Essentials of Time Series for Time Series Cross-Section Analyses

Lorena Barberia, University of São Paulo and Guy D. Whitten, Texas A&M University


Time series variables (e.g., presidential approval, public mood liberalism, GDP, inflation, education level) are extremely common in the social sciences. However, due to certain properties, these series cannot always be handled using standard regression approaches. This course serves as an introduction to the world of time series analysis. In this module, we will discuss the essentials of time series with a focus on preparing you for cross-sectional time series analysis.  We will explore the properties of time series (e.g., non-independence of observations, moving averages, unit-roots), and introduce strategies to test and model these data.

During the first four days, the course will involve about three hours of lecture time with breaks, then lunch, then three to four hours of hands on instruction in analysis that takes place in smaller groups using Stata. On the fifth day, students will present a specific project that applies the concepts introduced in the course.


This course runs January 14-18, 2019.

TEACHING FELLOW: Rodrigo Nakahara, University of São Paulo


Topic 1: For the first topic, we start with a review of standard regression assumptions. We then move into the basics of time series data. We will cover how to write time series notation, how to analyze time series data, and begin to discuss threats to inference common in time series data, including autoregression and non-stationarity.

Topic 2: For the second topic, we introduce ARIMA models, which are used for univariate time series. We will cover how to estimate these models, test for statistical issues, and make forecasts using these models.

Topic 3: For the third topic, we will focus on regression models more common in applied analysis (e.g.,, a dependent variable and multiple independent variables). We will discuss how autocorrelation is commonly treated in these models, lag structures, and interpreting the output from multivariate regression models.

Topic 4: For the fourth topic, we will discuss stationary and non-stationary data, and various approaches to modeling them. Much of this focuses on cointegration, which means that two or more non-stationary series to have a stable long-run relationship. Often, this is estimated using a model known as an error correction model. There has been a lot of recent discussions about cointegration and error correction, so we will spend some time talking about this debate.

Topic 5: For the fifth topic, we will have student presentations of a research project developed over the week. Everyone will provide feedback. If needed, we will also finish up any lectures.

A full-semester graduate-level course in multiple regression analysis.