
Now, go fuck…don’t ask me why.
