cross tabulation

What is cross tabulation used for?

A method to analyze the relationship between two categorical variables by comparing group percentages.

Chi-Square

• Tests if relationship exists

• p < 0.05 = significant

Cramer’s V

• Measures strength

• 0.1 weak / 0.3 moderate / 0.5 strong

What is the independent vs dependent variable in cross tabs?

• Independent = cause (grouping variable)

• Dependent = outcome being compared

When do you use row vs column percentages?

• Row % → if independent variable is in rows

• Column % → if independent variable is in columns

• ALWAYS follow the cause (INDEPENDENT) direction

How do you interpret a cross tab?

Compare dependent variable across groups of the independent variable

Example:
“Doctor group vs no doctor group → compare usage rates”

Why are column percentages often misleading?

They describe the outcome group (e.g., “among buyers…”)
→ They do NOT start from the cause
→ Cannot support causal interpretation

What does the chi-square test do?

Tests whether two categorical variables are independent or related

• p < 0.05 → significant relationship

• p > 0.05 → no relationship

What are marginal totals?

Totals for each variable separately (ignore relationship) / total number of cases in a category, ignoring any relationship between variables.

When do you use each? regression vs. cross tab

• Cross tab → categorical variables

• Regression → numerical variables + prediction

What does a regression coefficient mean?

tells you how much the outcome (Y) changes when the predictor (X) goes up by 1 unit.