NeuroNomixer logoNeuroNomixer
  • Home
  • Blog
  • Visual Guides
  • Authors
  • Contact
Sign InSign Up
NeuroNomixer logo
HomeBlogAuthorsContactPrivacy Policy

© 2026 NeuroNomixer — Built with Next.js & Tailwind CSS

Visual Guides/Student's t, Chi-Square & F
STATISTICS

Student's t, Chi-Square & F Distributions

Three indispensable sampling distributions. Understand why small samples demand heavier tails, why categorical tests use chi-square, and how ANOVA relies on the F-distribution. Adjust degrees of freedom and watch the shapes transform in real time.

t-slider adjusted
Other distribution explored
Convergence viewed
Use-case cards: 0/3

Sign in to track progress

t-Distribution: Symmetric, heavier tails than normal. Used with small samples and unknown σ.

t-distribution (df = 3)
Normal(0,1)
heaviertailheaviertail-4-2024
Degrees of Freedom3
1As df → ∞, t → Normal50
Mean0
Variance3.000
Support(-∞, +∞)
SymmetrySymmetric

Low df → very heavy tails. t-critical values are much larger than z-critical values.

t → Normal as df Increases

Each panel shows t (solid teal) vs standard normal (dashed gray). The badge shows the density overlap: the share of probability area the two curves have in common.

Which Distribution Should You Use?

Three key sampling distributions, each with a distinct use case.

When to Use t: Small Samples, Unknown σ

The t-distribution is your go-to whenever the population standard deviation is unknown and you're working with small samples.

  • One-sample t-test: Is the sample mean different from a hypothesised value?
  • Two-sample t-test: Do two groups have different means?
  • Confidence interval for mean when population SD is unknown
  • Paired t-test: Before-and-after measurements on the same subjects

Example

Sample of 15 students (mean GPA 3.2, s = 0.4). Test if GPA differs from 3.0. Use t with df = 14.

t → Normal as n grows

When to Use Chi-Square: Categorical Data

Chi-square tests work on counts and frequencies. The distribution itself is always positive and right-skewed.

  • Goodness-of-fit: Does the data match an expected distribution?
  • Test of independence: Are two categorical variables related?
  • Test of homogeneity: Do multiple groups share the same distribution?
  • Variance test: Is a sample variance equal to a hypothesised value?

Example

Roll a die 300 times. Test if it is fair. Use chi-square goodness-of-fit with df = 5.

Always non-negative, right-skewed

When to Use F: Variances & ANOVA

The F-distribution is defined as the ratio of two scaled chi-square variables. It arises whenever you compare variances.

  • Variance-ratio and Levene's tests: Are group variances equal? (Bartlett's test answers the same question but its statistic follows chi-square, not F)
  • One-way ANOVA: Do multiple group means differ significantly?
  • Regression F-statistic: Is the overall regression model significant?
  • Two-way ANOVA: Main effects and interaction effects

Example

Three production lines, 30 samples each. Test if they have equal output variances using Levene's test, an F-test with df1 = 2, df2 = 87.

Ratio of two chi-squares
← Data DistributionsProbability Fundamentals →