Hoffmann J. Linear Regression Analysis. Assumptions and Applications 2010 torrent

Textbook in PDF format

I have been teaching these methods for several years and have seen many students struggle and more succeed. Thus, I know how important it is that students are familiar with some standard statistical concepts before beginning to learn about linear regression analysis. Hence, the first chapter is designed as a quick and dirty review of elementary statistics. For the reader who has not been exposed to this material (or who has forgotten it), I recommend reviewing a basic statistics textbook to become familiar with means, medians, standard deviations, standard errors, z-cores, t-tests, correlations and covariances, and analysis-of-variance (ANOVA). I also suggest that the reader take some time to learn Stata, a statistical software package designed to carry out most of the analyses presented herein. It is a relatively easy program to master, especially for students who have used some type of spreadsheet software or other statistical software. I find that Stata combines the best of all software worlds: ease of use and a comprehensive set of tools. It allows users several ways to request regression models and other statistical procedures, including through a command line, user-defined files (called .do files; which are simple programs of instruction), and drop-down menus. In this presentation, we will rely on the command line approach, although I always encourage my students to write out programs using Stata .do files so they have a record of the commands used. Recording work in Stata’s log files is also strongly recommended.
The chapters follow the typical format for books on linear regression analysis. As mentioned earlier, we first review elementary statistics. This is followed by an introductory discussion of the simple linear regression model. Second, we learn how to interpret the results of the linear regression model. Third, we see how to add additional explanatory variables to the model. This transforms it into a multiple linear regression model. We then learn about goodness-of-fit statistics, model comparison procedures, and dummy explanatory variables. Fourth, we move into an in-depth discussion of the assumptions of linear regression models. We spend several chapters on exciting (okay, too strong of a word) and mystifying topics uch as multicollinearity, heteroscedasticity, autocorrelation, and influential observations. We finish the presentation by learning about the logistic regression model, which, as mentioned earlier, is designed for outcome variables that include only two categories (e.g., “Do you support the death penalty for murder?” 0 = no, 1 = yes)