BACKGROUND: In the numerical Stroop paradigm (NSP) participants decide whether a digit is numerically or physically larger than another simultaneously presented digit. This paradigm is frequently used to assess the automatic number processing abilities of children. Currently it is unclear whether an equally refined evaluation of numerical magnitude occurs in both controlled (the numerical comparison task of the NSP) and automatic (the physical comparison task of the NSP) numerical comparison in both children and adults. One of our objectives was to respond this question by measuring the speed of controlled and automatic magnitude processing in children and adults in the NSP. Another objective was to determine how the immature executive functions of children affect their cognitive functions relative to adults in numerical comparison. METHODS AND RESULTS: The speed of numerical comparison was determined by monitoring the electro-encephalographic (EEG) numerical distance effect: The amplitude of EEG measures is modulated as a function of numerical distance between the to-be-compared digits. EEG numerical distance effects occurred between 140-320 ms after stimulus presentation in both controlled and automatic numerical comparison in all age groups. Executive functions were assessed by analyzing facilitation and interference effects on the latency of the P3b event-related potential component and the lateralized readiness potential (LRP). Interference effects were more related to response than to stimulus processing in children as compared with adults. The LRP revealed that the difficulty to inhibit irrelevant response tendencies was a major factor behind interference in the numerical task in children. CONCLUSION: The timing of the EEG distance effect suggests that a refined evaluation of numerical magnitude happened at a similar speed in each age group during both controlled and automatic magnitude processing. The larger response interference in children than in adults suggests that despite the similar behavioural profile of children and adults, partially different cognitive processes underlie their performance in the NSP. Further, behavioural effects in the NSP depend on interactions between comparison, facilitation/interference and response-related processes. Our data suggest that caution is needed when using the NSP to compare behavioural markers of the numerical processing skills of children and adults.