Positive Affect (PA) and Negative Affect (NA) are important states related to one’s emotional experiences and engagement in life (Watson et al., 1988). Both child and parent versions of the Positive and Negative Affect Schedule for Children (PANAS-C/P; Laurent et al., 1999; Ebesutani et al., 2011) are among the most widely-used instruments measuring NA and PA.
Measuring NA and PA is an important area of research given that both have been found to relate to various forms of psychopathology (e.g., Ebesutani et al., 2011; Watson et al., 1995). Based on the Tripartite Model, and its subsequent two-factor model of emotion (cf. Selbo-Bruns et al., 2012), both anxiety and depression share the vulnerability of (high) NA, while depression is uniquely related to (low) PA. Previous researchers found that NA and PA scores can effectively discriminate between anxiety and depression in both adults and youth population (Chorpita & Daleiden, 2002; Clark & Watson, 1991).
Despite this clinical utility, there is a need for more studies to better understand how to use the PANAS-C subscales in different contexts. For example, (a) most studies on the PANAS-C/P have given less attention to the relationship between NA/PA and externalizing disorders. Second, (b) modern classification data analytic approaches have not yet been applied to the PANAS-C/P scales, despite recent studies showing that ‘data-mining’ approaches can lead to an enhanced ability to achieve meaningful classification (e.g., Ebesutani et al., 2012).
The current study will use a clinic-referred sample (Youth N = 268; Parent N = 174) who received assessment through a multi-informant, multi-method approach at a university research and training clinic. Youth ranged in age from 7 to 19 years (M = 13.56, SD = 2.67), were 68% boys, and had diverse diagnoses (e.g., Affective Disorders, Disruptive Behavior Disorders, ADHD).
This study has three major aims. First, we will (a) examine how NA/PA scores correspond with internalizing and externalizing disorders using logistic regression. We hypothesize greater correspondence with internalizing disorders based on child reported PANAS-C scales, and greater correspondence with externalizing disorders based on parent reported PANAS-C scales. This is based on the previous finding that children are more reliable reporting on their own internalizing symptoms, while parents are more reliable reporting on their children’s externalizing symptoms (De Los Reyes et al., 2005).
Second, we will (b) find the optimal cut-points for each NA/PA scale that maximize correspondence with each internalizing and externalizing disorder using Receiver Operating Characteristic (ROC) analyses. Finally, we will (c) identify a combination of cut-points across all PANAS-C/P subscales using Decision Tree analysis. We hypothesize that a series of cut-points will be identified across the PANAS-C/P subscales that will yield greater diagnostic classification than ROC-derived cut-points.
This study will provide important findings related to (a) better identifying youth with internalizing and externalizing problems based on their NA/PA scores, and (b) aiding in more accurate monitoring of client’s affective states across treatment periods.
Seojeong Kim– Research Assistant, Duksung Women's University, Richardson, Texas
Yongmin Sin– Duksung Women's University
Kelsie Okamura– University of Pennsylvania, Philadelphia, Pennsylvania
Brad Nakamura– Associate Professor, University of Hawaii at Manoa, Honolulu, Hawaii
Chad Ebesutani– Associate Professor, Duksung Women's University, Gangbuk-gu, Hawaii, South Korea