StatGardenREF. DESK

Calculators/Maths/Statistics

Statistics calculators

Averages, spread, distributions, confidence and correlation for data sets you paste straight in.

Descriptive statisticsMean, median, mode, spread and quartiles for a data set.Weighted averageAverage where each value carries its own weight.Z-score & percentileHow far a value sits from the mean, in standard deviations.Normal distribution probabilityArea under the bell curve between two points.Confidence interval for a meanMargin of error around a sample average.Required sample sizeHow many responses a survey needs.Correlation & linear regressionPearson r and the line of best fit for paired data.Binomial probabilityChance of exactly k successes in n independent trials.Probability of eventsCombines two event probabilities: and, or, neither.Odds ↔ probabilityConverts between fractional odds, decimal odds and percent.GPA / weighted gradeGrade point average across units with credit weights.Grade needed on the finalWhat the last assessment must score to hit a target.Coefficient of variationRelative variability of a data set: standard deviation as a percentage of the mean.SkewnessSample skewness: whether the data leans left or right of the mean.Poisson probabilityChance of a given number of independent events in a fixed interval, given the average rate.Geometric probabilityChance the first success happens on a given trial, for repeated independent trials.Hypergeometric probabilityChance of drawing an exact number of successes from a finite population, without replacement.Chi-square goodness of fitTests whether observed category counts match expected counts.Cohen's d (effect size)Standardised difference between two group means, in pooled standard deviations.Margin of error (proportion)Margin of error for a survey proportion at a chosen confidence level.Relative riskCompares event risk between an exposed and an unexposed (or treated and control) group.T-statistic (one or two sample)How many standard errors a sample mean is from the reference value, or from another mean.Z-test for a proportionTests a sample proportion against a hypothesised population proportion.Expected value of a discrete distributionMean and standard deviation of a discrete random variable from its outcomes and probabilities.Outlier fences (Tukey)The IQR-based cutoffs beyond which a data point counts as an outlier.CovarianceHow two variables move together, in sample or population form.Standard error of the meanHow much a sample mean is expected to vary from the true population mean.F-test for equal variancesCompares the spread of two samples using the ratio of their variances.Negative binomial probabilityChance that the rth success happens on exactly the kth independent trial.Percentile rankWhat percentage of a data set falls at or below a given value.Value at a percentileFinds the data value at any percentile using linear interpolation.Mean absolute deviationThe average distance of each data point from the mean (or median).Bayes’ theoremUpdates the probability of an event given new evidence.Class width for a histogramHow wide each bin should be, from either a target class count or Sturges’ rule.Spearman’s rank correlationMeasures a monotonic relationship between two variables using their ranks, ties included.Chebyshev’s theoremMinimum share of any data set within k standard deviations of the mean, whatever its shape.Two-proportion z-testTests whether two independent samples have significantly different proportions.Pooled standard deviationCombines two sample standard deviations into one pooled estimate for equal-variance tests.Z-test for a single meanTests a sample mean against a hypothesised mean when the population standard deviation is known Quartile deviationHalf the interquartile range: a robust spread measure, less sensitive to outliers than the SD.Range rule of thumb for standard deviationA quick estimate of standard deviation from just the minimum and maximum of a data set.Mean & standard deviation from grouped dataEstimates the mean and standard deviation of a frequency table using class midpoints.Empirical rule (68-95-99.7)The share of normal data expected within 1, 2 and 3 standard deviations of the mean.KurtosisSample excess kurtosis: how heavy-tailed a data set is versus a normal distribution.Diagnostic test accuracy metricsSensitivity, specificity, predictive values and accuracy from a 2×2 test result table.Birthday paradox probabilityChance that at least two people in a group share the same birthday.Bonferroni correctionAdjusts the significance threshold when running many statistical tests at once.Benford's law leading-digit checkCompares a data set's leading digits against the distribution natural data follows.One-way ANOVA (F-statistic)Tests whether three or more group means differ, from the raw data in each group.McNemar's testTests whether paired binary outcomes changed, using the two discordant-pair counts.Mann-Whitney U test (Wilcoxon rank-sum)Compares two independent samples for a shift in distribution, using ranks instead of means.MidrangeThe midpoint between the smallest and largest value in a data set, a quick measure of centre.Sum of squaresTotal squared deviation of each data point from the mean, the building block of variance.Confusion matrixTurns the four counts of a binary confusion matrix into accuracy, precision, recall and F1.Youden's indexCombines a test’s sensitivity and specificity into one score of diagnostic performance.Weibull distributionComputes the PDF, CDF, mean and variance of a Weibull distribution from its shape and scale.Rayleigh distributionComputes the PDF, CDF, mean and variance of a Rayleigh distribution from its scale parameter.Fisher's exact testRuns a two-tailed Fisher exact test on a 2×2 contingency table of counts.Process capability indexCalculates Cp and Cpk from specification limits, the process mean and standard deviation.Degrees of freedomWorks out degrees of freedom for a one-sample, two-sample or chi-square test.Index of qualitative variationMeasures diversity among categorical frequencies on a 0 to 1 scale using the IQV formula.Beta distributionPDF, mean and variance of a beta distribution from its two shape parameters.Monty Hall problemWin probability from staying versus switching after the host reveals losing doors.Statistical power analysisAchieved power for a given sample size, and the sample size needed for a target power.Error propagationPropagates the uncertainty of two measurements through a sum, difference, product or quotient.

Other Maths categories