This 21-item parent-report measure was designed to rapidly assess and screen for elevated symptomatology in children following exposure to a stressful and/or traumatic event. It is not intended to be a diagnostic instrument. It consists of behaviors that have been identified in the literature as associated with experiencing traumatic events and consists of 17 general behavior items and 4 trauma-specific items. The measure yields scores on the following scales: 1) Anxious/Withdrawn, 2) Fearful, and 3) Acting Out. Of the 4 trauma-specific items, 2 loaded on a separate Talk/Play factor.
Overview
PEDS
Saylor, C.F, Swenson, C. C., Reynolds, S.S., & Taylor, M. (1999). The Pediatric Emotional Distress Scale: A brief screening measure for young children exposed to traumatic events. Journal of Clinical Child Psychology, 28:1, 70-81. Saylor C.F. (2002). The Pediatric Emotional Distress Scale (PEDS). Available from Conway Saylor, PhD, Department of Psychology, The Citadel, 171 Moultrie Avenue, Charleston, SC 29409 (conway.saylor@citadel.edu).
Administration
4-item Likert rating scale: 1) Almost Never, 2) Sometimes, 3) Often, and 4) Very Often)
| Domains | Scale | Sample Items |
|---|---|---|
| Total | Acting Out | Has temper tantrums. |
| Anxious/Withdrawn | Seems sad and withdrawn. | |
| Fearful | Refuses to sleep alone. |
Parallel or Alternate Forms
The author notes that they are currently testing a version used in a post-9/11 media study that asks parents to rate retrospectively both behavior prior to the trauma and behavior since the trauma. Please contact the author should you want additional information.
Psychometrics
Means and standard deviations for 475 children (see "Population Used to Develop Measure") are presented in Saylor (1999) for the full sample and separately for trauma-exposed and non-trauma-exposed children by age (2- 5 and 6-10) and gender. This sample was not considered a normative sample because they were not represented that way in the article and because, with the exception of the Children’s Evaluation Center, they appear to be gathered from a predominantly White, upper-middle-class sample and do not appear to be a true normative sample. Analyses of means suggest that there were no differences based on gender. The only age difference is that younger children scored significantly higher on the acting-out factor. Age analyses examined age differences between children aged 2-5 and those aged 6-10.
Overall Sample: Anxious/Withdrawn = >9.5, Acting out = >13.5, Fearful = >8.5, Total = >27.5; Saylor et al. (1999) suggest using different cutoffs for scales based on maternal education.
| Type: | Rating | Statistics | Min | Max | Avg |
|---|---|---|---|---|---|
| Test-Retest- # days: 42 | Acceptable | Pearson's r | 0.55 | 0.61 | 0.58 |
| Internal Consistency | Acceptable | Cronbach's alpha | 0.72 | 0.85 | 0.77 |
| Inter-rater | Acceptable | Pearson's r | 0.47 | 0.65 | 0.59 |
| Parallel/Alternate Forms |
Saylor et al. (1999) report the data reported below and in the table. TEST-REST RELIABILITY: 6-8 week interval examined with 102 families from the Utah sample (see Population Used to Develop Measure) PEDS Total (.56), Anxious/Withdrawn (.58), Fearful (.55), Acting Out (.61) INTERNAL CONSISTENCY: (Cronbach’s alpha) PEDS Total (.85), Anxious/Withdrawn (.74), Fearful (.72), Acting Out (.78) INTERRATER RELIABILITY: Assessed examining correlations between 111 married mothers and fathers in the Utah sample. PEDS Total (.65), Anxious/Withdrawn (.58), Fearful (.47), Acting Out (.64) OTHER STUDIES A study of 29 homeless children (Page & Nooe, 1999) aged 2-10 (M=68.6 months, 59% White, 32% African American) used the 17-item PEDS and reported the following alphas: Total (.84), Anxious/Withdrawn (.62), Fearful (.21), Acting Out (.87). Given the low internal consistency of the Fearful subscale, they excluded it from further analyses.
The author notes that they recently collected norms on a sample of 800 children from Spain. Please contact the author for more information. From Saylor et al. (1999). Through discussions with expert investigators who had studied the effects of an earthquake in Southern California and a hurricane in South Carolina, 17 items were identified. Items were identified from checklists (Child Behavior Checklist, the Reaction Index, the Revised Children’s Manifest Anxiety Scale, Eyberg Checklist) as having been useful with children exposed to natural disasters. Four items were added that were considered characteristic of traumatized children based on the DSM-III-R criteria and outcome literature. Doctoral-level psychologists also read items and identified those that were redundant, inappropriate for children aged 2-10, or too difficult for a parent with an 8th-grade reading level to understand.
| Validity Type | Not known | Not found | Nonclincal Samples | Clinical Samples | Diverse Samples |
|---|---|---|---|---|---|
| Convergent/Concurrent | Yes | ||||
| Discriminant | Yes | Yes | |||
| Sensitive to Change | Yes | Yes | |||
| Intervention Effects | |||||
| Longitudinal/Maturation Effects | |||||
| Sensitive to Theoretically Distinct Groups | Yes | Yes | |||
| Factorial Validity | Yes | Yes |
Saylor et al. (1999): Concurrent validity analyses were conducted with 50 children being evaluated for possible sexual abuse victimization. PEDS Scores have been found to correlate with scores on related measures. Total PEDS and Acting Out scores were correlated with the ECBI (r=.62, p<.001 and r=.86, p<.001, respectively). Smaller correlations were seen with the PEDS Anxiety/Withdrawn and Fearful factors and the ECBI (r=.42, p<.007 and r=.32, p<.04, respectively). This would be expected, given that the ECBI does not measure internalizing/anxiety symptoms. Total PEDS was significantly correlated with PTSD symptoms assessed using the parents’ report on the Reaction Index (r=.62, p<.001) as was the Fearful factor (r=.59, p<.001), and the Anxious/Withdrawn Factor (r=.62, p<.001). The Talk/Play factor, however, was not related to the Reaction Index scores or to the ECBI. 1. In a study examining the effects of indirect exposure to the 9-11 attacks (Saylor, Cowart, Lipovsky, Jackson, & Finch (2003), PEDS total scores were related to exposure to negative media images (r=.23 p<.002) and positive media images (r=.22, p<.002). The PTSD subscale of the PEDS was also related to exposure to negative and positive media images (r=.24, p<.001 and r=.29, p<.001, respectively). 2. Page, & Nooe (1999) reported that homeless children scored higher compared to norms (with numbers >1 SD above mean). In this sample, they reported significant correlations among scales (with the exception of the Fearful scale, which did not have good internal consistency and was not used)?????: Anxiety and Acting Out (r=.58, p<.01), Acting Out and Total (r=.92, p<.01), and Total and Anxiety (r=.78, p<.01). A composite risk index was significantly related to PEDS Anxiety (r=.51, p<.01) and Total (r=.40, p<.05) scores, with demographic variables controlled. PEDS scores were significantly related to variables associated with being homeless, with all scores related to the number of cities in which the family had lived, and the Anxiety scale significantly correlated with the number of schools the child had attended (r=.49, p<.01). The measure has been used in a large study examining 1,739 children’s exposure to community violence. Of the 565 children for whom demographics were available: 35% African American, 23.6% Caucasian, 34% unknown. Age: 32.2% age 5 or under. Children were referred to the Children Who Witness Violence Program of Cuyahoga County, Ohio, for evaluation and services following exposure to violence. Violence included domestic violence (87%) and much smaller numbers for other violence including homicide, shooting, completed suicide, and sexual assault. Clinicians and caregivers completed the PEDS and both groups reported that high percentages of children aged 2-7 were experiencing clinically significant levels of anxiety (92% of clinicians and 85% of caregivers). In addition, both rated “8 of every 10 children to have clinically significant problems with acting out behavior.” 3. The PEDS has been shown to differentiate between different groups of children including trauma-exposed and non-trauma-exposed children (Saylor et al., 1999) and children with different types of trauma (Stokes et al., 1995). Swenson et al., (1996) compared preschoolers who experienced a Class IV hurricane 14 months earlier to peers with no history of exposure to natural disaster. Hurricane-exposed children showed significantly higher behavior problems, anxiety, and withdrawal. Stokes et al. (1995) compared PEDS scores for children who had experienced possible sexual abuse, a Class IV hurricane, severe negative life events, or no trauma. Children with possible sexual abuse had the highest scores on all PEDS subscales. Those with negative life events or exposure to a hurricane had higher scores than those with no known trauma. 4. Swenson, Brown, & Sheidow (2003) studied 37 children aged 6-13 (M=9.5, SD=2.1) with substantiated cases of physical abuse. 54% were female, 46% male; 62% were African American, 32% were Caucasian. Over time, significant reductions were found for PEDS scores and for scores on the Children’s Depression Inventory and Child Behavior Checklist, regardless of whether the child had received treatment. FACTOR ANALYSES (Saylor et al., 1999) Principal components analysis with an oblique rotation of the first 17 items of the scale suggested three factors based on the interpretability and simplicity of the structure. All factors had eigen values >1.5. The factors were labeled: 1) Acting Out, 2) Fearful, and 3) Anxious/Withdrawn. The full scale (all 21 items) was factor analyzed using data from trauma-exposed children (the Hurricane and Children’s Evaluation Center samples described under “Population Used to Develop Measure”). The authors report that for nontrauma- exposed children, parents did not complete the additional 4 trauma items and so were not included in this analysis. The factor analysis yielded identical factors for the 17 items. Two of the trauma items loaded on a fourth factor, labeled Talk/Play. One trauma item (“Avoids talking about the traumatic event when asked”) loaded on the Anxious/Withdrawn factor, and the other (“Seems fearful of things that are reminders of the trauma”) loaded on the Fearful factor.
| Not Known | Not Found | Nonclinical Samples | Clinical Samples | Diverse Samples | |
|---|---|---|---|---|---|
| Predictive Validity: | |||||
| Postdictive Validity: |
Saylor et al. (1999) report that PEDS scores alone do not appear to predict trauma-exposed versus non-trauma-exposed status, but PEDS scores blocked by maternal education correctly classified 78% of the cases; 9.5% false positive rate and 12.5% false negative rate.
1. Ohan, Myers, & Collett (2002) suggest that scores should not be used to determine whether the child has been abused or exposed to trauma because sensitivity and specificity studies have compared traumatized to nonclinical, nontraumatized samples but have not compared traumatized children to other clinical samples. 2. In addition, it should be noted that classification rates were based not only on PEDS scores but also on PEDS scores cutoffs based on maternal education. 3. The majority of the studies appear to have used the 17-item PEDS. There are 4 specific items that were added. The factor analysis reveals that 2 of the items load on a 4th factor, but studies do not appear to use this 4th factor. In addition, the data do not appear to support the validity of the factor (lack of correlations with either the Reaction Index or ECBI). However, this may be due to the fact that it has only two items. 4. While Saylor et al. (1999) report means and standard deviations for 475 children, the sample does not appear to be a representative sample, and caution should be used if these data are used for normative purposes.
Translations
| Language: | Translated | Back Translated | Reliable | Good Psychometrics | Similar Factor Structure | Norms Available | Measure Developed for this Group |
|---|---|---|---|---|---|---|---|
| 1. Spanish | Yes | Yes | Yes |
Population Information
POPULATION USED TO DEVELOP MEASURE From Saylor et al. (1999) Four samples of children were used in the psychometric study: 1. UTAH SAMPLE: 182 children aged 2-10 (M=7); 50% female; 91% White, 1% African American, 8% Other; 97% of mothers attended or completed college; 91% of mothers were married. Children were attending a university-sponsored school in Logan, Utah. 2. BOSTON SAMPLE: 64 children enrolled in a university-affiliated developmental kindergarten. Children were aged 2-10 (M=7.9); 50% female; 95% White, 5% African American; 93% of mothers attended or completed college; 92% of mothers were married. 3. HURRICANE SAMPLE: 179 children (mean age=3.6), recruited from two large private schools, who had been directly exposed to Hurricane Hugo (in Charleston, South Carolina); 47% female, 53& male; 99% White, 1% Other; 75% of mothers attended or completed college, 15% attended vocational school, and 10% attended or completed high school; 96% of mothers were married. 4. CHILDREN’S EVALUATION CENTER: 50 children (mean age=7.9), recruited from an outpatient evaluation center for children and adolescents, who may have been sexually abused. 59% female, 41% male; 91% Caucasian, 9% African American; 52% of mothers attended or completed high school, 17% attended vocational school, and 31% attended or completed college; 65% were married, 8% were divorced, and 21% had another marital status.
| Population Type: | Measure Used with Members of this Group | Members of this Group Studied in Peer-Reviewed Journals | Reliable | Good Psychometrics | Norms Available | Measure Developed for this Group |
|---|---|---|---|---|---|---|
| 1. Developmental disability | Yes | |||||
| 2. Disabilites | Yes | |||||
| 3. Lower socio-economic status | Yes | Yes | Yes | |||
| 4. Spicial needs | Yes | Yes | ||||
| 5. Rural populations | Yes | Yes |
Pros & Cons/References
1. It screen for symptoms commonly seen in trauma-exposed children. 2. It is brief. 3. Individuals items are clearly written and easy to understand. 4. Items are derived from widely accepted measures. 5. The measure is free.
1. As noted by Feeney, Foa, Treadwell, & March (2004), the measure does not assess all PTSD symptoms (it was not designed for this purpose), and therefore does not provide a true measure of PTSD symptomatology. Also, the Reexperiencing factor may be problematic, given that it consists of 2 items. 2. While psychometrics are promising, more research is needed. Concurrent validity was examined only with a clinical sample. Also, as suggested by Ohan, Myers, & Collett (2002), more research is needed to determine whether the PEDS can distinguish between traumaexposed children and other clinical samples. 3. Little research has been conducted with diverse samples. 4. While Saylor et al. (1999) report means and standard deviations for 475 children, the sample does not appear to be a representative sample, and caution should be used if these data are used for normative purposes. 5. Although Saylor et al. (1999) found good internal consistency for all scales, Page & Nooe (1999) found low internal consistency for the Fearful subscale (alpha=.21). 6. The measure was developed for children aged 2-10; however, examination of the mean age of the children in the different studies suggests that few studies have used the measure with younger children. More research is needed with younger children.
A PsychInfo literature search (7/05) of "Pediatric Emotional Distress Scale" or “PEDS” anywhere revealed that the measure has been referenced in 10 peer-reviewed journal articles. Two additional references were identified while conducting the review. A sampling of these articles is included below. 1. Drotar, D., Flannery, D., Day, E., Friedman, S., Creeden, R., Gartland, H., et al. (2003). Identifying and responding to the mental health service needs of children who have experienced violence: A community-based approach. Clinical Child Psychology & Psychiatry, 8(2), 187-203. 2. Feeny, N. C., Foa, E. B., Treadwell, K. R. H., & March, J. (2004). Posttraumatic stress disorder in youth: A critical review of the cognitive and behavioral treatment outcome literature. Professional Psychology: Research & Practice, 35(5), 466-476. 3. Ohan, J. L., Myers, K., & Collett, B. R. (2002). Ten-year review of rating scales. IV: Scales assessing trauma and its effects. Journal of the American Academy of Child & Adolescent Psychiatry, 41(12), 1401-1422. 4. Page, T. F., & Nooe, R. M. (1999). Relationships between psychosocial risks and stress in homeless children. Journal of Social Distress & the Homeless, 8(4), 255-267. 5. Saylor, C. F., Cowart, B. L., Lipovsky, J. A., Jackson, C., & Finch, A. J. J. (2003). Media exposure to September 11: Elementary school students' experiences and posttraumatic symptoms. American Behavioral Scientist, 46(12), 1622-1642. 6. Saylor, C. F., Swenson, C. C., Reynolds, S. S., & Taylor, M. (1999). The pediatric emotional distress scale: A brief screening measure for young children exposed to traumatic events. Journal of Clinical Child Psychology, 28(1), 70-81. 7. Stokes, S., Saylor, C. F., Swenson, C. C., & Daugherty, T. (1995). Comparison of children’s behaviors following three types of stressors. Child Psychiatry and Human Development, 26, 113-123. 8. Strand, V. C., Sarmiento, T. L., & Pasquale, L. E. (2005). Assessment and screening tools for trauma in children and adolescents: A review. Trauma, Violence, & Abuse, 6(1), 55- 78. 9. Swenson, C. C., Brown, E. J., & Sheidow, A. (2003). Medical, legal, and mental health service utilization by physically abused children and their caregivers. Child Maltreatment: Journal of the American Professional Society on the Abuse of Children, 8(2), 138-144. 10. Swenson, C. C., Saylor, C. F., Powell, M. P., Stokes, S. J., Foster, K. Y., & Belter, R. W. (1996). Impact of a natural disaster on preschool children: Adjustment 14 months after a hurricane. American Journal of Orthopsychiatry, 66(1), 122-130.
The author read the review and provided comments, which were integrated. He also provided permission to post the Spanish and English versions of the measure on nctsn.org