Introduction to the Structure of Data in Statistics

Section 1.1: The Structure of Data

Overview of Statistics

Statistics is the discipline concerned with collecting, describing, and analyzing data to gain information about a larger group than the one directly observed. This section introduces foundational concepts and terminology essential for understanding how data is structured and used in statistical analysis.

• Statistics: The process of collecting, describing, and analyzing data, often to make inferences about a broader population.

• Descriptive statistics: Methods for summarizing and describing the main features of a dataset.

• Inferential statistics: Techniques for making generalizations or predictions about a population based on a sample.

• Sample: The group from which data are actually collected.

• Population: The larger group from which the sample is drawn and about which conclusions are made.

• Census: Data collection from every member of the population.

Key Terms in Data Structure

• Case (Unit): An individual object (such as a person, animal, country, etc.) from which data are collected.

• Identifier (Label): A unique value used to distinguish each case (e.g., country name, ID number).

• Variable: A characteristic or attribute measured for each case. Variables can be responses to survey questions, results of measurements, or other recorded features.

• Value of a Variable: The specific measurement or category recorded for a variable in a particular case.

• Record: The complete set of values for all variables for a single case.

• Observation: The value of a variable for a particular case; sometimes used interchangeably with 'record' or 'data point'.

Example: In a study measuring the body mass of mice, each mouse is a case, 'BodyMass' is a variable, and the value for mouse #7 might be 15g.

Data Tables

Data are often organized in tables, where:

• Rows represent cases (units).

• Columns represent variables.

• Cells contain the values of variables for each case.

This format is sometimes called a unit record data table or a data frame (in programming contexts).

Types of Variables

• Categorical Variable: Divides cases into groups or categories. Each case falls into exactly one category. The values are the category labels.

• Quantitative Variable: Takes numerical values for which arithmetic operations (such as addition and averaging) make sense. Quantitative variables often have units (e.g., inches, dollars).

Example (Categorical): In a survey, the variable 'Smoke' might have values 'Yes' or 'No'. Example (Quantitative): The variable 'Income' measured in thousands of dollars.

• Numbers can sometimes represent categories (e.g., 1 = Yes, 0 = No), but this does not make the variable quantitative.

• Proportions calculated from categorical data (e.g., the proportion of students who prefer a certain award) are quantitative summaries, but the underlying variable remains categorical.

Relationships Between Variables

Statistics often investigates whether and how variables are related. A relationship exists if the values of one variable can help predict or explain the values of another.

• Explanatory Variable: The variable used to explain or predict changes in another variable.

• Response Variable: The variable whose variation is being explained or predicted.

Example: In a study on penguins, the type of tag (metal or not) is the explanatory variable, and whether the penguin died is the response variable.

Sometimes, the distinction between explanatory and response variables is not clear-cut, especially in observational studies.

Using Data to Answer Questions

Data can be used to answer questions about single variables or about relationships between variables. Some questions are about causality, while others are about association.

• Single Variable Question: What proportion of people wash their hands in public restrooms? (Categorical variable: Yes/No)

• Relationship Question: Does watching cat videos improve mood? (Could involve both categorical and quantitative variables)

Not all relationships imply causation. Two variables may be related due to a third variable or by coincidence.

Summary Table: Key Terms and Definitions

Term	Definition	Example
Case (Unit)	Individual object from which data are collected	A student in a survey
Variable	Characteristic measured for each case	Height, Income, Smoke
Value	Recorded measurement or category for a variable	"Yes" for Smoke, 15g for BodyMass
Categorical Variable	Divides cases into groups	Gender, AwardPreference
Quantitative Variable	Numerical values with meaningful arithmetic	Age, Income
Explanatory Variable	Used to explain or predict another variable	Type of tag in penguin study
Response Variable	Outcome being explained or predicted	Penguin survival

Additional info:

• In experiments, the explanatory variable is often manipulated to observe its effect on the response variable.

• In observational studies, variables are observed as they naturally occur, and causality is harder to establish.