Construct validity of the Child-Initiated Pretend Play Assessment-For 3-year-old Brazilian children.

Introduction: Play is an indication of children's development and how they function. In occupational therapy it is regarded as an important occupation of childhood. Assessment of a child's play should be included in the test batteries of occupational therapists, who understand the construct validity of the assessment they have chosen. Our aim was to provide evidence of reliability, internal consistency and hypotheses testing of the construct validity of the cross-culturally adapted version of the Child-Initiated Pretend Play Assessment.

Methods: Two hundred typically developing Brazilian children aged 3 years were evaluated individually using the Child-Initiated Pretend Play Assessment.

Results: The internal consistency showed a Cronbach's alpha coefficient to Percentage of Pretend Play Actions (PEPA) of 0.86 and NOS of 0.81. There were significant differences between the younger children (aged 36-41 months) and the older children (aged 42-47 months) for elaborate play across both conventional-imaginative play (p = .002), symbolic play (p = .012), and the combined score (p = .005). There were significant differences between younger and older girls for elaborate play with symbolic play materials (p = .009) and elaborate play overall (p = .039). There were significant differences between young and older boys for elaborate play with conventional toys (p = .006) and elaborate play overall (p = .025). There were no significant differences for object substitution or imitated actions.

Conclusion: The measurement properties of the cross-culturally adapted version of the Child-Initiated Pretend Play Assessment for 3-year-old Brazilian children identified evidence for response processes, internal structure, with discussion of consequences of testing for 3-year-old Brazilian children.