Data Description

Categorical Variables: Summarizing and Interpreting Data

In statistics, understanding how to summarize and interpret categorical data is essential for describing patterns and making comparisons. This section covers the key concepts and methods for analyzing one or two categorical variables, including the use of proportions, percentages, and visual displays.

• Categorical Variable: A variable that places each individual or item into one of several groups or categories (e.g., movie genre, education level).

• Distribution of a Variable: Describes the values a variable takes and how often it takes those values. Applies to both categorical and quantitative variables.

Summarizing One Categorical Variable

• Proportion in a Category: The ratio of the number of cases in a category to the total number of cases (sample size). This is a key summary statistic for categorical data.

• Notation:

  • Sample proportion:

  • Population proportion:

• Formula for Sample Proportion:

• Expressing Proportions:

  • As a decimal (e.g., 0.246)

  • As a percentage (e.g., 24.6%)

  • As a count per 1,000, 10,000, or 100,000 (e.g., 9.5 per 1,000)

• Conversions:

  • Proportion to percentage: multiply by 100 (move decimal two places right)

  • Percentage to proportion: divide by 100 (move decimal two places left)

  • Proportion to count per 1,000: multiply by 1,000 (move decimal three places right)

  • Proportion to count per 100,000: multiply by 100,000 (move decimal five places right)

• Mode: The category with the greatest proportion; useful for identifying the most common category.

• Visual Display: Bar charts are commonly used to visualize the distribution of a single categorical variable.

Example: Out of 844 Hollywood movies released from 2019 to 2023, 208 were dramas. The sample proportion is , or 24.6%.

Interpreting Proportions in Different Formats

• 0.0095 as a proportion = 0.95% as a percentage = 9.5 per 1,000 = 95 per 10,000 = 950 per 100,000

• Choose a power of 10 for counts so that the numbers are easy to interpret (not too large or too small).

Summarizing Two Categorical Variables

When analyzing two categorical variables, we are often interested in whether the variables are associated and how to summarize their relationship. The primary tool for this is the two-way table.

• Two-Way Table: A table that displays the counts for combinations of two categorical variables. One variable's categories are listed as row headers, the other's as column headers. Each cell shows the count for that combination.

Example: In a survey of 2,625 US adults, respondents were asked if they believe in 'one true love' and about their education level. The two-way table shows counts for each combination (e.g., 363 people with a high school education agreed with the statement).

Types of Proportions in Two-Way Tables

• Overall (Table) Proportion: Proportion of the entire sample in a specific cell.

  • Formula:

  • Example: 13.8% of the sample had a high school education and agreed with 'one true love'.

• Column Proportion: Proportion within a column (i.e., among those with a specific value of the column variable).

  • Formula:

  • Example: 38.6% of people with a high school education agreed with 'one true love'.

• Row Proportion: Proportion within a row (i.e., among those with a specific value of the row variable).

  • Formula:

  • Example: 49.4% of people who agreed with 'one true love' had a high school education.

Visualizing Relationships Between Two Categorical Variables

• Stacked Bar Charts: Used to visualize the relationship, but 100% stacked bar charts (where each bar represents proportions summing to 100%) are preferred for comparing proportions across groups.

Assessing Association Between Two Categorical Variables

• Compare column proportions for each row. If the proportions differ greatly, the variables are associated.

• Example: In a nutrition study, 23.3% of smokers used vitamins regularly, compared to 41.2% of non-smokers. The large difference suggests an association between smoking status and vitamin use.

• Describe the association by referencing the relevant proportions.

Summary Table: Types of Proportions in Two-Way Tables

Additional info: When using software like StatKey, you can switch variables to compute either row or column proportions, depending on which variable you want to use as the explanatory variable. This flexibility helps in making predictions and understanding associations.