6. Randomization Methods: Simple, Blocked, Stratified, Concealed
After this lesson you can
- State the three things randomization buys and the one it does not guarantee
- Show how simple randomization can drift out of balance
- List the six sequences of a block of four and explain block predictability
- Stratify a schedule by a prognostic factor without emptying the blocks
- Judge allocation concealment and explain central allocation
Video coming soon
Lessons publish daily from November 17, 2026.
This lesson explains what randomization buys and does not guarantee, then builds simple, permuted-block and stratified schedules with the fictional Oakhaven numbers. It defines allocation concealment, shows how it fails, and separates it from blinding.
Questions and answers
Questions a learner would ask about this lesson. Think through your own answer before you open each one.
- Why isn't alternation (one to the program, one to usual care) randomization?
Show answer
Alternation is predictable, so anyone who knows the last assignment knows the next one, and the enroller can steer participants. It keeps the counts equal but not the arms comparable. Randomization assigns each participant by a chance process that no one can predict or influence.
- What does randomization buy, and what does it not guarantee?
Show answer
It buys three things: balance of known and unknown factors in expectation, a basis for the probability statements, and no selection into arms. It does not guarantee balance in the particular trial in front of you, and it does not blind anyone to the arm.
- How do permuted blocks work?
Show answer
The sequence is built from blocks in which each arm appears equally often in random order. A block of four with two arms has six possible sequences (four choose two): AABB, ABAB, ABBA, BAAB, BABA and BBAA. This keeps the arms within two of each other and lands at 125 and 125, at the price that the last allocation in a block can be predicted.
- What is stratified randomization, and how did Oakhaven use it?
Show answer
Stratification runs a separate schedule within each level of a strong prognostic factor chosen before enrollment. Oakhaven stratifies by age band: 150 participants under 50 split 75 and 75, and 100 aged 50 and over split 50 and 50. Too many strata leave many blocks unfinished, which gives no balance.
- How is allocation concealment different from blinding?
Show answer
Concealment acts before enrollment and keeps the enroller from knowing the next assignment; it is always possible. Blinding acts after enrollment and hides the arm; it is sometimes impossible. Oakhaven uses central allocation, where the clinic enrolls first and then requests one logged assignment from a central system.
Practice quiz
- Question 1How many distinct sequences does a permuted block of four with two arms have?
- Question 2In a fixed block of four, the first three assignments were A, B, A. What must the fourth be?
- Question 3Simple randomization of 250 participants could produce:
- Question 4In the Oakhaven stratification, the under-50 stratum randomizes:
- Question 5Which of these is NOT one of the three things randomization buys?
- Question 6Allocation concealment protects against:
- Question 7A good stratification factor is:
- Question 8The safest concealment method among these is:
- Question 9An enroller can see that the next slot is the program and steers sicker participants away from it. What happens to the measured effect?
- Question 10Why are fixed blocks of four safe in the Oakhaven trial?
Saved in your browser only: no account, no server.
Design of Clinical Studies is education only. Every lesson follows the fictional Oakhaven Adherence Trial with invented numbers; the course uses no real data and cites no real studies.