Adult Attention-Deficit/Hyperactivity Disorder (ADHD) is prevalent but often overlooked and undertreated. Left untreated, it is linked to increased risk of untoward outcomes including unemployment, relationship breakups, substance use, driving accidents and other mental health conditions. Several brief screening tools have been developed for adult ADHD. The most frequently used is the World Health Organization's Adult ADHD Self-Report Scale (ASRS V1.1). Here, we show in two independent population samples (UK: N = 642, USA: N = 579) that the tool resulted in considerable overestimation of ADHD, indicating probable ADHD in 26.0% and 17.3% of participants, as compared to expected prevalence of 2.5%. The estimated positive predictive value was only ~11.5%. Both samples had normal levels of trait impulsivity as assessed using the Barratt Impulsiveness Scale. The data indicate that using the ASRS in general population samples will result in 7–10 times over-identification of ADHD. We use these results to highlight how such tools should most appropriately be used. When being used to determine possible cases (such as for onward referral to an ADHD specialist) they should be complemented by clinical assessment – we give examples of how non-specialists might determine this. When measuring ADHD symptoms dimensionally, researchers should be mindful that the ASRS captures impulsive symptoms other than those due to ADHD. Lastly, we note the need to screen for impulse control disorders (e.g., gambling disorder) when using such tools to measure ADHD, be it for onward referral, or for dimensional research studies.