A five-pass checklist for reading a randomised trial

Reading a randomised trial becomes easier when you give each pass a different job. On the first pass, you are finding the question. On the next, you are checking how the comparison was organised. Only then do you decide what the results can support. This order keeps a memorable abstract sentence from becoming the answer before you have examined the evidence behind it.

The worksheet below is designed for a real paper, not for memorising research terminology. Open the article alongside its registry record and any available protocol or statistical analysis plan. Make a short note at each step. When the information is unavailable, write not reported or not located rather than filling the gap with a favourable assumption. Missing reporting is different from a demonstrated design problem, but both affect what a reader can verify.

You do not need to calculate a new statistical analysis to use this method. The main task is to keep the study's question, participant flow, result and conclusion aligned. At the end, try writing a two-sentence appraisal: one sentence stating the supported finding and another identifying the most consequential limitation. That exercise is more revealing than collecting a list of technical words, and it gives you a record you can revisit when a later paper or headline claims to overturn the result. Keep your completed note beside the original paper for future comparison.