Reading focus
Syllabus: How Can We Study Change Within People?
Identify the planned assignment mechanism, permitted phase changes and ethical constraints behind each design. Keep hypothetical randomization separate from what the original studies did.
Why you’re reading it. Connect a single-case analysis to the comparison the design actually supplies.
A question to consider. For an intervention with lasting effects, how could staggered onset strengthen inference without requiring withdrawal, and which timing decisions would need to be made in advance?
Why this reading
Repeated observation of an individual can be organized to test an intervention. Tanious and Onghena compare single-case designs and connect their planned randomization to the comparisons that an analysis can justify.
Before you read
A baseline records the outcome before an intervention condition. Reversal, alternation, multiple-baseline and changing-criterion designs organize comparisons differently. A randomization test evaluates results against the assignments permitted by the actual design.
Read for
- Identify what is randomized and what supplies experimental control in each design.
- Distinguish designs actually randomized from illustrations whose randomization is hypothetical. Consider visual patterns, magnitude and uncertainty together.
Consider afterwards
- An intervention cannot ethically be withdrawn and may have lasting effects. Which design could strengthen an individual-level comparison, and what would need to be decided before observations begin?
Editorial context
The authors disclose that some worked randomization analyses assume procedures absent from the original studies. Those calculations illustrate methods rather than establish valid randomized evidence retrospectively. Single-case designs face their own limits, including carryover, few permissible assignments and uncertain generality. They complement other designs; a small p value does not establish clinical importance or a mechanism.