sensitivy analysis

[Anthony Babinec]

The question is ill-posed. "Sensitivity analysis" can mean exploring the

implications of a fitted model. For example, suppose you are predicting

whether or not a disease occurs as a function of gender, age, and other

predictors. You obtain an estimated model using logistic regression, say.

Then, you create a series of "plug-in" cases, for example, Male-Age 43,

Female-Age21, plug these cases into the model, and obtain the predicted
outcome.

For categorical predictors, you could use the possible categories. For

numeric predictors, you could use minimum, maximum, and quartile values.



Anthony Babinec

[Dale Glaser]

In regard to Anthony's comment below, there is a very nice chapter on sensitivity analysis in  Gelman, Carlin, Stern & Rubin (1995..though there is a later 2nd ED).  Bayesian Data Analysis.

Anthony Babinec <[hidden email]> wrote:  The question is ill-posed. "Sensitivity analysis" can mean exploring the

implications of a fitted model. For example, suppose you are predicting

whether or not a disease occurs as a function of gender, age, and other

predictors. You obtain an estimated model using logistic regression, say.

Then, you create a series of "plug-in" cases, for example, Male-Age 43,

Female-Age21, plug these cases into the model, and obtain the predicted
outcome.

For categorical predictors, you could use the possible categories. For

numeric predictors, you could use minimum, maximum, and quartile values.



Anthony Babinec



Dale Glaser, Ph.D.
Principal--Glaser Consulting
Lecturer/Adjunct Faculty--SDSU/USD/AIU
President, San Diego Chapter of
American Statistical Association
3115 4th Avenue
San Diego, CA 92103
phone: [phone]
fax: [phone]
email: [hidden email]
website: www.glaserconsult.com