Free statistics help
Stuck on a statistics question?
Tell us what you're trying to figure out, in plain words. Someone on our team will read it and reply by email within two business days.
- No jargon required. Describe your data and your goal the way you'd explain it to a colleague.
- Theses, studies, program evaluations, surveys, business data.
- Rather work it out yourself? Try the test chooser below.
Interactive guide
Which statistical test should I use?
Answer a few plain-language questions. You'll get a suggested test, what to check before you trust it, and links to free tools.
A starting point, not a substitute for looking at your data. Design details such as clustering, missing data, and multiple comparisons can change the right answer.
Free resources
Free tools and guides
These live on the DASS site and are free to use.
Tools
Power analysis calculators
Plan sample size for t-tests, ANOVA, correlations, chi-square, and multilevel designs.
Open calculators →APA 7 results formatters
Turn statistical output into clean, copy-ready APA sentences. Runs in your browser.
Open formatters →Check a reported p-value
Recompute the exact p from a t, F, chi-square, or correlation statistic and compare it to what was reported.
Check a p-value →Guides
- Build a Raw-Score-to-T-Score Crosswalk for a Short Form
- Reviewer 2 Criticized Your Shortened Scale? How to Defend It
- How to Shorten a Survey Scale Without Losing Validity
- The Data Audit Checklist to Run First
- FIML vs. Multiple Imputation: Which to Use?
- When a Reviewer Challenges Your Model
- Power for Cluster-Randomized Trials Isn't Simple
- How Much Does a Statistical Consultant Cost?
Answers
Recent questions, answered
Plain-language answers to statistics questions people are asking.
How do you simulate a prediction interval from a fitted regression model?
A simulated prediction interval must combine estimation error and individual scatter. The recipe for one new case and for a group average, in R and Python.
Read the answer →What does it mean to control for a variable in regression?
How adding a covariate to a regression model 'holds it constant', why this replaces sorting people into subgroups, and what controlling can and cannot fix.
Read the answer →REML or ML: which should a likelihood ratio test of fixed effects use?
Restricted likelihoods from models with different fixed effects cannot be compared. Why, what goes wrong, and the ML-then-REML workflow, with R and Python code.
Read the answer →Need more than an answer?
AskStats is run by DASS. If your question turns into a project (study design, data analysis and visualization, or training for your team), book a free 30-minute strategy call.