<Preface>

Statistics or data analysis has been a challenging subject for many students since it is based on mathematics, and there are difficult concepts related to probability. I experienced similar frustration when I was in college and graduate school. Ironically, despite such difficulties, my interest in data analysis grew and led me to my current research and teaching. Since I was a graduate student, I have taught data analysis courses for more than 15 years, and I have been trying to find a better way to teach this subject. From my teaching, I have learned a new path to teaching data analysis. This textbook is the outcome of my experience as a student and a teacher in the data analysis courses.

This book is an introductory data analysis textbook for college and graduate students who have not studied this subject before. Different from other textbooks in data analysis, it focuses on the methods that are commonly used in quantitative reports and research papers in social science. Particularly, it covers how the sample mean and the regression model can be widely applicable to the cross-sectional data for various purposes. On the other hand, this book does not cover some of the conventional empirical methods that could be well replaced by the OLS regression analysis, such as the Chi-square test, ANOVA, and ANCOVA. Also, to help students understand, this book includes practical examples and exercises.

This textbook is designed for a one-semester course. Most of the instructors would cover all chapters over 15 weeks by teaching one chapter within one to two weeks. After the course with this book, students would be able to empirically analyze various topics in social science using the sample mean and regression model.

February, 2024

Haeil Jung

<Contents>

Chapter 1 How do we examine our interests with data?: Distribution and mean

• Understanding our world with data

• Mapping what we want to study into numbers

• Less likely or more likely? Think about the probabilities of events

• Which group of subjects do we want to study?: The population of interest and the random sample

• Random sample assumption and sampling methods

• What useful information can we have from a sample?: sample mean and sample variance

• Normal distribution and its application: One of the most popular and useful distributions

• Alternative measures to mean: median and mode

• Chapter Summary

• Exercises

Chapter 2 Do more with the sample mean: Inference

• Sampling distribution of the sample mean and the Central Limit Theorem

• The confidence interval (CI) for the population mean μ

• Hypothesis test for the population mean μ

• How to choose an appropriate sample size in the survey for inference

• Chapter Summary

• Exercises

Chapter 3 Examining the relationship between the two quantitative variables I: Correlation coefficient and introduction to the OLS regression analysis

• Covarience and correlation coefficent

• Introduction to the OLS regression analysis

• Chapter Summary

• Exercises

Chapter 4 Examining the relationship between the two continuous variables II: Inference in the OLS regression analysis

• The normally of the error term and the sampling distribution of the OLS estimator

• The linear regression model when the sample size becomes larger

• The Confidence Interval (CI) for the regression parameter β₁

• Hypothesis test for the regression parameter β₁

• Chapter Summary

• Exercises

Chapter 5 Handling two or more explanatory variables in OLS regression analysis I: Multivariate Regression Analysis

• Partialling out and multicollinearity in multivariate regression analysis

• Omitted variable bias in the linear regression model

• Adding an explanatory variable and the efficiency of OLS estimators

• Chapter Summary

• Exercises

Chapter 6 Handling two or more explanatory variables in OLS regression analysis II: Hypothesis tests and more in Multivariate Regression Analysis

• Hypothesis tests in multivariable regression analysis

• Adjusted R-squared

• Chapter Summary

• Exercises

Chapter 7 The OLS regression analysis when comparing the outcomes of the two or more groups: Use of binary explanatory variables

• Estimating group differences in an outcome variable

• Estimating group differences in an outcome variable without the constant

• Estimating group differences using an interval variable

• Estimating group differences in a slope coefficient

• Estimating group differences in all explanatory variables

• Estimating the nonlinear relationship between an explanatory variable and an outcome variable

• Subsample analysis based on exogenous explanatory variables

• Chapter Summary

• Exercises

Chapter 8 Developing and completing the OLS regression analysis by using rescaling and functional specifications

• Rescaling of the outcome and explanatory variables

• Linearity in the OLS analysis

• Linear and nonlinear specifications in the OLS analysis

• Choosing specifications by considering three different types of causal paths

• General rules for including additional variables and making specifications in multivariate regression analysis

• Chapter Summary

• Exercises

Chapter 9 The OLS regression analysis when the variance of the error term depends on the explanatory variables: Heteroscedasticity

• Chapter Summary

• Exercises

Chapter 10 The regression analysis when the outcome variable is binary: LPM, Logit, and Probit

• Linear Probability Model (LPM): Using OLS when the outcome variable is binary

• The estimation of logit and probit models

• Statistical inference and goodness of it for probit and logit models

• Chapter Summary

• Exercises

Appendix

A. Software programs for data analysis: SPSS, SAS, Stata, R

B. How to do a reliable empirical study

C. z distribution table: standard normal curve tail probabilities

D. t distribution table: critical values of the t distribution

E. Chi-square distribution table: critical values of the Chi-square distribution

F. F distribution table: critical values of the F distribution

<Author>

Haeil Jung is a professor in the Department of Public Administration at Korea University in Seoul, South Korea. He earned his PhD degree in Public Policy from the University of Chicago, Chicago, USA. Before assuming his current role, he was an assistant professor in the Paul H. O'Neill School of Public and Environmental Affairs at Indiana University, Bloomington, IN, USA, from 2009 to 2015. Additionally, from 2012 to 2020, he served as a consultant for the World Bank, where he played a key role in the evaluation of the early childhood education program in Indonesia. His research expertise lies in policy analysis and program evaluation, particularly focusing on poverty, inequality, and related social policy interventions. He has authored numerous peer-reviewed research articles on diverse topics such as early childhood education, college education, labor market participation, immigration, fertility, obesity, incarceration, COVID-19, and empirical methods, making significant contributions to these fields. Along with his research, he has a comprehensive teaching background. He has taught introductory, intermediate, and advanced data analysis courses, as well as social policy courses, at the University of Chicago, Indiana University, and Korea University.