ZStat
Sign in Start free

Statistics and machine learning for research.

Design the questionnaire, collect the responses lawfully, run the analysis, and export an APA 7 results chapter in Word.

t tests · ANOVA · regression · factor analysis · SEM · survival · panel data · meta-analysis · classification · forecasting · clustering

Start free See a whole study

Free for your first study, with no time limit and no card.

zstat — analysis

independent_t . outcome: score . group: method

result

Students taught by the new method scored higher (M = 68.4, SD = 9.1) than those taught conventionally (M = 61.2, SD = 10.4), t(148) = 4.51, p < .001, d = 0.74, 95% CI [4.05, 10.35].

Levene p = .31 . Shapiro-Wilk p = .22 . n = 150

independent_t . outcome: score . group: arm

wrong test

A t test compares two groups. "arm" has three: Control, Treatment A, Treatment B.

Use a one-way ANOVA.

panel_regression . outcome: roa . entity: firm

result

The fixed effects specification is reported: Hausman chi-square(3) = 21.7, p < .001. Leverage predicted return on assets, B = -0.18, SE = 0.04, p < .001.

Durbin-Watson 1.94 . 40 firms . 8 years

open study for collection

not lawful yet

No ethics reference, no retention date, no withdrawal route. Three of twelve NDPA conditions unmet.

The first respondent is already too late.

reliability . 4 items

result

The digital literacy scale was reliable, Cronbach alpha = .84, 95% CI [.79, .88], four items, n = 150.

No item raises alpha if removed

train classifier . 86 features

leakage

Feature "final_grade" correlates with the outcome at r = 1.00.

It would score perfectly here and fail completely in use.

What it writes for you — and what it refuses to write.

86statistical procedures
13machine learning tasks
12NDPA checks before collection
₦0for a student, always

Who uses it

Written for the question you are actually asking.

Each field has its own guide: the questions in the words you would use, the procedure that answers each, and the mistakes characteristic of that discipline.

Medicine and clinical research Clinical trials, diagnostic studies, case-control and cohort work, hospital audit. Public health and epidemiology Prevalence surveys, outbreak work, health systems research, programme evaluation. Psychology and behavioural science Scale development, experiments, questionnaire studies, intervention trials. Education Achievement studies, teaching interventions, curriculum evaluation, assessment research. Agriculture and food science Field trials, varietal comparison, storage and processing studies, extension research. Urban planning, estate management and architecture Land use, valuation, housing studies, post-occupancy evaluation, accessibility and transport. Economics, finance and accounting Firm performance, market studies, policy evaluation, organisational research. Engineering, manufacturing and quality Process control, quality assurance, reliability, operations research. Sociology, development and social work Community studies, programme evaluation, poverty and livelihood research, policy work.

How a study moves

Six stages. You cannot skip one.

Analysis needs collection. Collection needs a locked questionnaire. ZStat holds the order so you do not have to.

Design the questionnaire

Items, scales and choice lists, checked for leading wording, missing universes and response formats that will not analyse.

Clear the compliance gate

Ethics reference, lawful basis, a specific purpose, a retention date and a working withdrawal route. Twelve NDPA 2023 conditions, checked before the first respondent.

Collect

A link that works on a phone. Consent recorded against the wording it was given for. Every answer written with an authorisation the database itself checks.

Analyse

86 procedures, each running its assumption checks whether or not you asked, and each result stored with the digest of the data behind it.

Write it up

An APA 7 results chapter as a Word document, with the Method section assembled from what the study already recorded.

Answer for it

A supervisor link that needs no account. An append-only log of every analysis. Withdrawals honoured, retention enforced on a schedule.

What you get

Six things a statistics package will not do for you.

Not a feature list. These are the moments where a study goes wrong quietly, and what ZStat does at each one.

You pick the wrong test
It names the groups it actually found and points you at the right procedure, instead of running something that returns a number anyway.
Your questionnaire has a problem
Leading wording, missing universes and response formats that will not analyse are caught before the instrument is locked.
You start collecting too early
Ethics reference, lawful basis, purpose, retention date and a working withdrawal route: twelve NDPA 2023 conditions, checked before the first respondent.
Someone asks where a figure came from
Every result is stored with the SHA-256 of the data behind it, so "is this the same result as last month" has an answer.
You have to write it up
APA 7 tables and sentences, the Method section assembled from what the study already recorded, exported to Word.
Your supervisor wants to see it
A read-only link they open without an account, showing the results and the assumption checks behind them.

Plans

A plan limits how much, never how correctly.

Every check and refusal is on the free plan. Charging for the part that stops you publishing something wrong would mean the people who cannot pay get the tool that fails silently.

Free

₦0

No time limit

  • All 86 procedures
  • Every assumption check
  • One study, 150 responses
  • Collection and withdrawal
Start free

Researcher

Per year

One person

  • Unlimited studies
  • 20,000 responses each
  • Exports with no footer
  • Supervisors and the desktop app
Compare

Institution

Per year

A department, or the university

  • Your own subdomain
  • Your data in its own database
  • Institution-wide oversight
  • Training and support
What is included

Self-hosted

Licensed

On your own server

  • Runs inside your network
  • Respondent data never reaches us
  • Your backup and retention
  • Annual licence
ztech.ng

Run your first analysis today.

No card, no trial period. Every procedure, every assumption check and the full questionnaire designer are on the free plan.

Create an account Sign in