From the manual (Sheras et al., 1998)
1. Convergent Validity was derived from correlations between all domains of SIPA and Parenting Alliance Inventory (Abidin, 1988). n= 713; all correlations significant at p<0.01: Adolescent Domain (=-.48), Parent Domain (r=-.59); Adolescent/Parent Relationship Domain (r=-.42), Total (r=-.57). As expected, parents classified as high stress based on PSI cutoffs had significantly lower scores on the PAI than other parents did.
2. In a sample of 64 parents from the normative sample, all SIPA scales were related to aspects of the marital relationship (negative correlations with the Dyadic Adjustment Scale).
3. Validity was also examined in a clinical sample. The Adolescent Domain was significantly correlated with Child Behavior Checklist (CBCL): Internalizing (r=.53, p<.01), Externalizing (r=.59, p<.01), and Total Scores (r=.60, p<.01), whereas the Parent Domain was correlated only with Internalizing (r=.23, p<.05).
The Adolescent/Parent Relationship Domain and Index of Total Stress were both related to Internalizing, Externalizing, and Total Problems.
SIPA scores were also related to scores on the Coping Responses Inventory (with positive correlations for all SIPA scales with Cognitive Avoidance and Seeking Alternative Rewards and negative correlations for all SIPA scales, except Parent Domain, with Positive Reappraisals.
All SIPA scales were negatively correlated with the Cohesion Scale of the Family Adaptation and Cohesion Evaluation Scales III (Adolescent Domain: r=-.38, p<.01; Parent Domain: r=.29, p<.01; Adolescent/Parent Relationship Domain: r=-.52, p<.01; Total: r=-.42, p<.01).
In addition, as expected, only the Parent Domain was related to the Personality Assessment Screener (r=.50, p<.01), a measure of emotional and behavior problems in adults.
4. Discriminant Validity was established by comparing different known groups within the normative sample. All group differences were significant at p<0.001. Parents with a history of mental health treatment had higher scores than parents with no such history. Parents of adolescents with a history of mental health treatment also had higher scores than did parents of adolescents with no such history.
Group comparisons were also significant for parents of adolescents with a history of delinquent behavior and those with no such history and parents of adolescents with a history of suicidal behaviors and those with no such history.
5. Scores for a clinical sample were significantly higher than those for the normative sample. However, scores did not differentiate among clinical groups (Conduct Disorder, Oppositional Defiant Disorder, and Attention Deficit/Hyperactivity Disorder), but this may have been due to the sample size
and to comorbidity in the clinical groups.
6. As noted above under Construct Validity, the measure was developed using factor analysis.
7. Javis, Trevatt, & Drinkwater (2004) conducted a treatment outcome study of 26 participants. SIPA scores showed sensitivity to change related to treatment with pre- to post-test t-tests showing significant declines for Total Stress, Adolescent Domain and its scales, and the Adolescent-Parent Relationship Domain.