2

It's common for poll analysts to discuss the RCP average, and well-known that the RCP average is a simple arithmetic mean of the polls under consideration.

But what are the other rules? In particular I'm interested in:

  1. What rules govern how many polls should be averaged (this seems inconsistent between various types of elections RCP is currently covering)?

  2. What rules govern whether a poll will be included in the RCP average?

  3. What's the minimum number of polls that needs to be available before they will publish an average?

1

Not very well

It is not clear that any "rules" apply. Statistically their technique is on very shaky ground, see Why is an average of an average usually incorrect?

Their track record really isn't anything to brag about.

For further information, see Lies, damned lies, and statistics.

|improve this answer|||||

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .