Steply

Statistics formula sheet

31 formulas · what each one means · a free lesson for each

All formula sheets

steplyclass.com/formula-sheets/statistics/ · free, step-by-step lessons for every formula

Center and spread

Mean
x = Σxn
μ for a whole population.Lesson: Mean, median, mode, range →
Mean of grouped data
x ≈ ΣfxΣf
x = the middle of each class, f = its frequency.Lesson: Mean of grouped data →
Interquartile range
IQR = Q3 − Q1
The spread of the middle half.Lesson: IQR and outliers →
Outliers
below Q1 − 1.5 × IQR or above Q3 + 1.5 × IQR
Lesson: IQR and outliers →
Population standard deviation
σ = √Σ(x − μ)2N
Lesson: Standard deviation →
Sample standard deviation
s = √Σ(x − x)2n − 1
Divide by one less than how many.Lesson: Standard deviation →
Frequency density
frequencyclass width
The height of a bar when the classes are not the same width.Lesson: Histograms with unequal class widths →

Lines of best fit

Line of best fit
ŷ = mx + b
It goes through the mean point (x̄, ȳ). ŷ (y-hat) is the predicted y.Lesson: Line of best fit →
Residual
residual = y − ŷ
Actual minus predicted.Lesson: Residuals →
Correlation coefficient
r = Σ(x − x)(y − y)√Σ(x − x)2 × Σ(y − y)2
From −1 to 1: the sign gives the direction, near ±1 is strong.Lesson: The correlation coefficient →

The normal curve

z-score
z = x − μσ
How many standard deviations above (+) or below (−) the mean.Lesson: z-scores →
68-95-99.7 rule
μ ± σ: 68% μ ± 2σ: 95% μ ± 3σ: 99.7%
Lesson: The normal curve →
Margin of error (95%)
x ± margin
The true value is likely inside this interval.Lesson: Margin of error →

Random variables

Mean (expected value)
μ = Σ x P(x)
Lesson: Random variables →
Standard deviation
σ = √Σ(x − μ)2 P(x)
Lesson: Random variables →
Poisson
P(x) = μx e−μx!
μ = mean number of events in the interval.Lesson: The Poisson distribution →
Geometric
P(x) = (1 − p)x−1 p
The first success on try x.Lesson: The geometric distribution →
Sample means (CLT)
σx = σ√n
Means of big samples are normal, with a smaller spread.Lesson: Sample means and the central limit theorem →
z for a sample mean
z = x − μσ/√n
Lesson: Sample means and the central limit theorem →

Confidence intervals

Mean, σ known
x ± z* × σ√n
z* = 1.645 (90%), 1.96 (95%), 2.576 (99%). The ± part is the error bound EBM.Lesson: Confidence interval for a mean →
Mean, σ not known
x ± t* × s√n, df = n − 1
t* from the t-table.Lesson: Confidence interval with t →
Proportion
p′ ± z* × √p′q′n, p′ = xn, q′ = 1 − p′
Lesson: Confidence interval for a proportion →
Sample size
n = z*2 σ2EBM2
Always round up.Lesson: How big a sample? →

Hypothesis tests

Test for a mean (σ known)
z = x − μ0σ/√n
Lesson: Hypothesis test for a mean →
Test for a mean (t)
t = x − μ0s/√n, df = n − 1
Lesson: Hypothesis test with t →
Test for a proportion
z = p′ − p0√p0 q0 / n
Lesson: Hypothesis test for a proportion →
Decide
p-value < α → reject H0
Lesson: Hypothesis test for a mean →
Two proportions (pooled)
pc = x1 + x2n1 + n2
Lesson: Comparing two groups →
Paired samples
t = dsd/√n
d = the differences, after − before.Lesson: Paired samples →
Chi-square
χ2 = Σ(O − E)2E
O = observed count, E = expected count.Lesson: Chi-square →
Expected count (two-way table)
E = row total × column totalgrand total
Lesson: Chi-square →

More formula sheets