How accurat is the EA crowd in their predictions? For example, dose it exist any research on the brier score of the EA crowd? A common trend in most research on crowd sourcing show that independence between decisions by the individuals of the crowd is a important factor that makes the crowd more accurate (L. Hong, 2004) (J.P. Simons, 2011). Futhermore, some research show that the more negative corrolated the decisions are, the better (C. P. Davis-Stober, 2014). In other terms, the more disagreement between the individuals within the crowd the better their predictions are. Knowing this, what potential biases may the EA crowd have that would destrupt the collectiv precision of the group? Dose it exist some topics on witch the EA crowd is in concerningly high agreement on?
To summarize, my two main questions are:
What would be the expected accuracy if one where to crowd source the EA crowd?
I recommend adding more detail to this question, so that people have a better sense of what you are looking for. It might also help to explain why you are asking, whether there is specific data you are looking for, etc.
Not sure if this is what you have in mind, but Metaculus records the track record of its users’ predictive accuracy (see Brier score for the Community prediction).
A lot of their users are EAs I think, though it is probably not a representative sample.
[Question] What is the wisdom of the EA crowd?
How accurat is the EA crowd in their predictions? For example, dose it exist any research on the brier score of the EA crowd? A common trend in most research on crowd sourcing show that independence between decisions by the individuals of the crowd is a important factor that makes the crowd more accurate (L. Hong, 2004) (J.P. Simons, 2011). Futhermore, some research show that the more negative corrolated the decisions are, the better (C. P. Davis-Stober, 2014). In other terms, the more disagreement between the individuals within the crowd the better their predictions are. Knowing this, what potential biases may the EA crowd have that would destrupt the collectiv precision of the group? Dose it exist some topics on witch the EA crowd is in concerningly high agreement on?
To summarize, my two main questions are:
What would be the expected accuracy if one where to crowd source the EA crowd?
What potensial biases may the EA crowd have
Sources (lazy form):
L. Hong: https://www.pnas.org/content/101/46/16385.abstract
J. P. Simons: https://www.jstor.org/stable/10.1086/658070?seq=1#page_scan_tab_contents
C. P. Davis-Stober: https://arxiv.org/ftp/arxiv/papers/1406/1406.7563.pdf
I recommend adding more detail to this question, so that people have a better sense of what you are looking for. It might also help to explain why you are asking, whether there is specific data you are looking for, etc.
Not sure if this is what you have in mind, but Metaculus records the track record of its users’ predictive accuracy (see Brier score for the Community prediction).
A lot of their users are EAs I think, though it is probably not a representative sample.