Revising opinions in statistics: Difference between revisions
wikademia>Henrygb No edit summary |
typo (distribtion -> distribution) |
||
| Line 3: | Line 3: | ||
*A [[statistical model]] for the data generating process is assumed. The model might specify that the data follows a normal distribution with an unknown mean. | *A [[statistical model]] for the data generating process is assumed. The model might specify that the data follows a normal distribution with an unknown mean. | ||
*The researcher describes his opinion of a [[prior probability|prior]] distribution for the unknown [[parameter]]s of the model. So the prior | *The researcher describes his opinion of a [[prior probability|prior]] distribution for the unknown [[parameter]]s of the model. So the prior distribution of the unknown mean might be a [[normal distribution]] centered at 10 with a [[standard deviation]] of 2. | ||
*The data is then observed, and with the [[likelihood function]] of the observed data and the probabilistic description of his opinion, the researcher can calculate (using [[Bayes' theorem]]) the appropriate opinion consistent with both sources of information. This is called the [[posterior probability|posterior]] distribution. | *The data is then observed, and with the [[likelihood function]] of the observed data and the probabilistic description of his opinion, the researcher can calculate (using [[Bayes' theorem]]) the appropriate opinion consistent with both sources of information. This is called the [[posterior probability|posterior]] distribution. | ||
Revision as of 20:33, 21 January 2005
Researchers who use personal probability can proceed as follows:
- A statistical model for the data generating process is assumed. The model might specify that the data follows a normal distribution with an unknown mean.
- The researcher describes his opinion of a prior distribution for the unknown parameters of the model. So the prior distribution of the unknown mean might be a normal distribution centered at 10 with a standard deviation of 2.
- The data is then observed, and with the likelihood function of the observed data and the probabilistic description of his opinion, the researcher can calculate (using Bayes' theorem) the appropriate opinion consistent with both sources of information. This is called the posterior distribution.