Preschool children's recall of both experienced and non-experienced activities was examined across three interviews. One hundred and six children aged 4 to 5 years (58 males, 48 females) from both low and high socioeconomic status areas participated in an event called the Deakin Activities, which consisted of two experienced activities. One or two days later, the children were asked to recall what happened in the two activities and an activity they had not experienced which was suggested to have occurred along with the experienced activities. Next, children were given false suggestions about one of the experienced (true-biased) activities and the non-experienced (false) activity. For the remaining experienced (true-unbiased) activity, no questions were asked. Three and eight days after the activities were presented, children were again required to recall all three activities in their own words while a variety of suggestive techniques were used as encouragers. The results revealed that irrespective of the SES group, assent rates across the true and false activities became more similar after the first interview. Furthermore, children's narratives about the false activity became more similar in detail, structure and quality to their narratives about the true activities across interviews. However, the rate of fantastic/improbable details was higher for the false activity compared to the true activities, children reported more interviewer suggestions about the false activity than the true-biased activity, and there were fewer confabulation errors reported about the false activity compared to the true activities. The implications of the results are discussed. Copyright © 2003 John Wiley & Sons, Ltd.