Using Econometrics: A Practical Guide

Citation

Studenmund, A. H. (2016). Using Econometrics: A Practical Guide (7th ed.). Pearson.

Chapter Summary

The seventh edition has sixteen chapters and two appendices (answers and statistical tables). Pearson describes it as an introduction to elementary econometrics that concentrates on single-equation linear regression, omits matrix algebra, and moves proofs and calculus into footnotes and exercises. The chapter sequence falls into three blocks: building and testing the regression model (Chapters 1–5), the problems that arise in applied work (Chapters 6–11), and extensions beyond the basic model (Chapters 12–16). All examples and exercises in this edition are estimated with Stata.

Chapter 1: An Overview of Regression Analysis

The opening chapter introduces econometrics and regression analysis, the single-equation linear model that the rest of the book develops.

Chapter 2: Ordinary Least Squares

Presents ordinary least squares (OLS), the estimation technique on which the whole book rests.

Chapter 3: Learning to Use Regression Analysis

As its title indicates, a chapter on putting regression to work in practice, placed before the formal assumptions of Chapter 4.

Chapter 4: The Classical Model

Sets out the classical model, the set of assumptions under which OLS is studied. The later chapters on specification, multicollinearity, serial correlation and heteroskedasticity deal with situations in which these assumptions are not met in real data.

Chapter 5: Hypothesis Testing and Statistical Inference

Covers testing hypotheses about regression coefficients and statistical inference; Pearson lists the F-test and confidence intervals among the topics expanded in this edition.

Chapters 6–7: Specification: Choosing the Independent Variables; Specification: Choosing a Functional Form

Two chapters on specifying an equation: which independent variables to include, and which functional form to give it. Pearson describes the book's approach as practical and example-oriented, and these chapters, with Chapters 3 and 11, are where that orientation is most visible in the contents.

Chapters 8–10: Multicollinearity; Serial Correlation; Heteroskedasticity

The three classic departures from the classical assumptions, each with its own chapter. The seventh edition's additions include the Prais–Winsten approach to generalized least squares (a remedy for serial correlation) and the Breusch–Pagan test (a test for heteroskedasticity).

Chapter 11: Running Your Own Regression Project

Devoted, as its title says, to carrying out one's own regression study, drawing together the material of the first ten chapters.

Chapters 12–16: Time-Series Models; Dummy Dependent Variable Techniques; Simultaneous Equations; Forecasting; Experimental and Panel Data

The final five chapters extend the single-equation model: dynamic time-series models (with the Dickey–Fuller test among the edition's expanded topics), models in which the dependent variable is a dummy, simultaneous-equation systems, forecasting, and experimental and panel data. Pearson notes that notation and explanations were simplified in this edition, particularly for dynamic equations, dummy dependent variables, instrumental variables and panel data.

Key Concepts

Chapters 1–5: The Regression Model

• Ordinary least squares: the estimator that chooses coefficients to minimise the sum of squared residuals.
• Classical assumptions: the conditions under which OLS has its desirable properties, used as the reference point for the rest of the book.
• Hypothesis testing and statistical inference: tests and confidence intervals for regression coefficients, including the F-test.

Chapters 6–11: Applied Problems

• Specification: choosing the variables and the functional form of an equation.
• Multicollinearity, serial correlation, heteroskedasticity: departures from the classical assumptions, each with its own tests and remedies.
• Generalized least squares: estimation that corrects for serial correlation, here in the Prais–Winsten form.

Chapters 12–16: Extensions

• Dynamic and time-series models: equations with lagged variables, and tests such as Dickey–Fuller for non-stationarity.
• Dummy dependent variables: models for yes/no outcomes.
• Simultaneous equations: systems in which explanatory variables are jointly determined with the dependent variable; instrumental variables are among the topics Pearson says were clarified in this edition.
• Panel data: observations on the same units over time.

Critical Analysis

Strengths

• Practical and intuitive by design. Pearson presents the intuitive approach, with no matrix algebra and proofs and calculus moved to footnotes and exercises, as the book's defining feature; the edition is pitched at beginning students, regression users wanting a refresher and practitioners looking for a reference.
• Built around doing a project. Specification has two full chapters, and Chapter 11 is devoted to running one's own regression project, which suits courses where students must produce an empirical paper.
• Learning support. The publisher lists Interactive Regression Learning Exercises that give feedback on modelling decisions, new Econometric Labs in the appendices for hands-on work in and after class, roughly three times as many answers to exercises in Appendix A as in the previous edition, and a website with all datasets, a Stata guide, slides and answers to the labs.
• Stata throughout. All text examples and exercises are estimated with Stata, with an explanation of the software in an appendix and on the website, so students learn a standard package alongside the methods.

Limitations

• Single-equation focus. The book's core is single-equation linear regression; simultaneous systems, limited dependent variables, time series and panel data are each given a single chapter.
• Little modern causal-inference coverage in the contents. No chapter is devoted to regression discontinuity or differences-in-differences; experiments and panel data share the final chapter.

Where It Fits in the Course

Within this course list Studenmund is the most practice-oriented regression textbook and the gentlest of the full-length texts, since it avoids matrix algebra and calculus in the main text. Stock and Watson's Introduction to Econometrics and Wooldridge's Introductory Econometrics cover similar ground with more statistical theory and a broader range of applications; students who finish Studenmund comfortably can move to either. Newbold and colleagues' Statistics for Business and Economics can supply the probability and inference background that Chapter 5 assumes.

Angrist and Pischke's Mastering 'Metrics complements rather than replaces it: Studenmund teaches how to specify, estimate and diagnose a regression, while Mastering 'Metrics explains which research designs (randomised trials, instrumental variables, regression discontinuity, differences-in-differences) make causal conclusions credible. Their Mostly Harmless Econometrics is a later, graduate-level step for students heading into applied microeconomic research.

How to Use This Book

The book suits undergraduates in a first econometrics course, students in applied fields who need to run regressions for a project, and practitioners refreshing their knowledge. Pearson states that it requires no matrix algebra; comfort with algebra and an introductory statistics course will make Chapter 5 easier.

• Core sequence: Chapters 1–7 and 11 make a workable core for a one-term course built around a regression project.
• Diagnostics: Chapters 8–10 follow naturally and are needed for any time-series or cross-section project with real data.
• Electives: choose among Chapters 12–16 according to the course; Chapter 13 (dummy dependent variables) and Chapter 16 (experimental and panel data) are useful for microeconomic projects, Chapters 12 and 15 for time-series work.

Work the exercises with Stata and check answers in Appendix A; the interactive regression learning exercises and Econometric Labs are good preparation for writing up an independent project. Students planning a causal study should read Mastering 'Metrics alongside Chapters 6, 14 and 16.