Multiple perspectives on family relationships: a latent variables model

Area of Science:

• Psychology
• Family Studies
• Quantitative Research Methods

Background:

• Scholarly skepticism exists regarding the reliability of family member reports on interpersonal dynamics.
• Concerns include potential biases such as social desirability affecting familial reports.
• Accurate assessment of parent-child relationships is crucial in clinical and research settings.

Purpose of the Study:

• To evaluate the reliability and bias of mother, father, and child ratings of parent-child negativity.
• To determine if familial reports contain valid "true score" components.
• To assess the utility of latent variables modeling for analyzing such data.

Main Methods:

• Employed a latent variables modeling approach.
• Analyzed data from 78 clinical families.
• Compared family member ratings with observer ratings of parental behavior.

Main Results:

• Family member ratings of parent-child negativity contain a significant "true score" component.
• This "true score" component correlates with observer ratings of parental behavior.
• Systematic rater effects were demonstrated, highlighting the need for statistical control.

Conclusions:

• Familial reports on parent-child negativity offer reliable insights beyond mere bias.
• Latent variables modeling effectively controls for rater effects in family data.
• This methodology is recommended for analyzing interpersonal relationship data within families.