Skip to content

Chi-Squared Tests

Chi Squared Tests

Goodness of Fit

Use when 1 variable, many categories, 1 sample

Hypothesis

\(h_0:p_1=p_2=p_n...\) or \(p_1=..., p_2=..., p_n=...\)

\(h_a:\) At least \(n\) proportions are wrong

Conditions

  • Expected count \((E) \ge 5\)
  • Random or representative

Equations

\[ \large \chi^2=\Sigma{\frac{(O-E)^2}{E}} \]

Independence

Use when 2 variables and 1 sample

Hypothesis

\(h_0:\) Variable 1 and variable two are independent

\(h_a:\) Variable 1 and variable two are dependent

Conditions

Random or representative, and expected count \((E) \ge 5\)

Equations

\[ E=\frac{(\text{row})(\text{column})}{\text{grand total}} \]

Homogeneity

Descriptors

Use when 1 variable, many categories, and multiple samples

Hypothesis

\(h_0:\) Category proportions are the same across each population

\(h_a:\) Category proportions are not the same across each population