> My ~50 years of life experience leads me to suspect the exact opposite.
Duration/number of datapoints is less important than sampling. What is your sampling?
My own sampling is from talking to people next to me on trains around south east England. This is obviously not a completely random distribution, but it's much more representative than my friends, family, colleagues and online acquaintances.
Duration/number of datapoints is less important than sampling. What is your sampling?
My own sampling is from talking to people next to me on trains around south east England. This is obviously not a completely random distribution, but it's much more representative than my friends, family, colleagues and online acquaintances.
Our filter bubbles are incredibly strong.