Background Childhood amnesia in adults can be defined as the relative paucity of autobiographical memories from the first years of life. An earlier study by Wessel, Schweig and Huntjens demonstrated that 'how' we ask for an earliest memory may bias adults' estimations of when the earliest childhood memory actually happened. They suggested that snapshot memories (i.e., mental pictures) were less sensitive to an age manipulation than event memories (i.e. narratives). We aimed at replicating and extending these findings using a Dutch community sample stratified for age, gender and educational level. Method Participants (N = 619) were randomized into one of three experimental conditions. Prior to recalling their earliest memory, participants in the early and late conditions were presented with examples referring to memories from age 1-2 or 5-6, respectively. The example memories in the control group did not contain any age cues. Participants reported the estimated age in their earliest memory and their strategy for arriving at this estimate. They also rated their memory's phenomenology (e.g. vividness). Independent judges rated memory type (e.g., snapshot memories). Results Compared to the control group, participants in the early condition estimated the age in their memory to be significantly earlier. The difference between the late and control conditions was too small to be of interest. We did not observe a statistically significant interaction between memory type and condition. Snapshot memories were from a younger age than event memories and showed differences with respect to phenomenology (e.g., emotional intensity). Conclusion The results of this community study replicate earlier findings that instructions including age cues influence estimates of age in earliest memories. Although snapshot and event memories seem to be qualitatively different, the idea that they respond differently to an age manipulation could not be corroborated.