Mixed Findings - In contrast to the support for the internal working model, findings from Zimmerman et al's (2000) longitudinal studies yielded mixed results. They found that childhood attachment type was not a good predictor of attachments in adolescence, and that life events often altered secure attachments to an insecure type in adulthood. Those who had changed attachment type had generally experienced a life event such as the loss of a parent, parental divorce or a life-threatening illness of a parent or child. Waters et al (2000) found that 44% of the infants who had experienced negative life events changed attachment classification from infancy to early adulthood. This suggests that although there is strong support for continuity of attachment, there is also evidence that positive or negative environmental changes can improve or compromise attachment type in later life.