Topics
Software/Simulat.
Bibliography
Meetings
Funding
Publications
Presentations
Unpublished
Papers
|
|
| Statistics and Art:
Sampling, Response Error, Mixed
Models, Missing Data, and Inference (Presentation) |
| |
Synopsis : |
We develop a linear unbiased predictor of the expected value of a first-stage sample unit (PSU) under simple random sampling with measurement error in a finite population. The predictor is based on the expanded random variables which identify both unit and position. |
| |
Contents : |
|
| |
Print : |
|
| |
Author : |
Stanek, E.J.III
Reviewer : none. |
| |
Related : |
|
Developing Best Linear Unbiased estimator in Finite Population Accounting for Measurement Error Due to Interviewer ( Presentation) |
|
Synopsis : |
Explore how to develoop Best Linear Unbiased Estimator in Finite Population Accounting for Interviewer Error in FPMM approach and Godambe's class of linear estimator. We also explore their connections. |
|
Contents : |
Abstract. Presented as defensing proposal at Bio/Epi department of Umass,August 29, 2008. |
|
Print : |
|
|
Author : |
Ruitao Zhang and Stanek, E.J.III
Reviewer : none. |
|
Related : |
|
Rio Talk |
| |
Synopsis : |
|
| |
Contents : |
|
| |
Print : |
|
| |
Author : |
Julio M. Singer and Stanek, E.J.III
Reviewer : none. |
| |
Related : |
|
| Interpreting variances in mixed
models predictors with measurement
errors (Presentation) |
| |
Synopsis : |
Interpreting variances in mixed models predictors with measurement error. |
| |
Contents : |
International Statistical Institute Meetings, August 27, Lisboa Portugal. |
| |
Print : |
|
| |
Author : |
Julio M. Singer and Stanek, E.J.III
Reviewer : none. |
| |
Related : |
|
| An approach to estimation and prediction with measurement error under simple random sampling (Presentation) |
| |
Synopsis : |
We develop a linear unbiased predictor of the expected value of a first-stage sample unit (PSU) under simple random sampling with measurement error in a finite population. The predictor is based on the expanded random variables which identify both unit and position. |
| |
Contents : |
Abstract. Presented at the 10th Annual Research Day, Umass,March 29, 2007. |
| |
Print : |
|
| |
Author : |
Ruitao Zhang and Stanek, E.J.III
Reviewer : none. |
| |
Related : |
|
| Covariate adjustment may not be better: thresholds of relative risk for reductions in MSE (Presentation) |
| |
Synopsis : |
Adjusted rates should not be used in all settings, and in particular, not when both RRs and sample sizes are small. |
| |
Contents : |
Abstract. Presented at the International Biometric Society Eastern North American Region 2006 Meetings, Tampa, Florida, March 26-29, 2006. |
| |
Print : |
|
| |
Author : |
Li, W. and Stanek, E.J.III
Reviewer : none. |
| |
Related : |
|
| Predicting treatment means in a one-way factorial design based on a potential observable random variable framework (presentation) |
| |
Synopsis: |
We refer to the predictors of the realized treatment level mean as the best linear unbiased predictors. Comparison of the expected MSE of these predictors with the simple sample treatment group mean illustrates the advantage of this approach. |
| |
Contents : |
Abstract. Presented at the International Biometric Society Eastern North American Region 2006 Meetings, Tampa, Florida, March 26-29, 2006. |
| |
Print : |
|
| |
Author : |
Xu, B. and Stanek, E.J.III
Reviewer: none |
| |
Related : |
|
| Predicting realized cluster means in unequally sized cluster populations (presentation) |
| |
Synopsis : |
We develop design-based predictors for unequally sized clustered finite populations based on two stage sampling. Such predictors can be closely related to those in balanced designs and outperform their mixed model counterparts. |
| |
Contents : |
Abstract . Presented at the International Biometric Society Eastern North American Region 2006 Meetings, Tampa, Florida, March 26-29, 2006. |
| |
Print : |
|
| |
Author : |
Stanek, E.J.III and Singer, J.M.
Reviewer : none. |
| |
Related : |
|
| When BLUPs are BAD |
| |
Synopsis : |
We contrast the usual mixed model BLUPs with predictors of random effects based on random permutation models, illustrating the superior performance of the latter. |
| |
Contents : |
Abstract . Presented at the International Biometric Society Eastern North American Region 2004 Meetings, Pittsburgh, PA. March 28-31, 2004. |
| |
Print : |
13 pages; 3/2004 (pdf, 121 k) |
| |
Author : |
Stanek, E.J.III, Singer, J.M., Reed, G., and Li, W.; Reviewer : none. |
| |
Related : |
|
| Predicting random Effects in Group randomized trials |
| |
Synopsis : |
We discuss mixed models used in prediction of realized random effects, applied to group randomized trials. |
| |
Contents : |
Presented at the Department of Biostatistics Seminar Series, Yale Medical School October 7, 2003. |
| |
Print : |
26 pages; (pdf 285k) |
| |
Author : |
Stanek, E.J.III |
| Predicting Random Effects of Realized Clusters in Finite Populations |
| |
Synopsis : |
We discuss mixed models used in prediction of realized random effects in finite populations. |
| |
Contents : |
Presented at the International Biometric Society Eastern North American Region 2003 Meetings, Tampa, Flordia. March 31, 2003. |
| |
Author : |
Stanek, E.J.III ; Singer, J.D. |
| What are mixed models? Whay are they used? An introduction. |
| |
Synopsis : |
We discuss the ideas behind fixed and random effects, and introduce and motivate mixed models and prediction of realized random effects in finite populations. |
| |
Contents : |
Seminar in the Dept. of Public Health, UMASS, Amherst, Nov. 15, 2002. |
| |
Author : |
Stanek, E.J.III |
| Estimating
Cluster Means in Finite Population Two-stage Clustered Designs |
| |
Synopsis : |
We discuss estimates of cluster means
in a two stage random permutation model that corresponds to usual
two stage finite population sampling. While the balanced case is
straight-forward, the unbalanced setting is more complicated. We
develop two possible solutions based on reducing the dimension of
the underlying data, and illustrate their use. |
| |
Contents : |
Abstract and draft Manuscript. Presented
at the International Biometric Society Eastern North American Region
2002 Meetings, Washington D.C. March 17-20, 2002. |
| |
Print : |
40 pages; 11/15/01; c01ed37.pdf |
| |
Author : |
Stanek, E.J.III and Singer, J.M.; Reviewer : none. |
| |
Related : |
|
| Much ado about nothing:
the mixed models controversy revisited. |
| |
Synopsis : |
We revisit the classic mixed model, and discuss the
controversy about testing for random block effects. We resolve this
controversy using two model definitions by focusing attention on
the definition of a block effect. |
| |
Contents : |
Poster, Slides, and Manuscript.
Presented at the VIII Congreso Latinoamericano de Probabilidad y
Estadistica Matematica, La Havana, Cuba. Nov 15, 2001. |
| |
|
13 pages; 11/15/01; Controver3v4.pdf |
| |
Author : |
Lencina, V., Singer, J.M., and Stanek, E.J.III; Reviewer
: none. |
| |
Related : |
|
| Optimal Estimation in
A Simple Random Sampling Model |
| |
Synopsis : |
We investigate a random permutation model for simple
random sampling using an expanded population framework, and discuss
estimation problems of relevance including the mean, the value for
a position in the permutation, and the value for a realized subject. |
| |
Contents : |
Poster and Slides. Presented
at the VIII Congreso Latinoamericano de Probabilidad y Estadistica
Matematica, La Havana, Cuba. Nov 15, 2001. |
| |
|
12 page; 11/15/01; |
| |
Author : |
; Reviewer : none. |
| |
Related : |
|
| Using the Web for Teaching
with Research |
| |
Synopsis : |
We describe two web sites developed for research projects
with an eye fir their use in teaching. |
| |
Contents : |
Abstract. |
| |
|
1 page; 11/14/00; sea00d4.doc |
| |
Author : |
Eds Stanek, Ira Ockene, Arnold Well; Reviewer : none. |
| |
Related : |
|
|
Research