Our expertise

In our consulting we perform custom analyses using statistical models that meet the needs of the specific project.  In particular, Statistical Innovations specializes in segmentation and latent variable modeling, and the analysis of high-dimensional data  involving large number of correlated predictors. For a sample of recent projects and clients see Our Clients.

With our Latent GOLD® software and our publications we pioneered the use of latent class models to identify meaningful segments, as well as the development of typing tools to classify new cases into the appropriate segment. During these analyses, we properly adjust for response style biases and other scale usage artifacts when necessary to obtain most meaningful segments and most accurate predictions.

We also pioneered the analyses of high-dimensional data with our CORExpress® package, which implements our Correlated Component Regression (CCR) models  to obtain reliable predictions even when there are more variables than cases!

For more information on Consulting with Statistical Innovations see Getting Started.

Jay Magidson

Dr. Jay Magidson founded Statistical Innovations in 1981. His clients have included A.C. Nielsen Co., Chemical Bank, National Geographic Society, and Pfizer. He taught statistics at Tufts University and Boston University and is widely published on the theory and applications of multivariate statistical methods. Dr. Magidson designed CORExpress®, SI-CHAID®, GOLDMineR®, and is the co-developer (with Prof. Jeroen K. Vermunt) of Latent GOLD® and XLSTAT-LG (with Prof. Jeroen K. Vermunt and Thierry Fahmy).