Dear colleagues:
A new version of RegressIt, a very powerful and user-friendly Excel add-in for multivariate descriptive analysis and linear and logistic regression, is available for free at this web site:
https://regressit.comWhether or not you are fond of Excel as a tool for data analysis, please read on, particularly if you are an R user. A quick overview of RegressIt's features can be found in this set of slides:
https://regressit.com/RegressItSlideShow.pdfI developed it for use in an advanced course on regression and time series forecasting taught to MBA's and grad students in other programs at Duke, and it is also intended for wider use in teaching (including introductory courses, workshops, and webinars) as well as in applications. It has a sophisticated but easy-to-use interface that is designed to take maximum advantage of the interactivity and multiple dimensions of the Excel environment, with some unique features to support teaching and grading as well as to make your own analysis easier and more enjoyable. It includes very extensive teaching notes that can be embedded in regression model output worksheets in the form of cell comments, as well as a great deal of internal documentation. It's also very good for live demonstrations and computer exercises in the classroom or online.
Versions of RegressIt have been distributed to the public since 2014, and over the last year or so it has undergone significant enhancements, including an innovative menu system for model space navigation and output display, the ability to run on Macs as well as PC's, an interface for performing both linear and logistic regression analysis in R, and a separate PC version for logistic regression within Excel. The logistic regression version of the program generates highly interactive output worksheets in which cutoff values in classification tables and ROC charts can be dialed up and down after fitting a model, a useful tool for sensitivity analysis and for demonstrating the properties of a logistic model.
For those who teach or otherwise work with R (or who would just like to try it out), RegressIt includes a menu interface that allows Excel to be used as a front end for running linear and logistic regression models in R with many testing options. It writes and executes script files that produce customized output in R and which also send nicely formatted output back to Excel, allowing R to serve as a computational back end for work carried out mainly in Excel. No typing of R code is necessary, so this feature can be used by students without any programming experience, and it will give them a largely error-free opportunity to play around in RStudio. See this page and its links for details:
https://regressit.com/r-interface.htmlWhatever your prior feelings about statistical analysis in Excel, please take a look at RegressIt and try out its features to see how it provides a very effective platform for teaching and applying regression analysis, one that could even be integrated into an R-based curriculum. You may find it to be a fun sandbox in which to play with linear and logistic models even if it is not your primary tool. I welcome your feedback, and please pass this information along to colleagues who may be interested.
--Bob
------------------------------
Robert Nau
Professor Emeritus
Fuqua School of Business, Duke University
Durham NC
------------------------------