Every feature from all five editions for 15 days.
Standard edition from US$ 155 a year · 30-day money-back guarantee.
Excel’s Descriptive Statistics tool produces a table of the mean, standard deviation, skewness and kurtosis and stops. No Q-Q plot, no normality test, no one-sample test against a hypothesised value, no confidence interval on the median. Every analysis starts with the distribution of the data, and what it shows decides which test or model is appropriate — skipping this step is how wrong conclusions happen.
What does the distribution look like? Is it symmetric or skewed? Are there outliers pulling the mean away from the median? Is it normal enough for a parametric test, or should you use a non-parametric alternative? Descriptive statistics, distribution plots, normality assessment and one-sample tests belong in the same workflow, for continuous and discrete variables alike.
Location, spread and shape decide which test or model is appropriate, so characterise them together before you model. Mean, median, SD, CV%, skewness, kurtosis, quantiles, geometric and harmonic mean are reported in one output. See immediately whether the distribution is symmetric, how spread out it is and whether the tails are heavier or lighter than normal.
Summary statistics compress the data; the plots show what the numbers leave out. A histogram shows the shape, a box plot the median, quartiles and outliers, and a dot plot every observation. A CDF plot shows the cumulative distribution with a Kolmogorov-Smirnov confidence band. Each plot answers a different question — use whichever combination the data require.
A parametric test assumes a normal population, so check that assumption before you choose one. The normal Q-Q plot with its Lilliefors confidence band shows visually where the data depart from normality; Shapiro-Wilk, Anderson-Darling and Kolmogorov-Smirnov make it a formal test. If the distribution is not normal, transform the variable within the analysis and reassess — or switch to a non-parametric test.
Does the mean differ from a target value? Student’s one-sample t-test answers when normality holds, the Wilcoxon or sign test when it does not, and a Z test when the population SD is known. Mean and median are estimated with confidence intervals; the Hodges-Lehmann pseudo-median with a Tukey CI gives a robust location estimate. When dispersion is the question, a χ² test and confidence interval for the variance answer it.
A categorical variable needs a different summary: how often each category occurs, and whether the proportions match what was expected. Frequency tables give the frequency, cumulative, relative and cumulative relative frequency; bar plots with a cumulative frequency line and pie plots show them. Binomial exact and score Z tests handle a single proportion, and Pearson χ² and likelihood ratio G² tests compare a multinomial distribution with hypothesised proportions. Each proportion is estimated with a Clopper-Pearson exact or Wilson score CI.
See descriptive statistics and distribution analysis in detail — summary statistics, histograms, Q-Q plots, normality tests and one-sample tests.
4 pages
Continuous distributionDescriptive statistics are one part of a complete statistical analysis toolkit. The Standard edition also includes ANOVA and ANCOVA, simple and multiple regression, logistic regression, PCA and factor analysis, hypothesis testing, correlation and categorical data analysis. See everything in the Standard edition →
Learn what to do when the distribution is not tidy: testing normality and what to do when it fails and the assumptions and costs of non-parametric tests.
Try it on your own data first. The 15-day trial is every feature from all five editions — install it and start straight away.
Standard edition: US$ 155 per year or US$ 395 for a perpetual licence. Every purchase carries a 30-day money-back guarantee. Need a quote for purchasing? Add the licence to the cart and save it as a PDF quote.
