Comment on: 'An intervention for sensory difficulties in children with autism: a randomized trial' by Schaaf et al. (2013).

Hsieh et al. (2014) wrote a short letter about a sensory-integration RCT by Schaaf’s team.

They praised the study’s design, then listed four ways to make the next trial even tighter.

The authors did not collect new data.

They simply argued that future RCTs should: use blinded observers, give every group the same treatment time, share the protocol online, and compare OT-sensory to another active OT program.

Cook et al. (2020) and Aydin et al. (2022) echo the same worry: observer bias and weak metrics can sink a study. Cook shows how to double-check momentary time-sampling; Aydin offers PCES to tie effect size to preset performance criteria.

Hickey et al. (2021) also waves the rigor flag, but for autism-screening studies. They say you must count harms (like false positives) just as hard as benefits—another flavor of "measure better."

No contradictions here. Each paper targets a different step in the pipeline—K et al. fix RCT design, Cook and Aydin fix single-case measurement, Emily fixes screening policy—but all push the same core idea: tighter methods yield clearer answers.

If you ever help plan or interpret sensory research, keep K’s four checks in mind. Ask: Are outcome videos coded by staff who don’t know the child’s group? Is every arm getting equal minutes? Is the manual free to read? Is the control group also getting real OT, not just “usual care”? Slip any of these and the results may look good but mean little. Use the list as a quick scorecard when you review a new SI study—or when you design one yourself.