Deborah Glueck Apps

GLIMMPSE is an open-source application for calculating power andsample size. GLIMMPSE has been designed so that researchers andscientists with varying levels of statistical training can haveaccess to reliable power and sample size calculations. For optimumusability, GLIMMPSE provides two different modes. In Guided Modeusers receive step-by-step guided instructions for entering data inorder to obtain power and sample size outputs. In Matrix Mode usersreceive less guidance, and are assumed to possess in-depthstatistical training. By using GLIMMPSE, researchers and scientistscan calculate reliable values for power and sample size for designswith normally distributed outcomes. GLIMMPSE supports a variety ofmultilevel and longitudinal studies. GLIMMPSE can calculate powerand sample size for common statistical tests and models including:One sample t-test Paired t-test Two sample t-test Analysis ofvariance (ANOVA) Analysis of covariance (ANCOVA) Repeated measuresanalysis of variance Multivariate analysis of variance (MANOVA)Multivariate analysis of covariance (MANCOVA) The web-based versionof GLIMMPSE is available at http://glimmpse.samplesizeshop.org.Theoretical details are available athttp://www.jstatsoft.org/v54/i10. Development Team GLIMMPSE wasdeveloped by Dr. Sarah Kreidler at the University of ColoradoDenver. Dr. Sarah Kreidler leads a team of software engineersincluding Dr. Aarti Munjal and Ms. Uttara Sakhadeo. Dr. Munjal andMs. Sakhadeo have been instrumental in the extension of GLIMMPSE tomobile platforms. Statistical support and technical writing wasprovided by Dr. Keith Muller, Dr. Deb Glueck, Dr. Anna Barón, Mr.Zacc Coker-Dukowitz, and Brandy Ringham. Acknowledgements GLIMMPSEextends an existing program called POWERLIB. Bercedis Peterson andKeith Muller wrote POWERLIB for SAS PROC MATRIX (Version 0) in1983. In 1992, Lynette Keyes and Keith Muller converted POWERLIB toSAS/IML (Version 1). Keyes also created a new manual, based on thePROC MATRIX manual. Douglas Taylor wrote code for creatingconfidence limits for power values based on random estimates in2001, which Keith Muller then used to enhance POWERLIB. In 2003,Jacqueline Johnson and Keith Muller created Version 2 of POWERLIB(Johnson et al. 2009). Additionally, James Slaughter greatlyimproved the _PROBF module, and Matthew Gurka conducted extensivetests. At the same time, the related manual was significantlyrevised from the previous version. GLIMMPSE extends POWERLIB toinclude the work of Glueck and Muller (2003). Development ofGLIMMPSE 1.0.0 was funded by an American Recovery and Re-investmentAct supplement (3K07CA088811-06S) for NCI grant K07CA088811. Boththe parent grant and supplement were awarded to the University ofColorado Denver, (Deborah Glueck, PI). Additional funding wasprovided by NIDCR 1 R01 DE020832-01A1 to the University of Florida(Keith Muller, PI; Deborah Glueck, University of Colorado site PI).The funding from NCI allowed the development of the GLIMMPSEsoftware. The additional funding from NIDCR permitted extensivearchitecture changes which will eventually support power for thegeneral linear mixed model, as well as supporting beta testing andsoftware release activities. Additional support for Dr. Muller wasprovided in part by NIHNIDCR grants U54-DE019261, NIHNCRR grantK30-RR022258, NIHNHLBI grant R01-HL091005 and NIHNIAAA grantR01-AA013458-01.