Reference

Statistics Formula Sheet

Every formula from the course, with a plain-English reading next to each one. Print it, fold it, tape it inside your notebook. Click any course heading to revisit the lessons where the formula is a toy you can play with.

Need a critical value? → Tables & calculators

More one-page printables: the which-test cheat sheet, the APA reporting cheat sheet, and the assumption-checks cheat sheet.

Stats 1: Introduction to Statistics

ConceptFormulaRead it as
Meanx̄ = Σxᵢ / nAdd everything, divide by the count.
Sample variances² = Σ(xᵢ − x̄)² / (n − 1)Average squared distance from the mean (n − 1 because it's a sample).
Standard deviations = √s²Typical distance from the mean, back in the original units.
Z-scorez = (x − μ) / σHow many SDs a value sits above or below the mean.
General addition ruleP(A or B) = P(A) + P(B) − P(A and B)Add both, then take the overlap back out once. Drop the last term only when the events cannot both happen.
General multiplication ruleP(A and B) = P(A) · P(B | A)The second factor is a plain P(B) only when the two events are independent.
Discrete random variableμ = Σ x·p    σ² = Σ (x − μ)²·pWeight every value by its probability. The SD is √σ², as always.
Binomial mean and SDμ = np    σ = √(np(1 − p))The shortcut for X ~ B(n, p). The two rows above reach the same answers the long way.
Standard error of the meanSE = s / √nHow much x̄ wobbles from sample to sample.
z-interval (σ known)x̄ ± z* · σ/√nz* = 1.645 / 1.960 / 2.576 for 90 / 95 / 99% confidence.
Margin of errorm = z* · σ/√nHalf the width of the interval. Confidence level, σ and n are the only three things that move it.
Sample size for a marginn = (z* · σ / m)²Round up. Halving the margin costs four times the data.
One-sample z (σ known)z = (x̄ − μ₀) / (σ/√n)The gap between sample and claim, measured in standard errors.
Confidence interval (mean)x̄ ± t* · SEEstimate ± margin of error; t* from df = n − 1 (≈ 2 for 95%).
One-sample tt = (x̄ − μ₀) / (s / √n)Distance from the reference value, in standard-error units. df = n − 1.
Paired tt = d̄ / (sd / √n)The one-sample t, run on each person's difference score.
Independent tt = (x̄₁ − x̄₂) / SEdiffThe gap between two means over its standard error (Welch: no pooling).
Two-sample standard errorSEdiff = √( s₁²/n₁ + s₂²/n₂ )Also the SE in the interval (x̄₁ − x̄₂) ± t*·SE. By hand use the conservative df = min(n₁ − 1, n₂ − 1); software uses Welch–Satterthwaite.
SD of a sample proportionSD of p̂ = √( p(1 − p) / n )The spread of p̂ from sample to sample, centered on the true p.
Confidence interval (proportion)p̂ ± z* · √( p̂(1 − p̂) / n )The interval is estimating p, so p̂ goes into the standard error.
One-proportion z testz = (p̂ − p₀) / √( p₀(1 − p₀) / n )The test assumes H₀ is true, so p₀ goes in instead of p̂.
Two proportions, pooled p̂p̂ = (x₁ + x₂) / (n₁ + n₂), SE = √( p̂(1 − p̂)(1/n₁ + 1/n₂) )Under H₀ there is only one proportion in the world, so both groups estimate it together.
Cohen's dd = (x̄₁ − x̄₂) / spooledMean difference in SD units. 0.2 / 0.5 / 0.8 ≈ small / medium / large.
Empirical rule68% · 95% · 99.7%Share of a normal distribution within ±1, ±2, ±3 SD.

Stats 2: ANOVA & Regression

ConceptFormulaRead it as
Between-groups SSSSG = Σ ni(x̄i − x̄)²Each group mean’s squared distance from the grand mean, weighted by group size.
Within-groups SSSSE = Σ (x − x̄i)² = Σ (ni − 1)si²Every score against its own group mean. The right-hand form works straight from group SDs.
Total SSSST = Σ (x − x̄)² = SSG + SSEEvery score against the grand mean; the identity is the arithmetic check.
ANOVA dfdfG = I − 1, dfE = N − I, dfT = N − 1I groups (some books write k), N observations in total. They add up.
Mean squareMS = SS / dfA sum of squares turned into a variance.
F-ratio (ANOVA)F = MSG / MSE, on F(I − 1, N − I)Signal over noise. With two groups, F = t².
Pooled SDsp = √MSEThe one spread every follow-up comparison reuses.
Eta-squaredη² = SSG / SSTShare of outcome variance explained by group membership.
Contrastψ = Σ aiμi, with Σ ai = 0A weighted combination of group means. The zero-sum rule is what makes it a comparison.
Contrast estimate and SEψ̂ = Σ aii,  SE = √( MSE · Σ ai²/ni )Sample means in place of population means; the error term is borrowed from the ANOVA table.
Contrast test and CIt = ψ̂ / SE on df = N − I,  CI: ψ̂ ± t* · SETested on the error df, not on the groups involved. The interval excludes zero exactly when the t clears t*.
Pairwise t after ANOVAt = (x̄i − x̄j) / ( sp√(1/ni + 1/nj) ), df = N − IThe two-sample t with the pooled SD from every group, so more df than the pair alone would give.
Bonferroni correctionαeach = α / mSplit your alpha across m comparisons.
Bonferroni critical valuet** = t1−α/(2m) on df = N − IThe stricter bar a corrected comparison must clear. With 6 comparisons on 12 df it is 3.153 against a plain 2.179.
Simultaneous CI(x̄i − x̄j) ± t** · sp√(1/ni + 1/nj)The whole family of intervals holds at the stated level together, not one at a time.
Two-way dfdfA = I − 1, dfB = J − 1, dfAB = (I − 1)(J − 1), dfE = N − IJ, dfT = N − 1I levels of A, J of B, IJ cells. The five add up, and the design can be read back out of them.
Marginal mean and main effect = mean of row i's cells; main effect of A = difference of the row marginsThe margin averages the cells, not the people, so it matches the raw average only when cells are equal.
Simple main effect (2 × 2)A at Bj = x̄2j − x̄1j; interaction = (A at B2) − (A at B1)Each main effect is the average of its own two simple main effects; the interaction is their difference.
Two-way SS (balanced, n per cell)SSA = nJ Σ (x̄ − x̄)², SSB = nI Σ (x̄·j − x̄)², SSAB = n ΣΣ (x̄ij − x̄ − x̄·j + x̄)²Margins against the grand mean, then the part of each cell its own row and column could not have told you.
Two-way error and totalSSE = Σ (x − x̄ij)², SST = SSA + SSB + SSAB + SSEEvery score against its own cell mean. With equal cells the four pieces add back to the total, which is worth a moment on any table you fill in.
Two-way F ratiosF = MSA/MSE, MSB/MSE, MSAB/MSEThree tests, one error term, each on its own df over N − IJ.
Partial eta-squaredηp² = SSeffect / (SSeffect + SSE)What software prints for a factorial. The denominator changes per row, so the values do not add up.
Chi-squareχ² = Σ (O − E)² / EHow far observed counts stray from what independence expects.
Expected countE = (row total × column total) / NEach cell's count if the two variables were unrelated.
Chi-square dfdf = (r − 1)(c − 1)Rows and columns of the contingency table.
Cramér's VV = √( χ² / (N · (min(r,c) − 1)) )Effect size for χ², from 0 to 1.
Pearson rr = Σ zxzy / (n − 1)The average product of paired z-scores.
Regression slopeb = r · (sy / sx)Predicted change in y per one-unit rise in x.
Intercepta = ȳ − b·x̄The least-squares line always passes through (x̄, ȳ).
R² = 1 − SSres / SStotShare of y's variance the model explains.

Stats 3: Advanced & Elective Topics

ConceptFormulaRead it as
Multiple regressionŷ = b₀ + b₁x₁ + … + bkxkEach bⱼ is that predictor's effect holding the others constant.
Adjusted R²1 − (1 − R²)(n − 1) / (n − k − 1)R² with rent charged per predictor.
VIFVIFj = 1 / (1 − R²j)How much overlap with other predictors inflates bⱼ's variance (worry near 10).
Interaction modelŷ = b₀ + b₁x + b₂m + b₃(x·m)x's slope becomes b₁ + b₃m — it depends on the moderator.
Indirect effecta × bX→M path times M→Y path; test it with a bootstrap CI.
Effect decompositionc = c′ + a·bTotal effect = direct + indirect (through the mediator).
Logistic modelln( p / (1 − p) ) = b₀ + b₁xA straight line on the log-odds scale.
Odds ratioOR = ebMultiplies the odds per one-unit rise in the predictor.
AIC / BIC−2·lnL + 2k  ·  −2·lnL + k·ln(n)Fit plus a complexity penalty; lower wins. BIC punishes harder as n grows.

Stats 3, continued: modern methods

ConceptFormulaRead it as
Bayes' theoremP(θ | data) ∝ P(data | θ) · P(θ)Posterior ∝ likelihood × prior. θ (theta) stands for whatever you are estimating.
Beta–Binomial updateBeta(a, b) → Beta(a + k, b + n − k)See k successes in n trials: add them straight into the prior.
Percentile bootstrap CI[2.5th, 97.5th percentile]…of the statistic across thousands of with-replacement resamples.
GLMg(μ) = b₀ + b₁x₁ + …Link g puts the line on the right scale: logit (binary), log (counts).
Intraclass correlationICC = τ² / (τ² + σ²)Share of total variance sitting between clusters.
Kaplan–MeierŜ(t) = Π (1 − dᵢ / nᵢ)Multiply survival through each event time (censored cases stay in nᵢ while observed). Π is a capital pi and means multiply, the way Σ (sigma) means add.
Cox modelh(t) = h₀(t) · eb₁x₁ + …eb is the hazard ratio for that predictor.
Meta-analysis weightwᵢ = 1 / vᵢ  (random-effects: 1 / (vᵢ + τ²))Precise studies pull the pooled estimate hardest.
share of spread beyond chance≈ 25 / 50 / 75% for low / moderate / high heterogeneity.

Formulas are the receipts, not the idea. If any row feels like symbols instead of a picture, that's the lesson to revisit. Every one of these has an interactive where you can watch the formula happen.