Psychometric properties of Korean Infant Sensory Profile 2

Abstract

Objectives

This study aimed to evaluate test-retest reliability and convergent validity of the Korean Infant Sensory Profile 2 (K-ISP2) and explore whether the sensory processing of infants differs according to age, gender, and culture.

Materials and methods

Participants were 140 Korean caregivers of infants aged 0–6 months. They completed the K-ISP2 and the Korean version of ‘What My Baby Is Like’ (K-WBL), a scale for measuring infant temperament, for convergent validity. The K-ISP2 was re-administrated for test-retest reliability with a mean interval of 16 days.

Results

The K-ISP2 showed good test-retest reliability (ICC = .84) and significant but weak convergent validity with the K-WBL (amenability/persistence, r = −.28, p < .001; adaptability, r = −.30, p < .001; reactivity, r = .24, p = .004; activity, r = .20, p = .02). There were no significant age and gender differences in the K-ISP2 total score. The K-ISP2 total score was similar to that of the Infant Sensory Profile 2 (ISP2).

Conclusions

This study provides reliability and validity evidence of the K-ISP2. Normative data of K-ISP2 total score can be established regardless of age or gender. The ISP2 has showed similar psychometric properties for populations of both Korea and the United States.

KEY MESSAGES

• Sensory processing is a fundamental component for adaptive responses to environment and begins to develop before birth.

• The Korean Infant Sensory Profile 2 (K-ISP2) can be used to assess sensory processing patterns of Korean infants, with evidence of reliability and validity.

• Total scores of the K-ISP2 can be interpreted with one normative data set regardless of age and gender.

Keywords:

• Psychometric testing
• sensory processing
• Infant Sensory Profile 2
• Korean Infant Sensory Profile 2
• cross-cultural comparison
• temperament

Introduction

A sensory system begins to develop before birth for sensory processing, which is defined as the registration, modulation, and organization of sensory input for adaptive responses to environments [1]. Sensory processing is a unique attribute of individuals and has a neurological basis in their genetic background [2]. Even newborns show individual differences in their reactions to sensory stimuli [3]. Infants who have a sibling with sensory processing disorder are at genetic risk of developing sensory difficulties [4]. Infants have a biological tendency to respond to sensory input, and gene-environment interactions shape their character and development [5].

Sensory processing underlies temperament, which is defined as an inborn individual difference characterized by reactivity and self-regulation of attention, emotions, and motor domains [6–8]. Infants with atypical sensory processing are more likely to have difficult temperaments [9]. For example, if infants have low sensory thresholds, they are easily overwhelmed, as reflected in short attention spans, negative emotions, and withdrawal behaviors [10,11]. Thus, sensory processing is closely associated with socio-emotional development [12]. Sensory processing is also important for motor and cognitive development in infancy. With appropriate sensory processing, infants can give adaptive responses and develop motor function [13,14], which contributes to their cognitive development [15]. Sensory processing problems in infants can negatively affect their motor and cognitive skills [4,14]. Therefore, sensory processing in infants at risk should be evaluated for the early identification of sensory-related problems [16].

Representative assessments of sensory processing in very young infants are the 0–6 months section of the Infant/Toddler Sensory Profile (ITSP) [17,18] and its revised version, the Infant Sensory Profile 2 (ISP2) [19,20]. The ITSP and the ISP2 are based on Dunn’s sensory processing framework [17–21]. The framework suggests two dimensions: neurological thresholds (low versus high) and self-regulation strategy (active versus passive) continua. The intersection between the two continua creates four quadrants: seeking of high threshold and active strategy, avoiding of low threshold and active strategy, sensitivity of low threshold and passive strategy, and registration of high threshold and passive strategy. On the basis of this framework, the ITSP and the ISP2 ask caregivers about their young children’s sensory experiences with activities in daily living.

Assessments can be used appropriately with an understanding of reliability and validity. In the development study of the ISP2 for American infants, reliability evidence including acceptable internal consistency (α = .75), good test-retest reliability (ICC = .86), and the standard error of measurement (4.94) was provided for a total score [19,20]. The development study also reported details of content and construct validity evidence [19,20]. Content validity was established through expert reviews and preliminary pilot studies. Construct validity was supported by a moderated correlation of the ISP2 and the ITSP 0–6 months section. The Vineland Adaptive Behavior Scales, Second Edition (Vineland–II) [22,23] was measured to examine the construct validity of the ISP2, but the sample size (n = 18) was not sufficient to derive results.

Applying the ISP2 in other countries requires cross-cultural adaptation to establish equivalence between the original and translated questionnaires [24]. Through a cross-cultural adaptation process that considers the correspondence of semantic, idiomatic, experiential, and conceptual aspects, the adapted version can be comparable with the original version [25]. Studies on the cross-cultural adaptation of the ISP2 were conducted in Iran [26], Brazil [27], and Korea [28,29].

The reliability and validity of an assessment should be re-examined when it is adapted for use in other cultural contexts. The culturally adapted ISP2 reported psychometric properties such as internal consistency and content validity [26–28]. However, there is no evidence for the convergent validity of the ISP2 construct with its similar construct, temperament. Convergent validity provides information about the extent to which a measure relates to other measures containing similar constructs [30]. In addition to convergent validity, age and gender differences in the ISP2 were not reported.

This study aimed to examine the test-retest reliability of Korean Infant Sensory Profile 2 (K-ISP2) [28,29] and its convergent validity with the Korean version of ‘What My Baby Is Like’ (K-WBL) [31,32], a measure of temperament. We predicted that the test-retest reliability of the K-ISP2 would be good, comparable to that of the ISP2 [19,20]. Since sensory processing problems are linked to a difficult temperament [6], we hypothesized that high scores of the K-ISP2 would be correlated with high levels of activity and reactivity subscales and low levels of amenability/persistence and adaptability subscales in the K-WBL. To determine whether one normative data of the K-ISP2 can be used for Korean infants, this study examined differences in the K-ISP2 by age and gender. Given that the ISP2 presented one normative data set [19,27], we hypothesized that the K-ISP2 total scores would not be different according to age and gender. In addition, to investigate whether the sensory processing of infants differs according to culture, this study examined whether there is a difference between the K-ISP2 total score of Korean infants and the ISP2 total score of American infants. With little difference being reported between the American and Brazilian cultures on ISP2 [27], we predicted that the K-ISP2 total scores would be similar to the ISP2 total scores.

Materials and methods

Design

This cross-sectional study was part of the Korean Sensory Profile 2 project. Data were collected between July 2019 and October 2021.

Participants and procedures

Participants were Korean caregivers of infants aged 0–6 months. A convenience sampling method was used to recruit participants from parents’ online clubs and communities in South Korea. The sample size was estimated using G*power program [33]. The caregivers of 140 infants were included in this study. Another ten participants were approached but they refused to complete the questionnaire. Inclusion criteria were (1) primary caregivers of an infant aged 0–6 months and (2) caregivers who were able to read and write Korean.

The caregivers of 13 infants responded to the questionnaires in paper form and the caregivers of 127 infants completed them online. It is assumed that caregivers would respond similarly to paper and online questionnaires given that the ISP2 has been already administered in paper and online forms in the U.S. The study was explained to all participants and their informed consent was obtained. The project was approved by the institutional review board of Soonchunhyang University (IRB No. 201810-SB-037). After filling out the consent form, the participants completed a form of socio-demographic information (e.g. infant’s birth date and gender, respondent, parental age, parental education level) and the K-ISP2 and the K-WBL questionnaires.

Measures

Korean Infant Sensory Profile 2 (K-ISP2)

The K-ISP2 [28,29] is a Korean standardized version of the ISP2 [19,20] that assesses sensory processing patterns in infants aged 0–6 months. This assessment is based on Dunn’s sensory processing framework including seeking, avoiding, sensitivity, and registration. The assessment is divided into six sensory sections (general, auditory, visual, touch, movement, and oral sensory processing). The K-ISP2 asks caregivers to rate the frequency of behavior on 25 items using a Likert scale (5 = Almost Always, 4 = Frequently, 3 = Half the Time, 2 = Occasionally, 1 = Almost Never, and 0 = Does Not Apply). The total score ranges from 0 to 125. The K-ISP2 only provides a total score for interpretation, and a high score indicates that an infant exhibits frequent behavior in the items. The total score is classified based on standard deviation: ‘Much Less Than Others’ (below −2 SD), ‘Less Than Others’ (between −2 SD and −1 SD), ‘Just Like the Majority of Others’ (between −1 SD and +1 SD), ‘More Than Others’ (between +1 SD and +2 SD), and ‘Much More Than Others’ (above + 2 SD). The K-ISP2 has content validity and acceptable internal consistency (Cronbach’s alpha = .78) [28].

We, the authors of K-ISP2, obtained permission for cross-cultural validation of the K-ISP2 from the copyright holder, the Korea Psychological Corporation. The K-ISP2 can be purchased without special qualifications at the Korea Psychological Corporation website (https://www.koreapsy.co.kr/). We were also permitted to translate and study the ISP2 from its author, Dr. Dunn, and the Pearson Psychological Assessment. The ISP2 can be purchased by healthcare or education practitioners at the Pearson Psychological Assessment website (https://www.pearsonassessments.com/).

Korean version of What My Baby Is Like (K-WBL)

The K-WBL [31] is a Korean adapted version of ‘What My Baby Is Like’ (WBL) [32], which is a questionnaire used to measure the temperament of infants aged 1–12 months based on the New York Longitudinal Study [34,35]. The K-WBL has four subscales: amenability/persistence, adaptability, reactivity, and activity. The instrument consists of 18 items rated on a 9-point Likert scale (9 = for extremely, 1 = for not at all) by a caregiver. The mean score for each subscale ranges from 1 to 9, and a higher score indicates a greater degree of temperament. The K-WBL showed the construct validity of four factors and good internal consistency (Cronbach’s alpha = .86) [31].

We obtained permission to use the K-WBL from the Korean author of the K-WBL and from the author of the WBL. The K-WBL and the WBL can be found in research papers for validation [31,32].

Data analysis

The data were analyzed using IBM SPSS Statistics Version 22 (IBM, Armonk, NY). A p-value less than .05 was considered significant in the two-tailed test. The normality of data was verified using the Kolmogorov-Smirnov test.

Test-retest reliability of the K-ISP2 total scores was calculated using the intra-class correlation coefficient (ICC), which was interpreted as excellent (> .90), good (.75–.90), moderate (.50–.75), and poor (< .5) [36, 37]. In terms of the convergent validity of the K-ISP2, the relationship of the K-ISP2 total score with the K-WBL subscales was analyzed using Pearson’s correlation coefficient, which was interpreted as strong (≥ .70), moderate (.40–.69), or weak (< .40) [38, 39]. An independent sample t-test was conducted to examine differences in the K-ISP2 total scores by age (0- to 3-month-olds versus 4- to 6-month-olds) and gender (boys versus girls).

To examine whether there is a difference between the K-ISP2 total score of Korean infants and the ISP2 total score of American infants, we used a report on the U.S. normative data set of the ISP2 [19]. The sample of 68 infants in the U.S. had the following characteristics [19]: sex (female, 34; male, 34), race/ethnicity (White, 43; Hispanic, 12; African American, 7; Asian 2; other 4), region (South, 31; West, 18; Midwest, 13; Northeast, 6), and parental education level (four-year degree or higher, 32; some college or technical school, 25; high school graduate, 10; no high school diploma, 1). The mean and standard deviation of the ISP2 in the U.S. sample were 50.9 and 9.9, respectively. The homogeneity of gender and parental education level between Korean and American samples was verified using Fisher’s exact test. Thereafter, to determine whether the K-ISP2 total score is different from the ISP2 total score, an independent sample t-test was performed by using mean, standard deviation, and the number of each group.

Results

A total of 140 data were collected for convergent validity between the K-ISP2 and K-WBL and for making comparisons by age, gender, and culture. After the first survey, 127 out of 140 participants completed the K-ISP2 again. These K-ISP2 pairs were used for test-retest reliability. The demographic characteristics of participants are presented in Table 1.

Table 1. Participants’ characteristics.

Characteristics	Validity and comparison study (n = 140) Mean (SD)/n (%)	Reliability study (n = 127) Mean (SD)/n (%)
Infant’s age (months)	4.1 (1.6)	3.9 (1.5)
Gender
Male	71 (50.7%)	65 (51.2%)
Female	69 (49.3%)	62 (48.8%)
Relationship with infants
Mother	135 (96.4%)	122 (96.1%)
Father	5 (3.6%)	5 (3.9%)
Parents’ age (years)
Mother’s age	32.6 (3.4)	32.7 (3.3)
Father’s age	34.9 (4.2)	35.1 (4.3)
Educational level of mother
High school	13 (9.3%)	10 (7.9%)
Bachelor’s degree	103 (73.6%)	95 (74.8%)
Master’s degree and above	24 (17.1%)	22 (17.3%)
Educational level of father
High school	14 (10.0%)	12 (9.4%)
Bachelor’s degree	103 (73.6%)	96 (75.6%)
Master’s degree and above	23 (16.4%)	19 (15.0%)

The test-retest reliability of the K-ISP2 total scores was good (ICC = .84). The mean interval between test and retest was 16.1 days (SD = 7.6).

Convergent validity was determined by the significant correlation between the K-ISP2 total scores and the K-WBL subscales. The K-ISP2 was positively correlated with the K-WBL reactivity (r = .24, p = .004) and activity (r = .20, p =.02), whereas the K-ISP2 was negatively correlated with the K-WBL amenability/persistence (r = −.28, p < .001) and adaptability (r = −.30, p < .001). Further correlation analysis was performed excluding infants aged 0–1 months who were not the target of the K-WBL. For 136 infants aged 1–6 months, the K-ISP2 was positively correlated with the K-WBL reactivity (r = .22, p = .01) and activity (r = .20, p = .02), whereas the K-ISP2 was negatively correlated with the K-WBL amenability/persistence (r = −.33, p < .001) and adaptability (r = −.36, p < .001). These results were similar to those of participants aged 0–6 months.

There was no difference in the K-ISP2 total scores of 0- to 3-month-olds (n = 62, M = 53.0, SD = 13.0) and 4- to 6-month-olds (n = 78, M = 53.6, SD = 8.8; t = 0.26, p = .79). There was no significant difference in the K-ISP2 total scores of boys (n = 71, M = 53.9, SD = 10.5) and girls (n = 69, M = 52.7, SD = 11.1; t = 0.67, p = .51).

Fisher’s exact test showed no differences in gender and parental education level between Korean and American infants (ps > .24), indicating homogeneity between the two samples. The K-ISP2 total scores (n = 140, M = 53.3, SD = 10.8) were not significantly different from the ISP2 total scores of the development study (n = 68, M = 50.9, SD = 9.9; t = 1.54, p = .12) [19].

Discussion

This study showed that the K-ISP2 has adequate test-retest reliability and significant but weak convergent validity to assess sensory processing in Korean infants. The fact that there were no age and gender differences in the K-ISP2 suggests that the K-ISP2 could have one set of normative scores for interpretation. The total score of K-ISP2 was not significantly different from that of ISP2, revealing that the sensory processing of Korean infants and American infants may be similar.

Test-retest reliability of the K-ISP2 was good, indicating that the K-ISP2 stably measured the sensory processing of infants across times. The ICC of the K-ISP2 (.84) is comparable with those of the ISP2 (.86) [19,20] and its Persian version (.88) [26]. These consistent coefficients suggest that test-retest reliability properties of the ISP2 were shared across cross-cultural contexts.

Evidence of convergent validity was presented in the relationship between the K-ISP2 and the K-WBL. The K-ISP2 was positively correlated with reactivity and activity, which means that frequent sensory-related behaviors may be linked to high responsiveness and activity level. In contrast, the K-ISP2 was negatively correlated with adaptability and amenability/persistence. This means that infants with atypical sensory processing showed a low level of rhythmicity, slowness to adapt to change, and less positive mood. Difficult temperaments include high activity, irritability, irregularity, slow adaptation, and intense negative reaction [35,40]. These challenging characteristics were found to be correlated with atypical sensory processing in this study, which is consistent with the previous finding that low thresholds to sensory stimuli are related to difficult temperament [6]. The results of the current study could also be aligned with previous findings that children with sensory processing abnormalities were more likely to exhibit maladaptive behaviors and less likely to have adaptive behaviors [19,41].

Not much evidence for convergent validity for the ISP2 has been reported in previous research [20]. The study on the development of the ISP2 showed the relationship of the ISP2 with its first edition, the ITSP, and the Vineland–II [19,20]. The correlation between the ISP2 and the ITSP 0–6 months section was significant in thirty infants, suggesting that a similar construct is shared between the two tests. The ISP2 was also correlated with the Vineland–II motor domain, but this result was inconclusive due to a small sample size of eighteen. The current study included a larger sample to present convergent validity. Furthermore, this study provides an understanding of the link between sensory processing and temperament, which has been rarely studied in very young infants.

Differences in the K-ISP2 total score by age and gender were not found. Until now, there has been no report on the ISP2 comparing boys with girls or younger infants with older ones. Given that the ISP2 has one normative data set for the total scores [19,27], we could speculate that overall sensory processing measured by the ISP2 might have been consistent across gender and age. Consistency in the K-ISP2 total scores on age and gender suggests that the K-ISP2 can have one normative data set for interpreting the test result.

In this study, the K-ISP2 was compared with the ISP2 to determine whether the sensory processing of Korean infants was different from that of American infants. There was no significant difference in the total score between the K-ISP2 and the ISP2. This result is in accordance with the finding of research on Brazilian and American infants [27]. A range of ‘Just Like the Majority of Others’ of the Brazilian version of the ISP2 (41-62) was almost the same as that of the ISP2 (41-61). Taken together, the overall sensory processing in infancy seems to be consistent across different cultural contexts. These findings also imply that the ISP2 construct can measure infants’ sensory processing in diverse cultures.

There are limitations in this study and we propose future research for addressing them. First, this study adopted the K-WBL to measure temperament, because it has been used in previous research among very young Korean infants [42,43]. Further studies are expected to provide evidence for the additional validity of the K-ISP2 using the Early Infancy Temperament Questionnaire [44] or the Infant Behavior Questionnaire-Revised [45], which are more internationally used. Second, since the K-ISP2 is a questionnaire, its result reflects caregivers’ perspectives on their infants. The relationship between the K-ISP2 and behavioral/performance measures of sensory processing could be investigated in future researches. Third, the sample size of this study was not large enough to determine constructs by conducting factor analyses. Large samples in future studies may provide information about factor structure of the K-ISP2. Fourth, because we used values of the ISP2 derived from a previous study, there were limitations in analyzing the cultural factors with various variables. To expand the present finding that sensory processing of infants is culturally consistent, further research is needed to investigate cultural aspects, including socio-demographic details and lifestyle. Last, the results of this study were based on a general infant population. Infants at risk, such as those who are born early and have a sibling with autism spectrum disorder, are more likely to have sensory processing difficulties [46]. Studies on the discriminative validity of the K-ISP2 are needed to support the early identification of atypical sensory processing.

This study has the following clinical implications. The K-ISP2 could evaluate sensory processing in Korean infants with the evidence of test-retest reliability and convergent validity. The K-ISP2, which measures cross-cultural aspects of sensory processing, may have one normative data set across age and gender.

Conclusions

The present study provides further evidence for the reliability and validity of the K-ISP2. These findings demonstrate that the K-ISP2 can measure sensory processing in Korean infants consistently and appropriately. The fact that there were no age and gender differences in the K-ISP2 indicates that the one normative data set of the K-ISP2 can be applied to interpret the test results. The similarity in the total scores of the K-ISP2 and the ISP2 suggests that the ISP2 construct measures common sensory processing aspects across different cultural contexts.

Author contributions

SK, YJ, MK, SJ, and EYK conceived and designed the study. SK conducted the surveys and collected data. SK and EYK analyzed and interpreted the data. SK, YJ, and EYK wrote the first draft of this manuscript. All authors contributed to manuscript revision and approved the submitted version.

Acknowledgements

We would like to thank parents for participating in this study. We are also grateful to the members of the Lifespan Occupational Development Laboratory of Soonchunhyang University for their assistance.

Disclosure statement

All authors receive royalties on sales of the Korean Infant Sensory Profile–2. This research was supported by the Korea Psychology Corporation.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Additional information

Funding

This work was supported by the Soonchunhyang University Research Fund and the Korea Psychology Corporation.