Teaching · For Teachers

Twelve classrooms are not 245 independent students.

Enter the cluster size, response rate and ICC. The organizer works out the design effect, what your sample is really worth and how many classrooms or schools you need.

Plan a sampleWhat you get
$12one-time · every device you sign in on
Buy · $12

Included with Caveat Classroom and Whole-school licenses. Or get it in the For Teachers toolkit, $79 for 72 items.

The cluster design

One design per study and date. The example is made up.

Loading the form. If nothing appears, turn on JavaScript.

What you get

Teaching · For Teachers$12 · included with Classroom and Whole-school licenses

What you finish with
Design effect, total and effective sample size, clusters needed, and cautions for the plan.
Who it is for
Teacher-researchers, district evaluators, program evaluators and graduate students.
Format
Browser form. One saved design per study and date; printed with the working or as a blank form; CSV; export and import a file.
Access
$12 one-time, yours on every device you sign in on; the tool opens once you are signed in with it. Included with Caveat Classroom and Whole-school licenses. Also in the For Teachers toolkit ($79) and the Teaching Suite pass ($199), and 1 more bundle. Everything you type stays in this browser; nothing is sent to Caveat.
Where it lives
The For Teachers space in Caveat Teaching.
Support
Questions about an order, access or a correction: team@caveatacademic.com.
alike inside, different between
Drawn for this page. Students in one classroom share a teacher and a schedule, so each extra student tells you less than a stranger would.

How to use it

  1. Choose the cluster unit and count what is available.
  2. Enter the average cluster size and expected response rate.
  3. Find an ICC for your outcome, or try a range.
  4. Enter the effective sample size from your power or margin-of-error calculation.
  5. Adjust clusters until the plan meets the target.

Plan a sample

The working

The design effect

Because people in one cluster resemble each other, a cluster sample carries less information than a simple random sample of the same size. The design effect, 1 + (m − 1) × ICC, is how many times larger the cluster sample must be.

Where ICCs come from

Published tables give ICCs for achievement by grade and subject; between-school values for achievement are often around .10 to .25. Your own earlier data are better still.

What this is not

A power analysis. Use one to set the effective sample size first.

Sources

  1. Kish, L. (1965). Survey Sampling. Wiley: the design effect.
  2. Hedges, L. V., & Hedberg, E. C. (2007). Intraclass correlation values for planning group-randomized trials in education. Educational Evaluation and Policy Analysis, 29(1), 60–87.
  3. Lohr, S. L. (2021). Sampling: Design and Analysis (3rd ed.). CRC Press: cluster sampling.
  4. Institute of Education Sciences, What Works Clearinghouse (2022). Procedures and Standards Handbook, Version 5.0: clustering corrections.

The example study and figures are made up. Support: team@caveatacademic.com.