This study was designed to investigate parameters of children's fear in terms of frequency of fearful thoughts and avoidance behaviour. It is suggested that current measures such as the Fear Survey Schedule for Children-Revised (FSSC-R) do not assess fearful behaviour in the sense of the occurrence of fearful responding in daily life, but rather reflect a negative affective response to the thought of occurrence of specific events. A modified version of the FSSC-R examined the frequency of fearful thoughts/feelings and avoidance activities amongst 376 children aged 7-12 yr. Contrary to predictions, it was found that children reported high levels of fearful thoughts and avoidance behaviour to those items identified as the greatest fears on the FSSC-R, namely fears of injury, illness, death and danger. These events were typically of low probability (e.g. earthquakes) and the question was raised as to what children are responding to when they are asked to rate their fearful responses. The same pattern of results was reflected in older compared to younger children. It is suggested that even when children are asked to rate frequency of fearful thoughts or avoidance behaviour, they tend to respond to fear questionnaire items according to their affective response to the image or thought of the stimulus situation rather than their actual fear responses. Both the FSSC-R and the modified version were found to discriminate between teacher nominated high and low fearful children and to correlate significantly with a self report measure of anxiety.