Course notes for HDFS 523
1 About This Book
2 Data Cleaning
3 Describing Longitudinal Data
4 Matrix Algebra
5 Ordinary Least Squares
6 Statistical Control
7 Linear Regression
8 Logistic Regression
9 Poisson Regression
10 Marginal Effects
11 Two-Occassion Change
12 Introduction to Growth
13 Growth Curve Modeling
14 Nonlinear Growth Curves
15 Dyadic Data Analysis
16 Cluster Analysis
17 Vector Autoregressive Models
Published with bookdown

HDFS 523: Strategies for Data Analysis in Developmental Research

8.1 Categorical Data in the Social Sciences

Linear regression is a workhorse procedure of modern statistics. Our introduction to regression in this class was framed around the idea of a continuous dependent (outcome) variable. However, categorical data is extremely common in many health, behavioral and social science applications.

8.1.1 Examples of Categorical Data

Binary Variables have two categories and are often used to indicate that an event has occurred or a characteristic is present. Are you sick? Did you vote in the last election? Are you married?
Ordinal variables have categories that can be ranked. Surveys often ask respondents to indicate their agreement to a statement, how frequently then engage in a behavior, or even educational attainment.
Nominal variables occur when there are multiple outcomes that cannot be ordered. For example, left or right handedness or occupation.
Censored variables occur when the value of a variable is unknown over some range of the variable. For example, measuring hourly wages might be restricted on the lower end by minimum wage laws.
Counts indicate the number of times that some event has occurred. How many drinks last week? How many people living in a house? How many years of education? Censored and count variables are often lumped in with more traditional categorical variables under the umbrella of limited dependent variables.