Evaluation in Ayres Sensory Integration ® (EASI) Tactile Perception Tests: Construct Validity and Internal Reliability

Abstract

Importance: Assessment of tactile perception is foundational for addressing aspects of occupational performance.

Objective: To evaluate the construct validity and internal reliability of four new tactile perception tests.

Design: Causal comparative groups design.

Settings: Homes, schools, and therapy practices across the United States.

Participants: Children ages 3 to 12 yr: typically developing (n = 174) and those with sensory integration concerns (n = 153).

Outcomes and Measures: Rasch analyses to evaluate construct validity; analysis of covariance to evaluate group differences.

Results: The Rasch model confirmed evidence of construct validity for each of the four tests. The typically developing group scored significantly higher than the clinical group on all tests (η² _p = .040–.105, p < .001). Person reliability indices and strata indicated moderate to strong internal reliability (Rasch person reliability indices = .69–.87; strata = 2.33–3.82).

Conclusions and Relevance: The Evaluation in Ayres Sensory Integration^® (EASI) Tactile Perception Tests are reliable and valid measures for assessing tactile perception in children ages 3 to 12 yr. Findings suggest that these tests are likely to be clinically useful and appropriate for children in this age range and may provide critical information regarding underlying sensory functions necessary for optimal occupational performance.

What This Article Adds: This article provides data supporting the reliability and validity of the EASI Tactile Perception Tests in a U.S. sample. These assessments can be used by therapists trained in their administration to assess tactile functions that may affect participation in activities, tasks, and occupations.

The Evaluation in Ayres Sensory Integration (EASI) Tactile Perception Tests are reliable and valid measures and can be used by therapists trained in their administration to assess tactile functions in children that may affect participation in activities, tasks, and occupations.

Touch is one of the earliest senses to develop, playing a critical role in human development across the lifespan (Ayres, 1972; Linden, 2015). Tactile perception provides essential sensory information that informs social communication (Ellingsen et al., 2016), body awareness (Tresilian, 2012), postural control (Hadders-Algra & Carlberg, 2008), and motor performance (Shumway-Cook & Woolacott, 2007). Furthermore, touch integrates with other senses, such as vision and proprioception, to guide actions (Ernst & Banks, 2002) and enables participation in activities that require seeing, reaching, touching (Corbetta & Snapp-Childs, 2009), and socializing (Cascio et al., 2019). Ayres et al. showed strong and significant associations between tactile perception and praxis (Ayres, 1989; Mailloux et al., 2011), and Smith Roley et al. (2015) found relationships among tactile perception, praxis, and social participation in a cohort of children with autism. These relationships explain why many children with poor tactile perception have trouble planning actions needed for social participation and daily life activities, such as donning clothing, playing with toys, or using a writing utensil. Thus, it is crucial that occupational therapists assess tactile perception when considering sensory factors that may affect participation in occupations.

Existing Measures of Tactile Perception in Children

Ayres developed many measures of sensory perception and praxis over several decades to help inform sensory integration theory. The Sensory Integration and Praxis Tests (SIPT; Ayres, 1989) include four tests designed to measure tactile perception. However, these tests have dated norms and can be inaccessible to many practitioners because of cost (Mailloux et al., 2018; Schaaf & Lane, 2015). Thus, there is a need for updated, accessible, psychometrically strong, performance-based assessments of tactile perception. In this study, we evaluated a new set of tests measuring tactile perception in children that are part of the Evaluation in Ayres Sensory Integration^® (EASI; Mailloux et al., 2018) measure. The EASI consists of 20 tests of sensory and motor functions, four of which measure tactile perception: Tactile Perception: Localization (TP:L); Tactile Perception: Designs (TP:D); Tactile Perception: Shapes (TP:S); and Tactile Perception: Oral (TP:O). The purpose of this article is to report the evidence for the construct validity and internal reliability of these four tests.

Method

We examined the evidence for construct validity and internal reliability of the four EASI Tactile Perception Tests with typically developing (TYP) children and children with sensory integration (SI) concerns using Rasch analysis. Approval for the study was obtained through the Thomas Jefferson University Institutional Review Board.

Participants

We studied a convenience sample of 327 children from 12 states in the continental United States, including 174 TYP children (78 boys, 96 girls) and 153 children with SI concerns (96 boys, 57 girls). There were more boys than girls in the SI sample, which is consistent with prevalence studies of many developmental conditions (Boyle et al., 2011), but no significant differences between the groups for age. We did not gather data on participant race, ethnicity, or socioeconomic status in this small study.

Inclusion criteria for both groups included chronological age between 3 yr, 0 mo, and 12 yr, 11 mo, and English as the primary language. We included children in the TYP group who were developing and performing within age expectations according to parent report; we excluded children if they had known or suspected medical concerns or behaviors suggestive of SI concerns or if they had received any prior SI intervention. We included children in the SI group who were identified by an occupational therapist, physical therapist, or speech-language pathologist as having SI concerns, on the basis of standardized sensory assessment data from tools such as the Sensory Processing Measure, the Sensory Profile, and the SIPT, as well as from clinical observations. We included children who met the criteria for SI concerns in the SI group, including those with autism spectrum disorder, attention deficit disorder, speech or language delay, anxiety disorder, or a regulatory disorder. We excluded those with a physical or intellectual disability or uncorrected vision or hearing deficits. Parent responses on a participant information form aided testers in verifying inclusion and exclusion criteria.

Procedures

Examiners included 67 occupational therapists, 1 physical therapist, and 1 speech-language pathologist. All were trained to competence in the ethical conduct of research and EASI test administration by means of online training modules. Examiners obtained parental consent (and assent from participants age 7 yr or older). Most children completed testing on the four tactile tests in one 1-hr session; 4 children required two or more sessions. Examiners submitted deidentified test scores to a confidential database.

Instruments

Table 1 describes each EASI Tactile Perception Test, lists the number of items, and describes the scoring criteria. Not all children completed each test; therefore, the final sample sizes for each of the four tests varied.

Table 1.

Description of Evaluation in Ayres Sensory Integration Tactile Tests and Scoring

Test	Description of the Ability Tested	Subscore	No. of Items	Scores
Tactile Perception: Designs (n _TYP = 159; n _SI = 134)	Ability to perceive and indicate the location of tactile stimuli applied to arms and hands	Accuracy	23	2 = correct, 1 = partially correct, 0 = incorrect
Tactile Perception: Localization (n _TYP = 157; n _SI = 145)	Ability to perceive and replicate designs drawn on arms and hands	Accuracy	20	1 = correct, 0 = incorrect
Tactile Perception: Shapes (n _TYP = 172; n _SI = 152)	Ability to perceive and match shapes and textures felt in one hand to both shapes seen and shapes and textures felt with the other hand	Accuracy	19	1 = correct, 0 = incorrect
Tactile Perception: Shapes (n _TYP = 172; n _SI = 152)		Time	19	Time at choice, recorded in whole seconds
Tactile Perception: Oral (n _TYP = 160; n _SI = 143)	Ability to perceive and match shapes felt in the mouth to shapes seen	Accuracy	11	1 = correct, 0 = incorrect

Note. SI = children with sensory integration concerns; TYP = typically developing children.

Data Analysis

We used Winsteps, Version 4.5.2, to conduct iterative Rasch analyses, revise the item set and scoring structure of each test, and evaluate the evidence for construct validity and internal reliability of the revised tests. The Rasch model is a latent-trait psychometric model that constructs linear measures from ordinal raw data. The Rasch model generates item-difficulty and person-ability measures along a single, unidimensional scale. Readers can find a detailed description of the Rasch model elsewhere (e.g., Bond et al., 2020). For the three tests with dichotomous items (TP:L, TP:S, and TP:O), we used the standard Rasch model. For TP:D, which uses a 3-point scale, we used the partial credit Rasch model (PCM). The PCM is appropriate for evaluating items with polytomous item response categories (Bond et al., 2020). We conducted an initial round of data analyses that led to item elimination and test revisions. We then conducted the final data analyses, reported in the Results section, on the revised tests.

Analysis of Construct Validity

For each of the tactile tests, we evaluated four Rasch-generated indicators of construct validity associated with the accuracy subscales of each test. These were point–measure correlations, goodness-of-fit statistics, ordinality of rating scales, and Wright maps.

We used point–measure correlations to evaluate whether items aligned in the same direction as the latent variable (i.e., a higher score on an item correlates with a higher score on the overall measure). We expected all items to demonstrate positive correlations.

Goodness-of-fit statistics indicated the degree to which the data conformed to the Rasch model. Mean square (MnSq) values indicated the degree of randomness in the measurement system; standardized Z (Z _std) values indicated the improbability of the data if the data fit the model perfectly (Linacre, 2002). Both MnSq and Z _std statistics are reported as infit (weighted or inlier sensitive) and outfit (unweighted) statistics. We expected 95% of the items to demonstrate MnSq values within the acceptable range of 0.5 to 1.5 and Z _std values within the acceptable range of −2 to 2. If both criteria were violated for either infit or outfit for any item, we considered the item to have failed to conform to Rasch expectations.

Regarding the ordinality of rating scales, we expected rating scale categories for each item (e.g., 0, 1, 2) to be used at least 10 times and to progress logically such that lower categories correlated with lower average measure scores more than did higher categories. If either assumption was violated, we considered collapsing rating scale categories. Finally, Wright maps allowed us to examine the match between sample ability and item difficulty.

We used known-groups analyses to evaluate data on each of the tactile tests for evidence of construct validity. We conducted analyses of covariance using Rasch person–ability measures as dependent variables, group (TYP and SI) as a fixed factor, and age as a covariate for each test. We expected children in the TYP group to score significantly higher on all tests and subtests than children in the SI group (α = .05). For TP:S, we also evaluated group differences on summed time to complete each item. We expected children in the SI group to have significantly longer summed time scores. To evaluate effect size, we calculated partial η² (η² _p) values; we considered η² _p values ≥.14 to represent large effects, values ranging from .06 to .14 to represent medium effects, and values ranging from .01 to .06 to represent small effects (Cohen, 1988).

Analysis of Internal Reliability

We evaluated two indicators of internal reliability. First, person-reliability indices suggest the reproducibility of relative measure locations (Linacre, 2017). We considered person-reliability indices ≥.8 to indicate strong internal reliability. Second, strata values indicate the levels of performance that the test can reliably distinguish. We determined strata on the basis of Winsteps-generated person–separation indices using the formula $strata = \frac{4 G + 1}{3}$ , where G = person–separation index (Linacre, 2017). Ordinarily, strata ≥2 indicate adequate internal reliability (i.e., separate good from poor performers). However, because we assumed that performance on these tests would increase with age, we sought at least three strata as a reflection of adequate internal reliability.

Results

The following sections describe Rasch analysis results for each of the Tactile Perception Tests.

Construct Validity

All items had uniformly positive point–measure correlations. All items on the TP:D, TP:L, and TP:O fit the Rasch model, whereas 16 of 19 (84.2%) TP:S items fit the model. The three misfitting items on the TP:S showed outfit statistics outside of our desirable range but infit statistics within the range (Table 2). Overall, 95.8% of items across all four tests met Rasch model expectations. Rating scales for all four tests progressed appropriately, and all categories were used at least 10 times for each item. Visual examination of Wright maps indicated adequate item spread and match between sample ability and item difficulty (see the Supplemental Appendix, available online with this article at https://research.aota.org/ajot). The Supplemental Appendix shows item measures and fit statistics for all of the items.

Table 2.

Misfitting Items

TP:S Item	Measure	Pt–Meas Corr.	Infit		Outfit
			MnSq	Z _std	MnSq	Z _std
			Item 1	−2.17	.47	0.87	−0.73	0.37	−2.03
Item 18	0.41	.30	1.37	5.43	1.59	5.08
Item 19	1.61	.36	1.28	4.36	1.63	5.15

Note. Boldface indicates goodness-of-fit statistics outside the desired criteria. MnSq = mean square; Pt–Meas Corr. = point–measure correlation; TP:S = Tactile Perception: Shapes test; Z _std = standardized Z value.

Table 3 contains the results of known-groups analyses for accuracy and time scores. The TYP group scored significantly higher on the total accuracy scores of all four tests. Effect sizes for total test accuracy ranged from small to moderate (η² _ps = .04–.105). As expected, the SI group had a significantly higher mean time score on the TP:S compared with the TYP group.

Table 3.

Known-Groups Analyses: Accuracy Scales and TP:S Time Scales

Test	Scores	M ^a (SE)		F	df	p	Effect Size (η² _p)
Test	Scores	TYP	SI	F	df	p	Effect Size (η² _p)
TP:D	Accuracy	0.23 (0.07)	−0.22 (0.06)	24.65	1, 290	<.001	.078
TP:L	Accuracy	1.35 (0.10)	0.50 (0.10)	35.20	1, 300	<.001	.105
TP:S	Accuracy	1.43 (0.09)	0.95 (0.10)	12.69	1, 303	<.001	.040
TP:S	Time, s	149.81 (5.57)	182.51 (6.01)	15.92	1, 300	<.001	.050
TP:O	Accuracy	0.39 (0.11)	−0.16 (0.11)	12.99	1, 300	<.001	.042

Note. SI = sensory integration; TP:D = Tactile Perception: Designs; TP:L = Tactile Perception: Localization; TP:S = Tactile Perception: Shapes; TP:O = Tactile Perception: Oral; TYP = typically developing.

Age-adjusted mean.

Internal Reliability

Table 4 provides the results of reliability analyses. Rasch reliability indicators (person-reliability and strata) indicated strong reliability for the TP:D (0.87 and 3.82, respectively). The TP:L, TP:S, and TP:O fell below our thresholds, suggesting moderate internal reliability.

Table 4.

Internal Reliability

Test	Person-Reliability Index	Strata
TP:D	.87	3.82
TP:L	.74	2.64
TP:S	.75	2.64
TP:O	.69	2.33

Note. TP:D = Tactile Perception: Designs; TP:L = Tactile Perception: Localization; TP:S = Tactile Perception: Shapes; TP:O = Tactile Perception: Oral.

Discussion

The findings from this study revealed strong evidence for construct validity for all of the Tactile Perception Tests. Correlations between item and total test measures (i.e., point–measure correlations) were positive, fit statistics for nearly all test items were within the a priori acceptable range, and the match between item difficulty and child ability was acceptable. Furthermore, the reliability indices and strata revealed moderate to strong evidence for internal reliability. Total accuracy scores from all four EASI Tactile Perception Tests differentiated between the TYP and SI groups, with the TYP group scoring significantly higher than the SI group.

As a result of these analyses, we identified three misfitting items on the TP:S. Items 18 and 19 are texture items: They require the child to match a 3D-printed texture held in one hand with a matching stimulus felt in the other hand in a field of four items. Both items demonstrated underfitting outfit statistics (MnSq > 1.5, Z _std > 2.0), which is likely the result of lucky guessing. We chose to retain these items but added an additional distractor to the test to decrease guessing. Item 1 required the child to match a square felt in one hand with a visual stimulus on a field of 16 items. This item demonstrated overfitting outfit statistics. Overfitting items are unlikely to degrade measurement (Linacre, 2002); therefore, we chose to retain this item. During future studies, we will closely monitor the performance of these items.

In comparison with the SIPT, the EASI provides an updated set of test items for tactile perception that are easily accessible and cost-effective. The EASI tactile tests are intended for a wider age range (3–12 yr vs. 4–8 yr for the SIPT) and will have a broader normative sample for comparison (an international sample for the EASI vs. a normative sample from only the United States and Canada for the SIPT). The EASI also includes a test of oral tactile perception (EASI TP:O), a function not tested by the SIPT. Given the important role that tactile perception plays in early development and its critical role in supporting praxis for successful participation in occupations (Ayres, 1989), assessment of tactile perception is essential for occupational therapists. Furthermore, best practice in occupational therapy supports the use of assessment data to guide intervention (Gillen et al., 2019). Assessment provides specific data about the factors that may be affecting occupational performance so that appropriate occupational therapy intervention can be implemented. Thus, when the EASI is finalized, it will provide an accessible, cost-effective means for valid and reliable testing of tactile perception, internationally, for children ages 3 to 12 yr.

Implications for Occupational Therapy Practice

This article provides evidence for reliability and validity of data gathered with the EASI Tactile Perception Tests in a U.S. sample and has the following implications for occupational therapy practice:

▪ Tactile functions are an important factor that may affect a child’s ability to participate in tasks, activities, and occupations; thus, assessment of these functions with the EASI can be useful for therapists.

▪ Accurate and precise assessment data are an important foundation for occupational therapy practice. The EASI Tactile Tests can provide useful information to determine whether tactile functions are affecting a child’s ability to participate in tasks, activities, and occupations, particularly children with suspected sensory integration deficits.

▪ The EASI Tactile Tests provide data that can be used as part of the assessment and treatment planning process in occupational therapy to create effective, individually tailored interventions that address tactile factors that may be affecting function and participation.

In addition, an international normative data study, currently in process, will reveal the status of the evidence for the validity and reliability of the EASI Tactile Perception Tests with a worldwide sample.

Limitations

Although this study is strong in its rigor, one limitation is that children in this study were from the United States only. In future studies, we will include representative samples from more than 75 countries to develop internationally normed data. Another limitation of this study was the length of the research versions of the tactile tests, because they did not yet have basal and ceiling levels by age or discontinuation criteria, which will be established once the larger normative data set is collected. These parameters, along with the possibility of computer adaptive testing, will enhance completion rates. In addition, because of the small sample size, we did not stratify results by age, race, diagnostic group, or other demographic variables for these initial analyses, but we plan to do so in future studies. We will also further evaluate additional aspects of validity and reliability, including test–retest and interrater reliability, as well as congruent validity, comparing the EASI to other performance-based tests and questionnaires.

Conclusion

The EASI Tactile Perception Tests are appropriate measures for assessing tactile perception in children ages 3 to 12 yr. Because tactile perception is associated with the motor planning and praxis abilities needed for successful participation in daily life, including social participation, these tests provide essential information for intervention planning for children with sensory integrative and other developmental disorders.

Supplemental Material

Supplemental Appendix

Supplementary material for Evaluation in Ayres Sensory Integration® (EASI) Tactile Perception Tests: Construct Validity and Internal Reliability

Supplementary material, sj-pdf-1-aot-10.5014_ajot.2023.050053.pdf for Evaluation in Ayres Sensory Integration® (EASI) Tactile Perception Tests: Construct Validity and Internal Reliability by Roseann C. Schaaf, Kelly Auld Wright, Zoe Mailloux, Patricia Grady, L. Diane Parham, Susanne Smith Roley and Anita Bundy in The American Journal of Occupational Therapy

Footnotes

Acknowledgments

We thank the testers, families, and children who participated in this study. We also acknowledge statistical assistance from Dr. Ben Leibly of Thomas Jefferson University; Dr. Steven Paul of the University of California, San Francisco; and Rachel Dumont for assistance with ethics board approvals.

References

Ayres

(1972). Sensory integration and learning disorders. Western Psychological Services.

Ayres

(1989). Sensory Integration and Praxis Tests: SIPT manual. Western Psychological Services.

Bond

T. G.

Yan

, & Heene

(2020). Applying the Rasch model: Fundamental measurement in the human sciences (4th ed.). Routledge. https://doi.org/10.4324/9780429030499

Boyle

C. A.

Boulet

Schieve

L. A.

Cohen

R. A.

Blumberg

S. J.

Yeargin-Allsopp

, . . . Kogan

M. D.

(2011). Trends in the prevalence of developmental disabilities in US children, 1997–2008. Pediatrics , 127 , 1034–1042. https://doi.org/10.1542/peds.2010-2989

Cascio

C. J.

Moore

, & McGlone

(2019). Social touch and human development. Developmental Cognitive Neuroscience , 35 , 5–11. https://doi.org/10.1016/j.dcn.2018.04.009

Cohen

(1988). Statistical power analysis for the behavioral sciences. Routledge Academic.

Corbetta

, & Snapp-Childs

(2009). Seeing and touching: The role of sensory-motor experience on the development of infant reaching. Infant Behavior and Development , 32 , 44–58. https://doi.org/10.1016/j.infbeh.2008.10.004

Ellingsen

D. M.

Leknes

Løseth

Wessberg

, & Olausson

(2016). The neurobiology shaping affective touch: Expectation, motivation, and meaning in the multisensory context. Frontiers in Psychology , 6 , 1986. https://doi.org/10.3389/fpsyg.2015.01986

Ernst

M. O.

, & Banks

M. S.

(2002). Humans integrate visual and haptic information in a statistically optimal fashion. Nature , 415 , 429–433. https://doi.org/10.1038/415429a

10.

Gillen

Hunter

E. G.

Lieberman

, & Stutzbach

(2019). AOTA’s top 5 Choosing Wisely^® recommendations. American Journal of Occupational Therapy , 73 , 7302420010. https://doi.org/10.5014/ajot.2019.732001

11.

Hadders-Algra

, & Carlberg

E. B.

(2008). Postural control: A key issue in developmental disorders. Mac Keith Press.

12.

Linacre

J. M.

(2002). What do infit, outfit, mean-square and standardized mean? Rasch Measurement Transactions , 16(2), 878. https://www.rasch.org/rmt/rmt162f.htm

13.

Linacre

J. M.

(2017). Reliability and separation of measures. https://www.winsteps.com/winman/reliability.htm

14.

Linden

D. J.

(2015). Touch: The science of hand, heart, and mind. Viking Press.

15.

Mailloux

Mulligan

Smith Roley

Blanche

Cermak

Coleman

G. G.

, . . . Lane

C. J.

(2011). Verification and clarification of patterns of sensory integrative dysfunction. American Journal of Occupational Therapy , 65 , 143–151. https://doi.org/10.5014/ajot.2011.000752

16.

Mailloux

Parham

Smith Roley

Ruzzano

, & Schaaf

R. C.

(2018). Introduction to the Evaluation in Ayres Sensory Integration^® (EASI). American Journal of Occupational Therapy , 72 , 7201195030. https://doi.org/10.5014/ajot.2018.028241

17.

Schaaf

R. C.

, & Lane

A. E.

(2015). Toward a best-practice protocol for assessment of sensory features in ASD. Journal of Autism and Developmental Disorders , 45 , 1380–1395. https://doi.org/10.1007/s10803-014-2299-z

18.

Shumway-Cook

, & Woolacott

M. H.

(2007). Motor control: Translating research into clinical practice (3rd ed.). Lippincott Williams & Wilkins.

19.

Smith Roley

Mailloux

Parham

L. D.

Schaaf

R. C.

Lane

C. J.

, & Cermak

(2015). Sensory integration and praxis patterns in children with autism. American Journal of Occupational Therapy , 69 , 6901220010. https://doi.org/10.5014/ajot.2015.012476

20.

Tresilian

(2012). Sensorimotor control & learning: An introduction to the behavioral neuroscience of action. Palgrave Macmillan. https://doi.org/10.1007/978-1-137-00511-3

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.23 MB