Welcome
What are you trying to do?
A study for my organisation in agencies, ngos and programme evaluation.
Your path
Create the study and set the discipline
The discipline decides what every later page shows you.
Decide whether people are involved
If your data is about machines, facilities or finances rather than people, declare no human subjects and the data protection conditions do not apply. The declaration is checked.
Collect or import
Collect through ZStat with consent and withdrawal, or bring a CSV, Excel or SPSS file you already have.
Analyse and compare
Between sites, periods, groups or programmes - with the effect size, because significance alone does not tell a manager whether it matters.
Produce the assurance pack
Generated from your own installation, for a funder or a data protection officer.
What people in your field ask
Define the indicator - sample defensibly - collect with consent - compare against baseline - report what the design can and cannot support.
- Did the programme change anything?independent t
- We could not randomise. Can we still claim an effect?propensity matching
- How many people are affected, and how sure are we?survey proportion
- Our survey used clusters. Does that change the numbers?survey mean
- Which communities are worst affected?getis ord
Watch for: Reporting a before-and-after difference as programme impact when anything else changed in the same period.
Create your first study The full guide for agencies, ngos and programme evaluation Start again