Drs Lu and Shen claim that our Trauma Screening Questionnaire (TSQ; Reference Brewin, Rose and AndrewsBrewin et al, 2002) is flawed because it omits avoidance and numbing symptoms and asks about symptom frequency using a simple ‘yes/no’ response format. It is puzzling then that the performance of the TSQ is superior to that of all comparable screening measures, including ones that follow Lu and Shen's recommendations. Their views are clearly contradicted by the data from the two studies we reported. Our reasons for designing the TSQ in the way we did were based on empirical and practical rather than theoretical considerations. In our original article we discussed some general principles for designing successful screening instruments, whereas Lu and Shen's comments seem more relevant to a diagnostic instrument. The two types of measure tend to be administered by different professionals, under different circumstances, and with different aims in mind. It seems to us that, as a screening instrument, what the TSQ gains in simplicity and clarity more than compensates for the absence of symptoms that may be difficult to understand and judgements that may be difficult to make.