When Good Confidence Intervals Go Bad: Predictor Sort Experiments and ANOVA
نویسنده
چکیده
A predictor sort experiment is one in which experimental units are allocated on the basis of the values of a predictor variable that is correlated with the response. Standard ANOVA analyses of predictor sort experiments can lead to confidence intervals whose actual coverages are poor matches to nominal coverages. Correct coverages can be obtained by adjusting confidence interval lengths by appropriate factors, or by performing analyses of covariance.
منابع مشابه
Comparison Of Hyperbolic And Constant Width Simultaneous Confidence Bands in Multiple Linear Regression Under MVCS Criterion
A simultaneous confidence band gives useful information on the reasonable range of the unknown regression model. In this note, when the predictor variables are constrained to a special ellipsoidal region, hyperbolic and constant width confidence bonds for a multiple linear regression model are compared under the minimum volome confidence set (MVCS) criterion. The size of one speical an...
متن کاملSharp Simultaneous Intervals for the Means of Selected Populations with Application to Microarray Data Analysis
Simultaneous inference is a challenge when the number of populations or the dimensionality is large. In some situations including microarray experiments, the scientists are only interested in the populations with the most extreme observations. In these cases, could we just focus on a few (K) selected observations and do the inference about the corresponding populations? Attempts including Cohen...
متن کاملArea specific confidence intervals for a small area mean under the Fay-Herriot model
‎Small area estimates have received much attention from both private and public sectors due to the growing demand for effective planning of health services‎, ‎apportioning of government funds and policy and decision making‎. ‎Surveys are generally designed to give representative estimates at national or district level‎, ‎but estimates of variables of interest are oft...
متن کاملA confidence-aware interval-based trust model
It is a common and useful task in a web of trust to evaluate the trust value between two nodes using intermediate nodes. This technique is widely used when the source node has no experience of direct interaction with the target node, or the direct trust is not reliable enough by itself. If trust is used to support decision-making, it is important to have not only an accurate estimate of trust, ...
متن کاملAnalysis of 2 Factorial Experiments with Exponentially Distributed Response Variable
This paper assesses and compares the performance of generalized linear models (GLM) with respect to log transformation and ANOVA approaches on the basis of coverage results and expected length of confidence intervals (CI) for the expected responses for exponentially distributed response variable. We also focus on the power functions of tests of hypothesis for testing significance of underlying ...
متن کامل