Altamont AdvisoryImplementation Integrity
Work email required.
Note from zs@altamontgroup.ca

Get now · Free

Sampling-Size Calculator

Free $0 · Get now

What it is. How many people to interview so the result can stand in a ToR.

What you do. Four questions. Calculate. See n=. Work email gets the note.

What you leave with. The n= for the group and precision you typed, plus the working for an inception pack.

Why this is not a webpage number. The reviewer asks why. This mails the working.

Calculate
Sampling note showing a recommended sample size a team leader can attach to an inception pack.
The branded note zs@altamontgroup.ca sends. A sampling note a team leader can attach.

What you walk away with

Recommended n= on this page after Calculate. Work email gets the branded PDF and editable Word from zs@altamontgroup.ca — the note a team leader can attach.

What it is not

Not a completed sample. Not a country pick-list. Not an evaluation. Not a live assignment. Not anyone else’s data. The page does not fill a sample for you.

Simple shows n= after Calculate. Work email unlocks the branded note from zs@altamontgroup.ca. No .txt. No optional email.

Simple

How large a sample?

Answer four questions. See the number on this page after Calculate.

Leave blank if the group is very large.
Percentage points. Default 5%.
How sure? Default 95%.
Default 50% — use this if you don’t know.
Non-response allowance (optional) Off by default. Adds 10% or 15% to the interview target.

Example only — 95% confidence, ±5 points, expected share 50%: starting n is 385 if the group is very large, about 370 if it is about 10,000. Your n= will differ if you change any of those.

Design Advanced design More options for means, two groups, clusters, and multiple sites — open only if you need them.

Step 1

What the n= is for

Required for Design. An n= without a stated use is a number, not a sample.

Optional. Flavours the note only. Who decides what, by when.

Step 2

Inputs

Empty where a statistician would leave them empty. Defaults only for confidence, p = 0.5, deff = 1, non-response = 0, and one site.

Confidence
Optional. Blank or 0 = large / infinite. No FPC.
Design effect
Minimum 1. Default 1.0 (simple random).
0–0.6 typical. Default 0. Invalid at 0.95 or above.
1–12. Labels are Site 1, Site 2, … never a country list.
Default 0. If > 0, floors apply and the total may exceed the formula n.

Share

Default 0.5 — most conservative (maximum variance).
Absolute, on the proportion scale. Example 0.05. Empty until you type it.
How the number is made
Enter the inputs for this mode. The formula will appear here as you type.

Yamane (a check, not the headline): n = N / (1 + N e²). For N = 10 000 and e = 0.05 that is 385.

Branded note

Accept and request the branded note

The recommended n= above stays on this page. Work email unlocks the branded PDF and editable Word from zs@altamontgroup.ca — assumption table, build chain, can/cannot, ToR lines.

Free. Simple shows n= after Calculate. Work email unlocks the branded note from zs@altamontgroup.ca. PDF + Word from zs@altamontgroup.ca.