Abstract
Keywords
The timing and attainment of developmental milestones, such as walking, talking, and solving problems, are important markers of neurological integrity in childhood (Kelly, Sacker, Schoon, & Nazroo, 2006). The lack of attainment of these developmental milestones, at the requisite time and in the expected order, is used by parents and health care professionals to identify developmental delay (Pachter & Dworkin, 1997; Roze et al., 2010). Although most children around the world reach these milestones at some point, much research has assessed whether children from different cultures develop in a similar manner (Golos, Sarid, Weill, Yochman, & Weintraub, 2011; Venetsanou & Kambas, 2010). A cross-cultural sequence of development appears to exist with regard to both sequence and timing (Geisinger, 1994); however, within this broad framework, cultural differences in various aspects of child development seem to exist (Lansdown et al., 1996).
To understand a child’s development, practitioners use a standardized developmental assessment, administering a set of tasks suitable to the child’s developmental phase and drawing conclusions about the child’s development on the basis of task performance, as compared with that of peers (Meisels, 1989). These tools aim to quantify individual behavior by comparison with norms, yet because of the extreme variability characteristic of pediatric development (Hadders-Algra, 2010), this process naturally invites an error of measurement. One important culprit of this error might be specific cultural nuances, particularly when an instrument is used in a culture other than the one for which it was developed (Geisinger, 1994).
Practitioners using standardized assessment tools must bear in mind that assessment tools are specific to the culture for which they are developed (Lansdown et al., 1996). Despite this, the World Health Organization has recommended cross-cultural sharing and translation of existing instruments because, in addition to being cheaper and faster, this process can facilitate international collaboration, information exchange, and cross-cultural comparison (Prado, Magalhães, & Wilson, 2009). Moreover, translated instruments can benefit the performance of cross-national, cross-language, and cross-ethnic comparative studies (Hambleton & Kanjee, 1995). To this end, it is essential to ensure that instruments developed in one culture are valid and reliable when translated versions are used with a culture other than that for which they were developed (Geisinger, 1994).
In recent years, many changes have occurred within the knowledge base and paradigms of the occupational therapy profession, including (among others) the Occupational Therapy Practice Framework (American Occupational Therapy Association, 2008) and the International Classification of Functioning, Disability and Health (ICF; World Health Organization, 2001). Within health contexts, the ICF has resulted in a shift from a medical model to a biopsychosocial model, fostering the understanding that true health lies in people’s optimal participation in activities important to them, within a range of environments. Optimal participation may be facilitated or hindered by supports and barriers resulting from the interaction between intrinsic and extrinsic components.
These developments within the profession’s knowledge base have led to changes in treatment emphases requiring that early intervention focus on functional abilities and participation, with the practitioner looking past motor skills while considering the child’s range of contexts. However, a review of the literature reveals a lack of standardized assessment instruments linking client factors and performance skills with the child’s ability to participate in various areas of occupation and in different contexts.
The Miller Function and Participation Scales (M–FUN) was developed in line with current professional paradigms (Miller, 2006). Published in the United States in 2006, the M–FUN assesses how a child’s motor competency affects his or her ability to participate in various life contexts and enables the practitioner to analyze the links between the child’s performance skills and participation (Miller, 2006).
The basis for this study was the importance of assessing whether this U.S.-developed instrument is suitable for use in Israel; although a number of instruments that separately assess motor function, participation, or both are used in Israel, few incorporate a comprehensive link between the two in a manner similar to the M–FUN. In addition, many internationally developed developmental assessments commonly used in pediatric practice, such as the Developmental Test of Visual Perception, second edition (DTVP–2; Hammill, Pearson & Voress, 1993) or the Miller Assessment for Preschoolers (Miller, 1982), are either outdated or have not been adequately assessed as to their suitability for use in Israel. Moreover, earlier studies assessing similar questions have shown differences between Israeli and U.S. populations on developmental measures (Schneider, Parush, Katz & Miller, 1995). As such, the purposes of this study were to ensure an accurate translation and adaptation of the M–FUN to the target population as well as to assess Israeli children using the M–FUN to ascertain the suitability of the U.S. norms with this population.
Method
Description of the Instrument
The M–FUN is a standardized, norm-referenced assessment of functional motor skills and a criterion-referenced assessment of participation of children between ages 2 yr 6 mo (2.6) and 7 yr 11 mo (7.11). The functional motor assessment includes two versions, one for children ages 2.6–3.11 and one for children ages 4.0–7.11. Although the tasks differ slightly between versions, the administration of the assessment follows the same protocol. Each child performs a series of visual–motor tasks in an age-appropriate workbook (2.6–3.11 or 4.0–7.11), then completes fine motor tasks while sitting at a table and gross motor tasks in a suitably sized room.
Three raw scores are calculated by summing up the scores for the tasks in each section (generating a visual–motor [VM], fine motor [FM], and gross motor [GM] raw score) that are each compared with a norm-referenced scaled score provided in the assessment manual (mean [M] = 10, standard deviation [SD] = 3) and divided across eight age bands (2.6–2.11, 3.0–3.5, 3.6–3.11, 4.0–4.5, 4.6–4.11, 5.0–5.11, 6.0–6.11, 7.0–7.11). In this manner, three separate scaled scores (VM, FM, and GM) are generated. Higher raw and scaled scores are both indicative of greater competency. No total score of any kind is generated for the M–FUN, enabling each subtest to be administered separately.
For the participation component, three observation questionnaires (parent or caregiver, kindergarten or classroom teacher, examiner) are completed, with the same questionnaires completed for all age bands. Respondents are asked to rate the child’s performance on a number of statements using a 4-point Likert scale, and the total score is summed, with higher scores indicative of better performance. For each questionnaire, the total score is compared with a criterion based on the child’s age; the home questionnaire has five age bands (2.6–2.11, 3.0–3.11, 4.0–4.11, 5.0–6.11, 7.0–7.11), the classroom questionnaire has three age bands (2.6–2.11, 3.0–4.11, 5.0–7.11), and the test questionnaire has four age bands (2.6–2.11, 3.0–3.11, 4.0–5.11, 6.0–7.11).
Evidence of sensitivity (positive predictive power ranging from 0.38 to 0.89) and specificity (negative predictive power ranging from 0.80 to 1.00) for the motor scores has been shown, and the M–FUN has been found to be correlated with other instruments assessing similar constructs (Pearson correlation coefficients [rs] between the M–FUN motor scores and the Miller Assessment for Preschoolers total score range from .47 to .62). Internal consistency (coefficient α ranging from .67 to .93), test–retest reliability (Pearson rs ranging from .77 to .82), and interrater reliability (Pearson rs ranging from .91 to .93) have been shown. Since its publication, further research assessing individual components of the test has also revealed adequate psychometric results; Diemand (2009) provided additional evidence of the criterion validity of the M–FUN VM score (Pearson r = .87) when compared with the DTVP–2.
Development of the Hebrew M–FUN
All components of the instrument (score sheets, administration instructions, workbooks, and participation questionnaires) underwent a thorough process of translation and back-translation (Beaton, Bombardier, Guillemin, & Ferraz, 2000; Van de Vijver & Hambleton, 1996). Discrepancies in translations were discussed by a group of four occupational therapists with extensive pediatric experience, and alternatives for items or sentences that were not culturally appropriate were selected (e.g., Item 38 of the class checklist, “carries tray in cafeteria,” became “carries items that require two hands to hold [large boxes, a tray with numerous items]”). Latin letter-writing tasks were adapted to ensure that the Hebrew tasks included the same number of letters and words as the English version. Although some characteristics of handwriting appear to be universal, specific features of the Hebrew handwriting process (Yochman & Parush, 1998) were considered. Moreover, tasks incorporating writing of individual letters ensured that the letters maintained characteristics similar to those of the original stimulus (e.g., directionality, type of line [curvy or straight], and number of lines; Figure 1). Finally, contact was made with the instrument’s author to ensure that the suggested adaptations were acceptable.
Adaptation from English letters to Hebrew letters with an attempt to retain the characteristics of the original English letters.
Study Design and Phases
This study had two phases. During Phase 1, we investigated the initial construct validity of the M–FUN when used in Israel in a pilot study that also aimed to assess differences between an Israeli sample and the U.S. standardization sample. Although the results of the pilot study (described next) were promising, we identified a need for further, in-depth investigation of the M–FUN with a larger sample, incorporating all age bands of the two versions of the instrument (Phase 2). Ethical approval to perform both phases of the study was granted through the ethics committee of the School of Occupational Therapy, Hebrew University of Jerusalem, and informed participant consent was obtained at every stage.
Phase 1: Pilot Study
Participants.
Thirty typically developing Israeli children (M age = 66.33 months, SD = 11.02) from three age groups (4.0–4.11, 5.0–5.11, 6.0–6.11) encompassing four norm-referenced M–FUN age bands (4.0–4.5, 4.6–4.11, 5.0–5.11, 6.0–6.11) were recruited by means of convenience sampling. Each age group included 10 children; the 4-yr-old group included 7 boys and 3 girls (M age = 53.80, SD = 3.52), the 5-yr-old group included 4 boys and 6 girls (M age = 66.30, SD = 3.65), and the 6-yr-old group included 4 boys and 6 girls (M age = 78.90, SD = 3.93). The sample was recruited in and around the Jerusalem area, based on defined inclusion criteria (fluency in Hebrew of both parent and child; no history of developmental delay) and exclusion criteria (referral or treatment of developmental assessment; treatment at the time of testing, in the past, or both).
Each Israeli child was age and gender matched (within a 2-mo radius) to a child from the U.S. standardization sample (U.S. group; n = 30; M age = 66.30 mo, SD = 10.94), whose scaled motor and raw participation scores were generously provided by the developers of the M–FUN. Ten children were included from each age group of the U.S. sample; the 4-yr-old group included 7 boys and 3 girls (M age = 53.90, SD = 3.35), the 5-yr-old group included 4 boys and 6 girls (M age = 66.10, SD = 3.48), and the 6-yr-old group included 4 boys and 6 girls (M age = 78.90, SD = 3.93).
Instruments.
Miller Function and Participation Scales—Hebrew Version
The VM, FM, and GM tests of the M–FUN were administered, and the three participation checklists (home, class, and test) were completed. Three raw and scaled motor scores and three total raw and criterion-referenced participation scores were calculated (based on U.S. norms) for each child.
Intake
We developed a brief questionnaire to gather demographic information for the purposes of this study (including age, gender, and the existence of exclusion criteria).
Beery-Buktenica Developmental Test of Visual–Motor Integration, Fifth Edition
The Beery-Buktenica Developmental Test of Visual–Motor Integration, Fifth Edition (Beery & Beery, 2004), is a standardized assessment of visual–motor skills used with children aged 2 to 18. The complete test battery was administered (the test of visual–motor integration [VMI], the test of visual perception [VP], and the test of motor coordination [MC], which requires the child to complete a series of fine motor control tasks). The psychometric properties of the Beery include evidence of reliability (overall coefficient α = .82, test–retest Pearson rs ranging from .63 [7-mo period] to .92 [2-wk period], and interrater Pearson rs ranging from .92 to .98) and validity (concurrent validity Pearson rs with the DTVP–2 ranging from .62 to .75 and construct-validity Pearson rs with age ranging from .84 to .89). Although the norms of the Beery were developed in the United States, they are “now commonly considered to be international norms” (Beery & Beery, 2004, p. 115).
Procedure.
All children were invited for an approximately 1-hr-long occupational therapy testing session at which parents signed an informed consent form for their child’s participation in the study. The Hebrew M–FUN and the three Beery subtests were administered by one of three occupational therapists who had undergone extensive training in the administration of the assessments. During this time, parents completed the intake questionnaire and the home checklist questionnaire of the M–FUN. Parents were given the class checklist questionnaire and a stamped, self-addressed envelope. They were requested to ask the child’s teacher to complete and return the questionnaire directly to the researchers. Once the Israeli sample had been recruited, the ages and genders of all participants were sent to the developers of the M–FUN and age- and gender-matched motor scaled scores and participation raw scores of children from the standardization sample were supplied for the purpose of direct comparison between groups.
Data Analyses.
We used SPSS for Windows (Version 19; IBM, Armonk, NY) for all analyses. Statistical procedures of the pilot study included calculation of Pearson rs to investigate relationships between variables, independent-sample t tests to assess group differences, and paired-sample t tests to compare age- and gender-matched U.S. and Israeli samples. For effect sizes, we considered d values of 0.2 to be small; 0.5, medium; and 0.8, large (Coe, 2002).
Phase 2
Although the pilot study results were encouraging, the sample was relatively small, only three of the older version M–FUN age bands were investigated, and the study was limited to children from one geographical area. Moreover, significant differences between the U.S. and Israeli samples for the VM score and the lack of correlation between age and the participation scores warranted further investigation.
Participants.
On the basis of the scaled motor scores of the M–FUN collected from the 60 children in the pilot study (30 Israeli children and 30 age- and gender-matched children from the U.S. sample), we conducted a power analysis with α = .05 (one-sided) and 80% power. On the basis of the VM score, the sample size required was 131; the FM score, 88; and the GM score, 45. Because the pilot study was based on four of the instrument’s eight age bands, the maximum size for the entire sample required was doubled, and we aimed to recruit 260 children over the eight normative motor age bands. The final sample included 267 Israeli children between the ages of 2.6 and 7.11 for all eight normative motor age bands (Table 1), who were recruited from throughout Israel by means of convenience sampling. The inclusion and exclusion criteria were the same as those of the pilot study in Phase 1.
Gender and Mean Ages of Children in Motor Norm- and Participation Criterion-Referenced Age Bands
Note. M = mean; SD = standard deviation.
Instruments.
The Hebrew M–FUN and intake questionnaire (as described in the Phase 1 section) were used for the second phase of the study.
Procedure.
Pediatric occupational therapists from two major health service providers with services throughout Israel underwent extended training in the administration and scoring of the M–FUN. During their training, the therapists watched videos of children completing the M–FUN tasks and were required to understand the scoring criteria and score the children accurately in accordance with them. Once their competence was ensured, these therapists assessed 267 typically developing children from throughout Israel, recruited through a convenience sample. Children were invited for a single testing session lasting up to 1 hr: The parents signed an informed consent form and completed the intake, and all components of the M–FUN were administered. Class checklist questionnaires were obtained in the same manner as in Phase 1.
Data Analyses.
We used SPSS for Windows (Version 19) for all analyses. We present percentages for the purposes of descriptive statistics. Pearson rs were calculated to assess relationships between variables. We used multivariate analyses of variance (MANOVAs), analyses of variance (ANOVAs; Scheffé post hoc analyses [Maxwell & Delaney, 2004]), and independent-sample t tests to assess between-groups differences. Comparisons with U.S. norms were based on data provided in the M–FUN examiner’s manual (Miller, 2006). For effect sizes, we considered d values of 0.2 to be small; 0.5, medium; and 0.8, large (Coe, 2002); we considered η2 values of .01 to be small; .06, medium; and .14, large (Cohen, 1988). Because two versions of the M–FUN assess motor function, we performed separate analyses for the younger and older children using raw scores and analyzed the total sample using scaled scores. For participation scores, we analyzed the entire sample together using total scores.
Results
Phase 1
Analysis of Israeli Sample.
Initially, we performed analyses using data from the 30 Israeli children. Because the M–FUN is a developmental measure, with performance expected to improve as the child ages, construct validity would be evidenced by correlations between the child’s scores and age. Pearson rs revealed strong, significant correlations between the VM (r = .85, p < .001), FM (r = .66, p < .001), and GM (r = .77, p < .001) raw scores and the child’s age. However, for the whole sample, Pearson rs between the child’s age and the three total scores of the participation checklists revealed no significant correlations (home, r = .30, p = .11; class, r = .23, p = .26 [n = 26]; test, r = .18, p = .34).
To assess criterion validity, we correlated the three raw motor scores of the M–FUN with the three raw motor scores of the Beery, revealing significant medium to high correlations between all scores (VM—VMI, r = .83, p < .001; VM–VP, r = .62, p < .001; VM–MC, r = .72, p < .001; FM–VMI, r = .59, p < .01; FM–VP, r = .54, p < .01; FM–MC, r = .50, p < .01; GM–VMI, r = .62, p < .001; GM–VP, r = .44, p = .01; GM–MC, r = .43, p = .02).
Comparison Between Israeli and U.S. Samples.
Because exact age matching was not possible for some of the children, we performed an independent-sample t test before comparing the age- and gender-matched Israeli and U.S. samples to assess age differences between the U.S. (M = 66.30 mo, SD = 10.94) and Israeli (M = 66.33 mo, SD = 11.02) groups; no age group differences were found, t(58) = −0.01, p = .99, d < 0.01. Thereafter, we performed paired-sample t tests for each of the motor scaled scores and participation raw scores; a significant difference was found only for the VM score, with the Israeli sample (M =11.37, SD = 1.83) showing a higher mean score than the U.S. sample (M =9.83, SD = 2.53), t(29) = 2.51, p = .02, d = 0.70. No differences were found for any other score.
Phase 2
Descriptive Statistics.
We analyzed parental education, a traditional socioeconomic status (SES) index (Raviv, Kessenich, & Morrison, 2004), as an indicator of SES (maternal education: 10.3% had 12 yr of schooling, 29.9% had 13–15 yr, and 59.8% had ≥16 yr; paternal education: 2.9% had <12 yr of schooling, 19.6% had 12 yr, 33.3% had 13–15 yr, and 44.2% had ≥16 yr). Because a higher than expected number of parents had completed further education, we performed analyses of the correlations between the number of years of maternal and paternal education and the VM, FM, and GM raw scores as well as the three participation scores. Results revealed no significant correlations with the exception of a weak correlation between maternal education and the test checklist score (r = .18, p = .01).
Regarding the distribution of children from different geographical areas of Israel, 39.0% of the sample were from the center of the country, 55.1% were from Jerusalem and the south, and the remaining 5.9% were from the north. Regarding the type of area in which the children lived, 60.2% lived in cities, 7.2% lived in small towns, 25.9% lived in villages or community settlements, and 6.7% were from a kibbutz (collective community).
Analysis of Israeli Sample Motor Scores.
To analyze construct validity, we correlated all motor raw scores for each M–FUN version with the child’s age in months using Pearson rs. For the younger version (n = 94), we found significant medium correlations for the VM (r = .57, p < .001), FM (r = .40, p < .001), and GM (r = .52, p < .001) raw scores. Likewise, for the older version of the test (n = 173), we found even higher correlations (VM: r = .79, p < .001; FM: r = .68, p < .001; GM: r = .78, p < .001). We also assessed correlations for the individual normative motor age band; we found very few significant correlations. For VM 3.0–3.5, r = .55, p = .001; VM 3.6–3.11, r = .32, p = .04; VM 5.0–5.11, r = .43, p < .01; FM 3.0–3.5, r = .43, p = .01; and GM 7.0–7.11, r = .48, p = .01.
For each M–FUN version, we performed a one-way MANOVA to assess age-group differences for the three motor scores to ascertain whether the raw scores improved with age, as would be expected with a developmental measure. Mean VM, FM, and GM scores by age band are presented in Table 2. For the younger version, we found significant differences in motor scores, with a large effect size, between the three age bands, Wilks’s Λ = 0.70, F(6, 174) = 5.80, p < .001, η2 = .17. One-way ANOVAs for each motor score revealed significant differences between age groups for the VM scores, F(2, 89) = 13.49, p < .001, η2 = .23; FM scores, F(2, 89) = 6.00, p < .01, η2 = .12; and GM scores, F(2, 89) = 14.47, p < .001, η2 = .24, with large effect sizes.
Independent-Sample t Tests Assessing U.S.–Israeli Group Differences for Each Motor Raw Score, by Normative Age Band
Note. df = degree of freedom; FM = Fine Motor; GM = Gross Motor; M = mean; SD = standard deviation; VM = Visual–Motor.
Scheffé post hoc analyses revealed that, for the VM score, the mean score of the 2.6–2.11 group was significantly lower than that of both the 3.0–3.5 (p = .05) and 3.6–3.11 (p < .001) groups (ds = 0.88 and 1.53, respectively). Likewise, the mean score of the 3.0–3.5 group was significantly lower than that of the 3.6–3.11 group (p = .02, d = 0.63). For the FM score, the mean score of the 2.6–2.11 group was significantly lower (p < .01, d = 0.93) than that of the 3.6–3.11 group. No other between-group differences were found. For the GM score, the mean score of the 2.6–2.11 group was significantly lower than that of both the 3.0–3.5 (p < .01, d = 1.25) and 3.6–3.11 (p < .001, d = 1.61) groups. No group differences were found between the 3.0–3.5 and 3.6–3.11 groups.
For the older version, an additional one-way MANOVA revealed significant motor score differences between the five age bands, with a large effect size, Wilks’s Λ = 0.24, F(12, 436.84) = 25.72, p < .001, η2 = .38. One-way ANOVAs showed significant age-group differences with large effect sizes for the VM score, F(4, 167) = 73.59, p < .001, η2 = .64; FM score, F(4, 167) = 36.26, p < .001, η2 = .46; and GM score, F(4, 167) = 74.35, p < .001, η2 = .64. Scheffé post hoc analyses revealed significant between-group differences between all pairs of groups on the VM score (ps < .01 for all comparisons with the exception of the 6.0–6.11 and 7.0–7.11 group, p = .04; ds ranged from 0.97 to 3.76). Likewise, for the FM score, we found significant between-groups differences with p < .01 (ds ranging from 0.83 to 2.75) for all pairs with the exception of 4.0–4.5 and 4.6–4.11 group (ns), 4.6–4.11 and 5.0–5.11 groups (p = .03, d = 0.77), and 6.0–6.11 and 7.0–7.11 groups (ns). For the GM score, we also found significant between-group differences with ps < .01 (ds ranging from 0.85 to 3.43) for all pairs with the exception of the 4.0–4.5 and 4.6–4.11 groups (ns) and the 4.6–4.11 and 5.0–5.11 groups (ns).
We also assessed correlations between the three raw motor scores for each test version. For the younger version, the VM score was highly correlated with both the FM (r = .75, p < .001) and GM (r = .68, p < .001) scores. Likewise, the FM and GM scores were highly correlated (r = .66, p < .001). For the older version, the VM score was highly correlated with the FM score (r = .78, p < .001) and GM score (r = .75, p < .001), and the FM and GM scores were also highly correlated (r = .76, p < .001).
Comparison Between Israeli and U.S. Motor Scores.
At the following stage of motor score assessment, we compared the Israeli scores with the U.S. norms. Initially, we performed independent-sample t tests between the two groups for the whole sample using scaled scores. The mean VM (n = 267; M = 11.29, SD = 2.85), FM (n = 266; M = 9.86, SD = 2.48), and GM (n = 264; M = 10.29, SD = 3.11) scaled scores were compared with the normative mean scaled score of 10 (SD = 3). On the basis of Table 4.13 of the examiner’s manual of the M–FUN (Miller, 2006, p. 100), the total U.S. sample was calculated (VM: n = 413, FM: n = 413, GM: n = 408). When the two groups were compared, we found significant between-group differences between the U.S. and Israeli samples for the VM score, t(678) = 5.58, p < .001, d = 0.44, with the Israeli sample performing better than the U.S. sample. For the FM score, t(677) = 0.63, p = .53, d = 0.05, and GM score, t(670) = 1.21, p = .23, d = 0.09, no group differences were found. Thereafter, we performed independent-sample t tests using the mean raw score for each age group and each motor score based on Table 4.13 of the U.S. manual. As shown in Table 2, significant differences were found predominantly for the VM raw score but not for the other two motor scores with the exception of one FM age band.
Analysis of Israeli Sample Participation Scores.
We performed analyses on the total scores on the participation checklist for the whole sample. Significant Pearson rs were found between the child’s age and the home (r = .59, p < .001), class (r = .34, p < .001), and test (r = .23, p < .001) checklists. Moreover, correlations assessed for each criterion-referenced participation age band revealed significant correlations for the four age bands of the home checklist (2.6–2.11: r = −.07, p = .79; 3.0–3.11: r = .25, p = .04; 4.0–4.11: r = .27, p = .05; 5.0–6.11: r = .26, p = .01) but not for the class and test checklists.
We performed one-way ANOVAs to assess age-group differences in participation scores. Significant differences were found between the five age groups for the home checklist, F(4, 250) = 24.27, p < .001, η2 = .28, with a large effect size. Scheffé post hoc analyses showed that the mean home participation score of the 2.6–2.11 group, 143.44 (SD = 19.74), was not significantly lower than that of the 3.0–3.11 (M = 146.28, SD = 23.41) or 4.0–4.11 (M = 158.09, SD = 20.08) groups, but it did differ significantly from those of the 5.0–6.11 (M = 171.94, SD = 9.59, p < .001, d = 1.45) and 7.0–7.11 (M = 180.24, SD = 16.18, p < .001, d = 2.04) groups. The 3.0–3.11 group differed significantly from the 4.0–4.11 (p = .04, d = 0.54), 5.0–6.11 (p < .001, d = 1.19), and 7.0–7.11 (p < .001, d = 1.69) groups. Finally, the 4.0–4.11 group differed significantly from the 5.0–6.11 (p < .01, d = 0.70) and 7.0–7.11 (p = .001, d = 1.21) groups.
We also found significant differences between the three age groups on the class checklist, F(2, 223) = 14.56, p < .001, η2 = .12, with a medium–large effect size. Scheffé post hoc analyses revealed that the mean class participation score of the 2.6–2.11 group (M = 108.78, SD = 14.23) was not significantly lower than that of the 3.0–4.11 group (M = 114.63, SD = 14.60) but differed significantly from that of the 5.0–7.11 (M = 124.06, SD = 14.93, p < .001, d = 1.05) group. Likewise, the 3.0–4.11 and 5.0–7.11 groups differed significantly (p < .001, d = 0.64).
The four test checklist groups were also found to differ significantly, F(3, 262) = 4.71, p < .01, η2 = .05, with a small–medium effect size. Scheffé post hoc analyses revealed that the only group differences in mean score were between the 3.0–3.11 (M = 64.44, SD = 12.85) and 6.0–7.11 (M = 69.88, SD = 6.59) groups (p < .01, d = 0.53).
We calculated Pearson correlation coefficients between the three raw participation scores, which showed that all participation scores were correlated with each other. The home checklist score was significantly correlated with scores on both the class (r = .45, p < .001) and test checklists (r = .34, p < .001). Likewise, the class and test checklists were significantly correlated (r = .34, p < .001).
Comparison Between Israeli and U.S. Participation Scores.
To further assess the participation scores, we compared the Israeli scores with the U.S. scores presented in the M–FUN manual (Miller, 2006). The total sample size for the home checklist was 325, the total for the class checklist was 183, and the total for the test checklist was 360. Because the mean total scores are provided on the basis of this division, we calculated independent-sample t tests for each participation checklist over the eight age bands; the results are presented in Table 3. Results showed no group differences for the majority of comparisons.
Independent-Sample t Tests Assessing U.S.–Israeli Group Differences for Each Participation Score by Normative Age Band
Note. df = degree of freedom; M = mean; SD = standard deviation.
Significant group differences at p < .05.
Correlations Between Motor and Participation Scores.
Finally, because the M–FUN aims to investigate how a child’s motor competency affects ability to participate in relevant contexts of his or her life, we assessed the extent to which the motor scores were correlated with the participation scores. Pearson rs between each mean motor scaled score and each participation checklist for the whole sample revealed significant (p < .01) yet low correlations between all scores, ranging between .18 and .36.
Discussion
In the current study, we aimed to develop an accurate Hebrew version of the M–FUN for use with Israeli children and to assess the suitability of the U.S. norms in Israel. The findings of the pilot phase lent support to the use of the adapted Hebrew M–FUN, with the correlations between the M–FUN and Beery scores providing evidence of criterion validity, yet identified a need for further, in-depth analyses of the instrument. The second phase of the study revealed correlations between age and motor scores as well as between age and participation scores, providing evidence for the underlying construct of developmental assessment. When comparing FM, GM, and VM scores with U.S. norms, we found no group differences for the FM or GM scores, yet noted differences for the VM score. With one exception, we found no differences when the Israeli participation scores were compared with U.S. criterion-referenced scores.
Although a number of standardized assessments are commonly used among pediatric occupational therapists in Israel to assess visual–motor functioning (including, among others, the DTVP–2 and the Beery) that have been shown to be suitable for use with Israeli children (e.g., the Beery’s norms “are commonly considered to be international norms”; Beery & Beery, 2004, p. 115), not many current, internationally developed assessments of fine and gross motor function have undergone a thorough process to establish the use of their norms in Israel. This study provides evidence for the validity of using the fine and gross motor norms of the M–FUN with Israeli children.
In the first phase, we adapted a Hebrew version of the instrument, following a thorough translation and adaptation process (Beaton et al., 2000; Van de Vijver & Hambleton, 1996). It is important to note that, during this process, the suitability of the ball used for the GM tasks was questioned. The M–FUN uses a baseball-sized ball, and it was argued that Israeli children are not familiar with games involving balls of this size, generating concern that their gross motor scores would be adversely affected. However, we deemed a change in ball size extreme for the pilot phase, a decision that was consequently supported by the findings of no differences between the U.S. and Israeli samples in GM function in the pilot and larger studies. Had the ball size been increased, Israeli children could feasibly have attained scores significantly higher than those of U.S. children and thus negatively affected the suitability of the U.S. norms.
A particular strength of the pilot study was the opportunity to analyze the scores of a sample of U.S. and Israeli age- and gender-matched children. Although this process enabled a more in-depth look at group differences for the motor and participation scores, it was unfortunate that the raw motor scores for the U.S. sample were not available. The comparison of the two groups using scaled scores meant that a level of sensitivity was lost, because each scaled score of the M–FUN incorporates multiple raw scores of the M–FUN. It is possible that a comparison using raw scores would have picked up on slighter nuances with regard to group differences. In the pilot study, the findings of group differences between the matched U.S. and Israeli samples for the VM score provided additional rationale for further assessment of the M–FUN with larger samples.
Findings regarding the VM score within the larger validation study revealed similar results. Although the U.S. FM and GM norms were found to be suitable for use with Israeli children, Israeli children were found to have higher VM scores than the U.S. norms. This finding might not be surprising in light of studies that have shown that different cultures have distinctive patterns and concepts as to what behaviors and skills to encourage in children (Katz, Kizony, & Parush, 2002). Visual perception might be an area of particular strength among Israeli children. Israeli children enter school at a relatively young age (Josman, Abdallah, & Engel-Yeger, 2006). Because the processes to enhance school readiness (Kramer & Hinojosa, 1999) at this young age appear to be based in educational practices that emphasize alphabetic skills such as letter recognition (Aram & Biron, 2004), it is possible that the visual–motor and perceptual skills of Israeli children are more developed than those of children in other countries. In any event, our findings should caution practitioners with regard to use of the VM norms of the M–FUN with Israeli children.
Construct validity was evidenced by the findings of significant positive correlations between age and motor scores when assessing the whole sample, as well as by significant age-group differences in motor scores. The lack of correlations with age within each of the eight age bands should not be cause for concern: A finding of age correlations within the separate normative bands would potentially raise doubts as to the reasoning for grouping these children together. For example, if children aged 4.0 were to score so differently from children aged 4.5 that an age correlation was noted within the age band, concern would be raised as to the validity of grouping these children within the same normative age bands.
Regarding the participation questionnaire scores, the study findings are encouraging. Participation is a particularly challenging construct to assess, and it might be assumed to be extremely culturally specific. Hence, when attempting to assess it using a standardized measure, it is particularly important to pay attention to cultural nuances. The finding that the U.S. criterion-referenced scores are suitable for use with the Israeli population seems to suggest that the questionnaires are general enough to overcome any overtly specific cultural factors of participation. Furthermore, although the pilot study did not reveal any age correlations for the total sample, the larger study indeed picked up on such correlations, which is encouraging because the assumption of the criterion-referenced scores is that older children should attain higher scores than younger ones.
Limitations and Future Research
This study had several limitations. Foremost, the children were recruited via convenience sampling, resulting in a somewhat skewed geographical representation of children, although it is noteworthy that a range of urban and rural lifestyles were incorporated. Moreover, many of the significant correlations found were in the low range; although this fact is important to note, recall that even low correlations, when consistent, lend support for the existence of underlying phenomena. Because the sample recruited was not even over the eight normative age bands of the test, some of the between-group comparisons should be considered with caution. Additionally, although we found no correlations between parental education (used as a measure of SES; Raviv et al., 2004) and the motor and participation scores, further research should investigate the use of the Hebrew M–FUN with a larger number of children from lower SES groups and with expanded samples in each of the individual age bands. Future research should also aim to establish the criterion validity of the Hebrew M–FUN, as well as assess its use with specific clinical populations. Moreover, the Hebrew M–FUN’s sensitivity and specificity should be investigated.
Conclusions
The study findings are encouraging for Israeli clinicians, despite the need to use the VM U.S. norms with caution because Israeli children were found to score higher than U.S. children on this measure. Because scoring is independent for each test component (Miller, 2006), Israeli clinicians can incorporate the M–FUN FM and GM sections into their battery of valid assessments. Likewise, findings regarding the participation questionnaires support their clinical use in Israel. Validated participation questionnaires means that this important factor of health (World Health Organization, 2001) can be adequately addressed in the assessment of young children. Moreover, the correlations between all motor and participation scores, as well as the test structure, appear suited to current Israeli clinical practice, enabling a useful assessment of the interaction between the child’s motor function and participation in a range of contexts.
Implications for Occupational Therapy Practice
The M–FUN has the potential to provide an important contribution to clinical occupational therapy practice in Israel.
Few internationally developed assessments of fine and gross motor function have been adequately assessed for use in Israel in a manner similar to that of the M–FUN.
When used with Israeli children, the U.S. FM and GM norm-referenced scores, as well as the criterion-referenced home, class and test participation scores, appear to be valid.
Practitioners should use caution when using the U.S. norm-referenced VM scores with Israeli children.
Footnotes
Acknowledgments
We thank Lucy Miller for her guidance and insight as well as Pearson Assessment for providing U.S. standardization sample data. We especially thank Esther Poupko, Batel Zmora, the occupational therapists of the Jerusalem district Pediatric Occupational Therapy Department of Clalit Health Services, and the pediatric occupational therapists of Maccabi Healthcare Services for their assistance in data gathering. We further thank Dennis Bernstein and Miri Barhak for their assistance and all the parents and children who participated.