Objectives: The concept of infant emotionality (temperament) is frequently used in the fields of developmental psychopathology, developmental psychology, and child and adolescent psychiatry. However, a valid assessment of the construct has to deal with some difficulties. Parent reports and behaviour observations of the interaction between caregiver and infant may be biased by parental characteristics, while laboratory assessment procedures often have not been sufficiently validated.
Methods: In the present study, three dimensions of temperament were assessed at three ages during the first year of life using three different measurement approaches. Convergent validity, discriminant validity and the associations of the temperament measures with maternal characteristics, i.e. depression, anxiety and educational status were analyzed. The study group consisted of 101 healthy first-born infants and their primary caregivers. At the ages of 4, 8 and 12 months, positive emotionality, negative emotionality and withdrawal/anxiety were assessed by means of a parent questionnaire, naturalistic behavioural observations of the caregiver-infant interaction, and by means of laboratory routines.
Results: Aside from just two exceptions, there were significant convergent correlations between the different measures at each age. Particularly the laboratory routines revealed a good degree of discriminant validity. Questionnaire scores often correlated with the mothers' characteristics. These correlations were independent of the associations with the observational data.
Conclusions: Therewith, parent reports include an objective as well as a subjective component.