Abstract
The Sensory Profile Caregiver Questionnaire (SP; Dunn, 1999) is a widely used pediatric assessment that provides a standard method for professionals to measure the possible contributions of sensory processing to children’s daily performance patterns by providing information about their tendencies to respond to stimuli and which sensory systems are likely contributing or creating barriers to functional performance. The SP contains >125 items organized into three main sections: (1) sensory processing, (2) modulation, and (3) behavioral and emotional responses.
The sensory processing section contains six item categories that measure children’s responses to information taken in through the sensory systems (e.g., auditory, visual, vestibular, tactile, oral). The modulation section contains five item categories that measure children’s ability to monitor and regulate information to generate an appropriate response to the situation (Dunn, 1997). The behavioral and emotional responses section contains three item categories that measure children’s emotional and behavioral responses to sensory experiences. The SP asks caregivers to record the frequency with which their child displays each item behavior on a 5-point Likert scale (1 = always, 2 = frequently, 3 = occasionally, 4 = seldom, 5 = never). For analysis, caregiver responses are totaled on a Summary Score Sheet (Dunn, 1999).
In its initial development, the SP produced two types of scores: section and factor. Section scores provided a visual summary of children’s sensory processing, modulation, and behavioral and emotional response abilities (Dunn, 1999). Factor scores were developed to capture children’s responses to sensory experiences on the basis of not solely their sensory systems but also other aspects of sensory processing (Dunn, 2006). Nine factor scores (Sensory Seeking, Emotionally Reactive, Low Endurance–Tone, Oral Sensory Sensitivity, Inattention–Distractibility, Poor Registration, Sensory Sensitivity, Sedentary, and Fine Motor–Perceptual), each containing items from multiple sensory systems, were used to characterize children by their responsiveness to sensory input (Dunn, 1999). Both sets of scores were analyzed using a classification system based on the performance of children without disabilities (N = 1,037). Children scoring ≥1 standard deviation below the mean were classified as typical, those with scores >2 standard deviations below the mean were classified as having a probable difference, and those with scores <2 standard deviations below the mean were classified as having a definite difference (Dunn, 1999).
The factor and section interpretations provided a standardized way to explain children’s sensory processing patterns to teachers and caregivers that was previously unavailable. By explaining children’s sensory processing problems to teachers, the SP promoted tolerance of children’s behaviors and sensitivity to the difficulties children with sensory processing problems experience (Case-Smith, 1997). The SP helped caregivers become more active participants in their child’s program, facilitated dialogue between the occupational therapist and caregiver about the child’s sensory needs and ways to adapt the environment to meet those needs, and provided caregivers with a greater understanding of their child’s behaviors and responses to certain situations and environments (Case-Smith, 1997). Occupational therapists used the factor and section interpretations to provide guideposts for planning interventions that could target which sensory system might be interfering with the child’s performance or which daily life tasks might be most difficult (Dunn, 1999). Although the SP proved to be a useful tool for professionals, many questions were raised about how to better understand and interpret its scores (Dunn, 2006).
In 2006, the Sensory Profile Supplement User’s Manual (Dunn, 2006) was published to add an updated system of analysis that would reflect a more current understanding of sensory processing and provide a better framework from which to speak to families and teachers. This new level of analysis contained four sensory quadrant scores (Registration, Seeking, Sensitivity, and Avoiding) based on Dunn’s (1997) Model of Sensory Processing. The quadrants measure the degree to which children miss, obtain, detect, or are bothered by sensory input. Data from the original norming sample were reanalyzed under the new quadrant model, resulting in an expanded classification system. The new classification system uses five categories to describe children’s scores along a continuum based on the normal distribution: much less than others, less than others, similar to others, more than others, and much more than others (Dunn, 2006).
The SP has undergone extensive psychometric testing to establish its validity and internal consistency (Dunn, 1999, 2006; Ermer & Dunn, 1998; Kientz & Dunn, 1997). Although the SP is frequently used in current practice to examine the sensory processing patterns of children with autism spectrum disorders, traumatic brain injury, and other conditions (Galvin, Froude, & Imms, 2009; Gere, Capps, Mitchell, & Grubbs, 2009; Hilton, Graver, & LaVesser, 2007; Rodger, Brown, & Brown, 2005; Watling, Deitz, & White, 2001; White, Mulligan, Merrill, & Wright, 2007), it is seldom used to examine intervention outcomes (Polatajko & Cantin, 2010). The SP has been found to detect intervention changes using section scores (Hall & Case-Smith, 2007), indicating that it has the potential to provide evidence of practice efficacy. The SP may not be a widely used outcome measure because it has not yet been examined for test–retest reliability. Without adequate test–retest reliability, practitioners cannot be sure of the stability of SP results over time.
In this study, we examined the SP’s test–retest reliability by comparing the section, factor, and quadrant scores after two administrations 7–14 days apart. We also used confirmatory internal consistency analyses to examine the results and compare them with Dunn’s (1999) original findings.
Method
The research was approved by the institutional review board (IRB) of the State University of New York Downstate Medical Center. The IRB approved using an information sheet to obtain implied consent. The information sheet notified prospective participants that by returning the anonymous SP form they were giving consent for their information to be used in this study. Prospective participants who were not interested in participating were instructed to not fill out or return the SP form. Participants were recruited from six nursery schools and child care centers in the New York metropolitan area. Inclusion criteria for the study included (1) primary caregivers of children ranging in age from 36 to 72 mo and (2) the completion of the SP on two separate occasions 7–14 days apart.
Procedures
Recruitment for the study was conducted in two phases. During the first phase, flyers were posted and sent home with children attending preschool and child care center recruitment sites. During the second phase, research packets were sent home with children at the recruitment sites. Each packet contained an information sheet, a direction sheet, and two SP forms. Interested participants were asked to complete the two SP forms 7–14 days apart. Participants were also asked to write on each SP form their child’s age in months and the date of completion. After completing both SP forms, participants were asked to enclose them in the envelope provided and return them to their child’s preschool or child care center.
Scoring and Data Analysis
The completed SP forms were scored according to the guidelines in the Sensory Profile User’s Manual (Dunn, 1999). Factor and section scores were calculated from SP forms, and quadrant scores were calculated from summary score sheets. To ensure accuracy, all forms were scored twice by occupational therapy graduate students who were trained by an experienced occupational therapist and demonstrated >.80 scorer agreement.
The data were analyzed using SPSS Version 17.01 (SPSS Inc., Chicago). Test–retest reliability was estimated using intraclass correlation coefficients (ICCs), and internal consistency was estimated using Cronbach’s α coefficients. According to Portney and Watkins’ (2008) guidelines, ICCs ≥ .75 are considered good, values between .50 and .74 are considered moderate, and ICCs <.50 are considered poor.
Results
Fifty-five caregivers of children ranging in age from 36 to 72 mo (M = 48.96, SD = 6.47) participated in the study. A cross-tabulation of children’s classifications across the four sensory quadrants during the pretest and posttest conditions depict a normal distribution with minimal skew (Table 1). Descriptive data on the quadrant, factor, and section pretest and posttest scores, ICCs, and Cronbach’s α coefficients are provided in Tables 2, 3, and 4. ICCs for the four quadrants were all in the good range (ICC[2, 1] = .80–.90). ICCs for the factor scores (ICC[2, 1] = .69–.88) and section scores (ICC[2, 1] = .50–.87) ranged from moderate to good. Cronbach’s α coefficients were high across the four quadrants (αs = .89–.95) and factors (αs = .82–.93) and moderate to high across the sections (αs = .67–.93).
Classification of Sensory Profile Quadrant Pretest and Posttest Scores (N = 55)
Descriptive Data, ICCs, and Cronbach’s α Coefficients for the Quadrant Scores
Note. N = 55; p < .001. ICC = intraclass correlation coefficient; M = mean; SD = standard deviation.
Descriptive Data, ICCs, and Cronbach’s α Coefficients for the Factor Scores
Note. N = 55; p < .001. ICC = intraclass correlation coefficient; M = mean; SD = standard deviation.
Descriptive Data, ICCs, and Cronbach’s α Coefficients for the Section Summary Scores
Note. N = 55; p < .001. ICC = intraclass correlation coefficient; M = mean; SD = standard deviation.
Discussion
We examined the test–retest reliability and internal consistency of the SP quadrant, factor, and section scores. Test–retest and internal consistency analyses revealed higher psychometric indexes across the four quadrants (ICCs[2, 1] = .80–.90, αs = .89–.95) than across the factors (ICCs[2, 1] = .69–.88, αs = .82–.93) and sections (ICCs[2, 1] = .50–.87, αs = .67–.93), suggesting that a quadrant-level analysis captures children’s sensory processing patterns more consistently than do factor- or section-level analyses. Compared with previous internal consistency analyses (αQuadrants = .87–.93, αFactors = .72–.92, αSections = .47–.90; Dunn, 1999, 2006), ours found higher α coefficients across all three types of scores. The higher α coefficients found in this study may be the result of a unique quality of the sample. Our sample was smaller in size (N = 55) than Dunn’s (1999) original sample (N = 1,037) and consisted of children in a narrower age range (3.0–6.0 yr) than Dunn’s (3.0–10 yr, 11 mo).
The lowest ICCs were estimated for the visual processing (ICC = .50) and auditory processing (ICC = .67) sections and the sedentary (ICC = .69) factor. Lower scores on these two sections and factor suggest that certain behaviors may be observed or conceptualized differently by caregivers over a short period of time. Alternatively, the larger difference in these scores may be explained by a slight maturation in the child over the 2-wk testing period, interventions, or modifications.
Our findings are helpful in considering the use of the SP to measure intervention outcomes. Analysis at the quadrant level suggests acceptable test–retest reliability and internal consistency, indicating that caregivers’ observations of their children are stable over time. When using the SP as an outcome measure, practitioners can have more confidence that changes detected are more likely the result of intervention than of measurement error or maturation; however, given several limitations of this study, use of the SP to measure intervention changes should be done with caution.
Implications for Occupational Therapy Practice
In this study, the SP quadrant scores were more reliable than section or factor scores and, thus, may be more clinically useful when using the SP as an outcome measure. Further study is required.
Limitations and Future Research
This study was limited by a small sample size of geographic convenience and a lack of demographic information about the study participants. More research is needed to further examine the SP’s test–retest reliability and its utility as an outcome measure. Future studies should include a larger sample with greater demographic representation.
Footnotes
Acknowledgments
We are grateful to all participating caregivers, nursery schools, and child care centers. We especially thank Joyce Sabari for her guidance and knowledge throughout the research process.