I think your problem is not that you couldn’t predict someone’s political views, but rather that you had far too much confidence in your predictive ability.
To fix this, you should generally have only weak expectations about what other people believe. Unless you know someone quite well, you shouldn’t be “shocked” to hear them express any view with a small but significant base rate in the population.
Yes! And on a related note, the overly confident prediction probably came about due to inaccurate beliefs about the correlation between things such as intelligence, educational level, socio-economic background, critical thinking skills, and political beliefs. If the degree to which these sorts of things are correlated were more accurately known (the correlation is lower than Elizabeth and most of us intuitively think), then she would have had far less confidence in her predictions and would have thought that a wider range of outcomes was feasible.
I think your problem is not that you couldn’t predict someone’s political views, but rather that you had far too much confidence in your predictive ability.
To fix this, you should generally have only weak expectations about what other people believe. Unless you know someone quite well, you shouldn’t be “shocked” to hear them express any view with a small but significant base rate in the population.
Yes! And on a related note, the overly confident prediction probably came about due to inaccurate beliefs about the correlation between things such as intelligence, educational level, socio-economic background, critical thinking skills, and political beliefs. If the degree to which these sorts of things are correlated were more accurately known (the correlation is lower than Elizabeth and most of us intuitively think), then she would have had far less confidence in her predictions and would have thought that a wider range of outcomes was feasible.
Time for a Bayesian update.