Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/50303034

MeasuringBehavioralRegulationinFourSocieties

ArticleinPsychologicalAssessment·March2011

DOI:10.1037/a0021768·Source:PubMed

CITATIONS

52

READS

59

12authors,including:

Someoftheauthorsofthispublicationarealsoworkingontheserelatedprojects:

RedLight,PurpleLight!DevelopingaSelf-RegulationInterventionforChildrenatRiskforSchool

DifficultyViewproject

Race-RelatedTeachingPracticesinEarlyChildhoodEducationViewproject

MeganMMcclelland

OregonStateUniversity

62PUBLICATIONS2,559CITATIONS

SEEPROFILE

AlanAcock

OregonStateUniversity

157PUBLICATIONS4,655CITATIONS

SEEPROFILE

XuezhaoLan

6PUBLICATIONS264CITATIONS

SEEPROFILE

FrederickJMorrison

UniversityofMichigan

116PUBLICATIONS5,912CITATIONS

SEEPROFILE

AllcontentfollowingthispagewasuploadedbyMeganMMcclellandon30November2016.

Theuserhasrequestedenhancementofthedownloadedfile.Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,lettingyouaccessandreadthemimmediately.

Measuring Behavioral Regulation in Four Societies

Shannon B. WanlessUniversity of Virginia

Megan M. McClelland and Alan C. AcockOregon State University

Claire C. PonitzUniversity of Virginia

Seung-Hee SonPurdue University

Xuezhao LanHarvard University

Frederick J. MorrisonUniversity of Michigan

Jo-Lin Chen and Fu-Mei ChenFu Jen Catholic University

Kangyi LeeSeoul National University

Miyoung SungSeokyeong University

Su LiChinese Academy of Sciences

The present study examined the psychometric properties of scores from a direct measure of behavioralregulation, the Head–Toes–Knees–Shoulders task (HTKS) with 3- to 6-year-old children in the UnitedStates, Taiwan, South Korea, and China. Specifically, we investigated (a) the nature and variability ofHTKS scores, including relations to teacher-rated classroom behavioral regulation; and (b) relationsbetween the HTKS and early mathematics, vocabulary, and literacy skills. Higher HTKS scores weresignificantly related to higher teacher ratings of classroom behavioral regulation in the United States andSouth Korea but not in Taiwan and China. Also, higher HTKS scores were significantly related to higherearly mathematics, vocabulary, and literacy skills beyond the influence of demographic variables andteacher-rated classroom behavioral regulation. These initial findings suggest that HTKS scores may beinterpreted as reflecting early behavioral regulation in these 4 societies and that behavioral regulation isimportant for early academic success in the United States and in Asian countries.

Keywords: behavioral regulation, academic achievement, measurement, Asia, preschool

The United States repeatedly scores below average on interna-tional academic assessments, whereas societies including Taiwan,South Korea, and China consistently score at or near the top(Baldi, Jin, Skemer, Green, & Herget, 2007; Mullis, Martin, Gon-zalez, & Chrostowski, 2004; Stevenson et al., 1990). In an effort tobetter understand how to improve U.S. children’s academic skills,many researchers have examined factors predicting achievement in

high-achieving societies (Huntsinger, Jose, Liaw, & Ching, 1997;Stevenson, Chen, & Lee, 1993). These studies, however, have notmeasured behavioral regulation (the integration of attention, work-ing memory, and inhibitory control), which significantly predictsacademic achievement in the United States throughout preschooland elementary school (Blair & Razza, 2007; McClelland, Acock,& Morrison, 2006; McClelland et al., 2007; Ponitz, McClelland,

This article was published Online First March 7, 2011.Shannon B. Wanless and Claire C. Ponitz, Curry School of Education,

University of Virginia; Megan M. McClelland and Alan C. Acock, De-partment of Human Development and Family Sciences, Oregon StateUniversity; Seung-Hee Son, Department of Child Development and FamilyStudies, Purdue University; Xuezhao Lan, Graduate School of Education,Harvard University; Frederick J. Morrison, Combined Program in Educa-tion and Psychology, University of Michigan; Jo-Lin Chen and Fu-MeiChen, Department of Child and Family Studies, Fu Jen Catholic Univer-sity, Sinjhuang City, Taiwan; Kangyi Lee, Department of Child Develop-ment and Family Studies, Seoul National University, Seoul, South Korea;Miyoung Sung, Department of Child Studies, Seokyeong University,Seoul, South Korea; and Su Li, Institute of Psychology, Chinese Academyof Sciences, Beijing, China.

We thank all of the participants, collaborators, and research assistants fortheir contributions. The Taiwan Social Skill Development Study was sup-ported by the U.S. Department of State Fulbright Student Scholarship; theRyoichi Sasakawa Young Leaders Fellowship Fund; the Oregon Sports Lot-tery; the Kappa Omicron Nu Honor Society; the Department of HumanDevelopment and Family Sciences, Oregon State University; and the Depart-ment of Child and Family Studies, Fu Jen Catholic University. The Universityof Michigan Pathways to Literacy Project was funded by the National Instituteof Child and Human Development and the National Science Foundation underGrants R01 HD27176 and 0111754, respectively.

Correspondence concerning this article should be addressed to ShannonB. Wanless, Center for Advanced Study of Teaching and Learning, Uni-versity of Virginia, 350 Old Ivy Way, Suite 300, Charlottesville, VA22903. E-mail: shannonwanless@gmail.com

Psychological Assessment © 2011 American Psychological Association2011, Vol. 23, No. 2, 364–378 1040-3590/11/$12.00 DOI: 10.1037/a0021768

364

Matthews, & Morrison, 2009). Moreover, few reliable and validmeasures of behavioral regulation have been developed and eval-uated outside of the United States. In the present study, scoresfrom a direct measure of behavioral regulation, the Head–Toes–Knees–Shoulders task (HTKS; Ponitz et al., 2009), were examinedwith children in the United States, Taiwan, South Korea, andChina, to investigate reliability and validity including relations toearly achievement.

Children in Taiwan, South Korea, and China are intently fo-cused on educational success, and their daily schedules are typi-cally filled with extracurricular academic classes and long hours ofstudying to improve their exams scores and class rankings (Bao,2004; Dwyer, 2004; Yi & Wu, 2004). Despite this emphasis onschool success, reliable and valid behavioral assessments foryoung children in these societies are limited. Early childhoodexperts in Asia have recently called for the development of cul-turally appropriate tools with scores that reliably and validlyreflect learning behaviors for these samples (Kim, Lee, Suen, &Lee, 2003; Tsai, McClelland, Pratt, & Squires, 2006). Given theimportance of behavioral regulation for children’s academic suc-cess in the United States, identifying a measure with scores thatreflect behavioral regulation would be especially valuable in so-cieties such as Taiwan, South Korea, and China, where academicsuccess is a heavily emphasized cultural expectation.

Defining Behavioral Regulation

In the present study, we define behavioral regulation, an aspectof self-regulation, as the integration of cognitive processes, includ-ing attention, working memory, and inhibitory control (McClel-land et al., 2007; McClelland & Wanless, 2008). Behavioral reg-ulation is especially relevant in school contexts. For example, achild with strong behavioral regulation can remember and followa classroom rule, such as waiting for his or her turn at the waterfountain, rather than using a more dominant response, such ascutting in line. Attention, working memory, and inhibitory controlindividually and collectively contribute to behavioral regulationand to the school success of young children.

Attention is defined as the ability to focus on a task whileignoring distractions (Rothbart & Posner, 2005; Rueda, Posner, &Rothbart, 2005). Research in the United States suggests that youngchildren need to attend to activities amidst distractions in order tosucceed in vocabulary acquisition and particularly in mathematics(Dixon & Salley, 2007; G. J. Duncan, et al., 2007; Howse, Calkins,Anastopoulos, Keane, & Shelton, 2003; McClelland, 2009). Work-ing memory is the ability to remember and apply informationwhile encountering and processing new stimuli (Gathercole &Pickering, 2000). This skill has been positively related to mathe-matics and language skills (Adams, Bourke, & Willis, 1999; Espyet al., 2004) and is particularly important during the transition toformal schooling (Senn, Espy, & Kaufmann, 2004). Finally, in-hibitory control is the ability to suppress one response in favor ofa more nondominant behavior (Dowsett & Livesey, 2000). Pre-schoolers are often expected to use inhibitory control to followclassroom rules, and this skill has been linked to mathematics andearly literacy outcomes (Blair & Razza, 2007; van der Schoot, etal., 2004). Two aspects of behavioral regulation, attention andinhibitory control, have also been included in the construct ofeffortful control in the temperament literature, which includes

emotional aspects of regulation (Rothbart & Sheese, 2007). Fur-ther, effortful control has been found to relate positively to aca-demic achievement and grade point average (Checa, Rodrıguez-Bailon, & Rueda, 2008; Deater-Deckard, Mullineaux, Petrill, &Thompson, 2009; Eisenberg, Sadovsky, & Spinrad, 2005; Liew,McTigue, Barrois, & Hughes, 2008; Zhou, Main, & Wang, 2010)as well as social development (Kochanska, Murray, & Harlan,2000). We focus on the behavioral aspects of regulation (e.g.,attention, working memory, inhibitory control) because they havebeen found to be most relevant to children’s academic success(Howse et al., 2003; McClelland & Ponitz, 2010; McClelland,Ponitz, Messersmith, & Tominey, 2010).

In the present study, we focus on behavioral regulation as theintegration of attention, working memory, and inhibitory controlbecause children in early learning settings are expected to orches-trate these skills for learning. This may be particularly true in Asiawhere classroom characteristics such as low teacher–child ratiosand frequent teacher lectures place high demands on behavioralregulation (Hsieh, 2004; Kim et al., 2003; Pang & Richey, 2007).In the present study, we examined the HTKS task of behavioralregulation, which integrates attention, working memory, and in-hibitory control in a simple game.

Behavioral Regulation and Early AcademicAchievement

Behavioral regulation predicts early academic success in pre-school and early elementary school concurrently and over time inthe United States and in Asia (Blair & Razza, 2007; Howse et al.,2003; McClelland et al., 2007; McClelland, Morrison, & Holmes,2000; Ponitz et al., 2009). For example, in the United States, onestudy found that gains in behavioral regulation, measured by theHTKS, over the prekindergarten year significantly predicted im-provement in early mathematics, vocabulary, and early literacyskills (McClelland et al., 2007). In Taiwan, 4-year-olds’ behavioralregulation on the HTKS also significantly predicted early mathe-matics and vocabulary skills (Wanless, McClelland, Tominey, &Acock, in press). Higher effortful control, which taps overlappingcomponents of behavioral regulation, was also found to relate tohigher grade point average in China (Zhou et al., 2010). Moreover,in some studies, the relation between behavioral regulation andearly academic achievement has been particularly pronounced forearly mathematics skills (Blair & Razza, 2007; G. J. Duncan, et al.,2007; Ponitz et al., 2009).

In the United States, research conducted with the HTKS tomeasure behavioral regulation has found that higher kindergartenbehavioral regulation in the fall significantly predicted highermathematics, vocabulary, and early literacy scores in the spring(Ponitz et al., 2009). For U.S. first graders, similar relations werepresent for scores on the HTKS predicting reading comprehensionand vocabulary skills (Connor et al., 2010). Taken together, thesestudies suggest that behavioral regulation measured by the HTKSis a unique and significant predictor of early school success, inWestern and perhaps in Eastern societies. This research, however,has not systematically examined the use of the HTKS in multiplesocieties by evaluating participants of similar ages and includingsimilar outcomes.

In addition, when examining links between direct measures ofbehavioral regulation and achievement, it is important to control

365MEASURING BEHAVIORAL REGULATION IN FOUR SOCIETIES

for teachers’ perceptions of children’s classroom behavioral reg-ulation (Matthews, Ponitz, & Morrison, 2009; Ponitz et al., 2009).This is based on prior research suggesting that teachers’ expecta-tions can influence children’s academic outcomes (Rosenthal,2002). For example, in one meta-analysis of studies conductedwith first through seventh grade teachers, those who had very littleprevious experience with children were found to be more easilyinfluenced by proposed expectations, which in turn had an averageeffect of one quarter standard deviation on children’s IQ scoresassessed months later (Raudenbush, 1994). This effect was partic-ularly strong in first and second grades. Because teacher percep-tions can influence children, they were controlled for in the presentstudy.

Considering Cultural Context

One theoretical tenet that guides this study is the belief thatpatterns of development are contextually specific and a skill suchas behavioral regulation must be examined within each society inorder to understand its unique properties and meaning (Cole, 1996;Shweder et al., 1998). Culturally specific parenting and teachingpractices vary across the United States, Taiwan, South Korea, andChina and may lead to children having different levels of behav-ioral regulation. For example, one study of effortful control in theUnited States and China found that although similar subscales ofeffortful control were observed in both cultures, children’s expe-riences may have led to differences found by the researchers, suchas differences by gender and with relations among subscales(Ahadi, Rothbart, & Ye, 1993). Further research in the UnitedStates and China revealed differences in aspects of children’seffortful control as early as infancy (Gartstein et al., 2006). In theTaiwanese culture, parents focus on teaching children to regulatetheir behaviors, especially around older family members (Hsieh,2004). Children are taught to eat quietly, sit still, and not speakwhen older family members are speaking. In contrast, parents inthe United States often encourage children to speak up and par-ticipate in family conversations at the dinner table (Martini, 1996).In China, cultural beliefs that may influence behavioral regulationare reflected in large class sizes and teaching practices where mostlessons involve group activities rather than individual learning(Pang & Richey, 2007). Young Chinese children are expected toregulate themselves and follow the behaviors of the group, evenwhen these behaviors may conflict with more dominant responses.In South Korea, however, a child-centered approach to earlychildhood education has become increasingly popular (McMullenet al., 2005). As these cultural differences suggest, studies ofbehavioral regulation conducted in the United States may lead usto new research questions about other societies, but they may notgeneralize to other groups. Thus, in the present study, we wereespecially interested in analyses within each society that addressedhow the HTKS functioned as an assessment of behavioral regula-tion.

Measuring Behavioral Regulation

Research has traditionally measured behavioral regulation withteacher or parent ratings, as an aggregated score of individualattention, working memory, and inhibitory control tasks, or withdirect measures (Bronson, Tivnan, & Seppannen, 1995; Carlson,

2005; Howse et al., 2003; Rothbart, Ahadi, Hersey, & Fisher,2001). Although these methods have been useful for understandingperceptions of children’s behavior, identifying individual compo-nents of behavioral regulation, and assessing specific populationsof children, they also have limitations. For example, teacher rat-ings have often differed from scores on other assessments includ-ing child and parent ratings and direct assessments of behavioralregulation (Kunter & Baumert, 2006; Loo & Rapport, 1998; Ma-hone et al., 2002; Mahone & Hoffman, 2007; Wall & Paradise,1981). In particular, teacher ratings in Asia have not alwayscorresponded to directly measured child behaviors (Jose, Hunt-singer, Huntsinger, & Liaw, 2000; Wanless et al., in press). Inaddition, parent ratings provide useful information but may also bebiased (Rothbart et al., 2001). Further, teacher-rated assessmentsoften use Likert scales, which may have limitations for cross-cultural comparisons because they rely on culturally based teacherexpectations for children’s behaviors (Heine, Lehman, Peng, &Greenholtz, 2002). As a result, direct assessments offer a uniqueperspective on child behaviors that may not be captured by teacherratings.

Aggregating scores from individual assessments of attention,working memory, and inhibitory control can also be problematic.Researchers in the United States and Asia have found relativelyweak relations among the different tasks measuring attention,working memory, and inhibitory control (Archibald & Kerns,1999; Espy & Bull, 2005; Oh & Lewis, 2008). In the present study,we measured the integration of these skills because children’sability to orchestrate these three skills has been shown to be a keypredictor of academic success (McClelland et al., 2007; Ponitz etal., 2009).

Although there are a number of reliable and valid direct mea-sures of self-regulation, many have been designed for the labora-tory or with clinical populations, or have limited usefulness inclassroom settings because of practical constraints (Fahie & Sy-mons, 2003; Pickering & Gathercole, 2004). For example, taskssuch as the go/no-go task require a computer to administer (Simp-son & Riggs, 2006) and tasks such as the Attention Network Taskor the Working Memory Test Battery for Children take a relativelylong time to administer (Pickering & Gathercole, 2004; Rueda etal., 2005). A number of tasks such as the day/night Stroop task(Gerstadt, Hong, & Diamond, 1994) and the Dimensional ChangeCard Sort (Muller, Dick, Gela, Overton, & Zelazo, 2006) requireless time, but children often successfully pass them by 5 years ofage (Carlson, 2005). As noted by Ponitz et al. (2009), one task(Luria’s Hand Game) is available that does not require specialmaterials or substantial amount of time to administer, but this taskis designed for early elementary school children and may not beappropriate for younger children (Fahie & Symons, 2003).

The HTKS task is a direct measure designed to capture theintegration of attention, working memory, and inhibitory control.In the task, children are instructed to perform the opposite of fourpaired commands: “touch your head” and “touch your toes”;“touch your knees” and “touch your shoulders.” The first pair ofcommands is used in the first half of the task, and the second pairis incorporated in the second half of the task. The task is uniquebecause it is a direct measure of children’s ability to use attention,working memory, and inhibitory control skills simultaneously toproduce a behavior in a social interaction, which is similar to howchildren often use these skills in classroom settings. Specifically,

366 WANLESS ET AL.

the HTKS aims to (a) tap attention skills by requiring children tofocus on the task and instructions, (b) tap working memory byremembering multiple task instructions while applying each ofthem (i.e., remembering to touch your head instead of your toeswhile responding to a command to “touch your knees”), and (c) tapinhibitory control by requiring children to stop their dominantresponse (to touch their toes when asked to touch their toes) andreplace this response with the more adaptive response (to touchtheir head when asked to touch their toes, as instructed at thebeginning of the task).

Research demonstrates that the HTKS is significantly related toindividual measures of attention, working memory and inhibitorycontrol. For example, higher HTKS scores have been significantlyrelated to higher parent ratings of attention and inhibitory controlin the United States (Ponitz et al., 2009) and to higher directlymeasured attention and working memory in Beijing (Lan & Mor-rison, 2008). Moreover, recent research has found evidence thatthe HTKS is significantly related to direct measures of attentionand inhibitory control in the United States (R. Duncan & McClel-land, 2010). The measure is also useful because it is quick toadminister, does not require special materials, requires limitedtraining, and yields reliable scores when the assessment is admin-istered in classroom settings (McClelland et al., 2007; Ponitz et al.,2008; Ponitz et al., 2009).

Goals of the Present Study

In the present study, the psychometric properties of scores on adirect measure of behavioral regulation, the HTKS, were analyzedfor young children in the United States, Taiwan, South Korea, andChina by examining (a) the nature and variability of behavioralregulation measured with the HTKS in each society, includingrelations with teacher-rated classroom behavioral regulation, and(b) relations to early mathematics, vocabulary, and early literacyskills.

We hypothesized that the HTKS would capture substantialvariability in behavioral regulation in the samples from all foursocieties, as has been found with previous research in the UnitedStates (Ponitz et al., 2009). In other words, we hypothesized thatscores would reflect the full range of the task. It was somewhatunclear whether scores on the HTKS would relate to teacher-ratedclassroom behavioral regulation in the Asian samples becauseresults from previous research have been mixed. In the UnitedStates, teacher-rated classroom behavioral regulation was signifi-cantly and positively related to the HTKS, but research conductedwith a simpler form of the HTKS (the Head-to-Toes task [HTT])in Taiwan did not find significant relations with teacher-ratedclassroom behavioral regulation (Ponitz et al., 2009; Wanless etal., in press). Similarly, in previous research in Taiwan, teachersrated children differently than did research assistants who directlyobserved children’s behavior (Jose et al., 2000). Therefore, it wasplausible that HTKS and teacher-rated classroom behavioral reg-ulation would be positively related in the present study but that therelation would be weaker in the Asian samples than in the U.S.sample. Finally, on the basis of previous studies conducted withthe HTKS in the United States and the HTT in Taiwan, weanticipated that scores on the HTKS would be positively related toearly mathematics, vocabulary, and early literacy skills.

Method

Participants

Data for the present study were collected in four societies—theUnited States, Taiwan, South Korea, and China—and consisted ofparents, teachers, and children who volunteered to participate.With all four samples combined, 814 children, 695 parents, andone teacher from each of 73 classrooms participated in the presentstudy (see Table 1 for descriptive statistics). The ages of thechildren ranged from 3.12 to 6.50 years old, with each sample’sage range overlapping with one another. Taiwan had the youngestmean age (4.56 years) and the United States had the oldest meanage (5.48 years). China had the largest age range (3.12 to 6.45years) and Taiwan had the smallest (3.89 to 5.00 years). Overall,the majority of children (91%) in the present study were either 4 or5 years old.

United States. Participants from the United States were re-cruited from two geographic locations: Michigan and Oregon.Data from both sites were combined for the present study (seePonitz et al., 2009, for a description of each U.S. site).

The children in the U.S. sample (N � 310) were between 4.14and 6.24 years old (M � 5.48, SD � 0.33; see Table 1) and werein 40 kindergarten classrooms. A total of 51% (n � 159 were girlsand 49% were boys (n � 151); the average mother’s educationlevel was some college. A small subgroup of the children hadmothers with a high school degree or less (11%; n � 35). Some ofthe children (4%; n � 13) from Oregon spoke Spanish as their firstlanguage and were given assessments in Spanish. In the overallU.S. sample, the majority of children were Caucasian (74%), andthe remaining children were Asian (7%), Hispanic (6%), or ofanother ethnicity (13%), including children identified as biracial.

Taiwan. Ages of the Taiwanese participants (N � 158)ranged from 3.89 to 5.00 years old (M � 4.56, SD � 0.29; seeTable 1). Children were from 10 preschool classrooms in Taipei(the capital city of Taiwan), and about half of the children weregirls (48%; n � 76). On average, mothers had between a highschool and a college degree, and slightly more than half of thechildren had mothers with a high school degree or less (51%; n �80). Most parents were born in Taiwan (77% of mothers, 100% offathers), with the remaining mothers originally from China (4%),Vietnam (4%), Indonesia (1%), or the Philippines (1%).

South Korea. Children in the South Korean sample (N �227) were between 3.58 and 6.50 years old (M � 5.05, SD � 0.85;see Table 1) and were from 16 preschool classrooms housed inthree child-care centers in Seoul (the capital city of South Korea)and Kyonggi province. Slightly less than half of the participantswere girls (40%; n � 91), and 60% (n � 136) were boys. Mothershad an average education level between a high school and acollege degree, and almost half of the children had mothers with ahigh school degree or less (43%; n � 98). All of the children in thissample were of South Korean descent.

China. Children in the sample from China (N � 119) werebetween 3.12 and 6.45 years old (M � 5.03, SD � 0.62; see Table1) and were from seven preschool classrooms in Beijing (thecapital city of China); 46% (n � 55) were girls, and 54% (n � 64)were boys. Information about mother’s education level was notavailable for the Chinese sample. All of the children were origi-nally from China.

367MEASURING BEHAVIORAL REGULATION IN FOUR SOCIETIES

Procedure

Behavioral regulation and academic data were collected fromchildren and teachers in all four samples and additional back-ground information was collected from parents in three samples(the United States, Taiwan, and South Korea). The Head–Toes–Knees–Shoulders (HTKS) and academic tasks (early mathematics,vocabulary, and early literacy) were given to children in the fall ofthe school year. The school year in South Korea and China,however, begins in the spring and ends in the winter, so the fallassessment represents the middle of the school year for thesechildren. The school year in the United States and Taiwan beginsin the fall and ends in the spring, so the fall assessment representsthe beginning of the school year for these children. All threeacademic domains were assessed in each sample except vocabu-lary in China and early literacy in Taiwan (see Table 1). In eachsample, research assistants visited the schools and assessed thechildren in unused classrooms, multipurpose rooms, or other quietspaces. Assessments were given in two sessions, with each sessionlasting between 15 and 40 min. In addition, in all four samples,teachers completed a questionnaire rating classroom behavioralregulation for each child in their classroom in the fall of the schoolyear, at the same time as the HTKS assessment, except in the

United States, where it was completed in the spring of the schoolyear. In other words, in the U.S. sample, the HTKS and academicassessments were administered at the beginning of the school yearand the teacher ratings at the end of the school year. In the SouthKorean and Chinese samples, the HTKS, academic assessments,and teacher ratings were collected in the middle of the school year.In the Taiwan sample, the HTKS, academic assessments, andteacher ratings were collected in the beginning of the school year.

Measures

All measures that were not previously used in each society weretranslated and/or back-translated by professors who were nativespeakers and also fluent in English, as well as bilingual graduatestudents from the society where the assessments were used. Inthe United States (Oregon only), assessments that were not previ-ously translated into Spanish were translated and back-translatedby bilingual research assistants and a professor of Spanish andused with participants who were identified by teachers as havingSpanish as their first language.

Background questionnaire. In the United States, Taiwan,and South Korea, parents completed background questionnairesasking about parent education level, prior child-care experience,

Table 1Descriptive Statistics by Society for Original and Imputed Data, Including Mean, Standard Deviation, and % Missing

Variable

United States (N � 310) Taiwan (N � 158) South Korea (N � 227) China (N � 119)

Original Imputed Original Imputed Original Imputed Original Imputed

Child ageM 5.48 5.48 4.56 4.56 5.06 5.05 5.02 5.03SD 0.33 0.33 0.29 0.29 0.85 0.85 0.62 0.62% 0.6 0.6 3.1 0.0

GenderM 0.49 0.49 0.52 0.52 0.60 0.60 0.54 0.54SD 0.50 0.50 0.50 0.50 0.49 0.49 0.50 0.50% 0.0 0.0 0.0 0.0

Mother’s educationM 3.32 3.25 2.45 2.45 2.63 2.67SD 0.70 0.77 0.45 0.61 1.10 1.10% 19.4 13.3 28.6

HTKSM 26.28 26.28 15.84 15.79 23.99 23.99 31.77 31.77SD 11.05 11.03 13.17 13.12 12.96 12.94 8.81 8.77% 0.0 0.6 0.0 0.0

CBRSM 4.02 4.00 3.90 3.89 3.75 3.75 3.76 3.76SD 0.68 0.68 0.62 0.62 0.72 0.71 0.77 0.77% 35.5 13.9 0.4 1.7

MathematicsM 434.00 434.04 18.11 18.10 12.77 12.85 5.95 5.95SD 15.33 15.30 8.65 8.60 3.87 4.01 2.76 2.76% 0.6 1.3 33.9 0.8

VocabularyM 477.41 477.45 38.38 38.35 13.71 13.99SD 13.47 13.45 15.76 15.74 6.07 6.22% 0.6 0.6 33.9

Early literacyM 375.31 375.37 28.81 28.33 23.61 23.61SD 30.18 30.16 30.49 31.77 20.27 20.20% 0.6 33.9 0.0

Note. HTKS � Head–Toes–Knees–Shoulders task; CBRS � Child Behavior Rating Scale. Mother’s education is coded as 1 � 0–8 years; 2 � �8 �12 years; 3 � �12 � 16 years; 4 � �16 years.

368 WANLESS ET AL.

child age, gender, and ethnicity. In China, some background in-formation was collected from school records.

Direct measure of behavioral regulation. The Head–Toes–Knees–Shoulders task (HTKS) was used to measure behavioralregulation, and children were given the task in their native lan-guage. Each of the 20 items was scored with 0 for an incorrectresponse (touching his or her head when asked to touch his or herhead), 1 for a self-corrected response (initially responding incor-rectly, but correcting himself or herself), or 2 for a correct response(touching his or her toes when asked to touch his or her head).Total scores on the HTKS range from 0 to 40 points. There are twoforms of the HTKS. Form A starts with head–toes commands, andForm B starts with knees–shoulders commands (Items 1–10). InItems 11–20, all four body parts are used in both forms of the task.In the present study, there were no significant differences betweenscores on the two forms in the United States, Taiwan, or Chinawhen controlling for age (ps � .05), which is consistent withprevious research in the United States and Taiwan (Ponitz et al.,2009; Wanless et al., in press). In South Korea, only one versionof the task (Form A) was used.

Research assistants were trained on the HTKS by studying thetask forms, watching videos of trained research assistants givingthe task to children, and practicing with other research assistants.In the United States, scores on the HTKS and the Head-to-Toestask (HTT), a simple version of the HTKS, have shown stronginterrater reliability (Connor et al., 2010; McClelland, 2009). Fur-ther, previous research of scores from the Head-to-Toes task inTaiwan has demonstrated strong interrater reliability (Wanless etal., in press). In the U.S., Taiwanese, and Chinese samples in thepresent study, examiners scored different children, so traditionalmethods of interrater reliability could not be calculated. However,there were no significant differences in these samples betweenexaminers in children’s HTKS scores after controlling for childage: U.S., F(141, 299) � 1.25, p � .05; Taiwan, F(40, 155) �1.08, p � .05; and China, F(28, 114) � 1.28, p � .05. In the SouthKorean sample, two research assistants rated the same children fora subsample of the participants (n � 72) and had good consistencyfor each HTKS item (ICC � .71, p � .001).

Teacher-Rated Measure of Classroom BehavioralRegulation

In all four samples, the Child Behavior Rating Scale (CBRS)was used to measure teacher-rated classroom behavioral regulation(Bronson et al., 1995). Items on the CBRS ask teachers to rate thechild’s typical behaviors when using materials, interacting withpeers, and completing tasks, on a scale ranging from 1 (never) to5 (usually/always). To determine whether a classroom behavioralregulation factor was present in each of the four societies, CBRSscores were analyzed with principal axis factor analysis with apromax rotation. In each of the four samples, the same 10-itemclassroom behavioral regulation factor emerged that has beenfound in other research in the United States (�s � .94–.95;Matthews et al., 2009; Ponitz et al., 2009), and in Taiwan (� � .94;Wanless, et al., in press). In the present study, scores on the CBRSclassroom Behavioral Regulation factor in each society had stronginteritem reliability (U.S.: � � .94; Taiwan: � � .94; South Korea:� � .94; and China: � � .95). This factor included items such as“Concentrates when working on a task; is not easily distracted by

surrounding activities” and “Completes learning tasks involvingtwo or more steps (e.g., cutting and pasting) in an organized way.”The mean score on the CBRS Classroom Behavioral Regulationfactor ranged from 1 to 5, with higher scores demonstrating higherlevels of classroom behavioral regulation.

Academic Achievement

United States. Mathematics, Literacy, and Vocabulary sub-tests from the Woodcock–Johnson Psycho-Educational Battery–IIITests of Achievement (WJ-III; Woodcock & Mather, 2000) wasused for the English-speaking U.S. children, and the BaterıaWoodcock–Munoz–R (Baterıa-R; Woodcock & Munoz-Sandoval,1996) was used for the Spanish-speaking U.S. children. Therewere 13 Spanish-speaking children out of the sample of 310children. Each child’s score was based on the norms for thespecific version of the test. The Baterıa–R standard scores areconsidered to be equivalent to the Woodcock–Johnson Psycho-Educational Battery—Revised (Woodcock & Johnson, 1989),which is the English version of the assessment prior to the oneused for the English-speaking children in our sample (WJ–III).Unfortunately, the Spanish version of the WJ–III was not availableat the time of testing, so the Baterıa–R was used. Children weretested on the same subtests (Applied Problems, Letter–Word Iden-tification, and Picture Vocabulary) for the WJ–III and theBaterıa–R. The sample to standardize the items on the Baterıa–Rconsisted of approximately 2,000 native Spanish-speaking individ-uals (Woodcock & Munoz-Sandoval, 1996). For the WJ–III, thesample consisted of a nationally representative U.S. sample of8,800 individuals (McGrew & Woodcock, 2001). When both theBaterıa–R and the Spanish equivalent of the WJ–III (Baterıa III;Munoz-Sandoval, Woodcock, McGrew, & Mather, 2005) wereadministered to children in previous research, scores on bothversions were highly correlated at r � .95 (Liew et al., 2008). Thisevidence suggests that the English WJ–III and Spanish Baterıa–Rscores were comparable and could be used in our analyses, as hasbeen done in previous research. The Applied Problems subtestassesses early mathematics and includes questions about quantity,time, money, and word problems. The Letter–Word Identificationsubtest measures early literacy and asks children to name lettersand read actual words, and the Picture Vocabulary subtest usespictures to assess expressive vocabulary. W-scores were used totake into account age at the time of assessment and to allow forcomparison of performance of children across a range of ages. Inthe United States, interrater reliability of scores from these subtestsis greater than .85 (Woodcock & Mather, 2000).

Taiwan. Early mathematics and vocabulary skills were as-sessed in Taiwan with measures that had previously been trans-lated into traditional Chinese and used in Taiwan. The Test ofEarly Mathematics Ability–2 (TEMA-2) measured relative mag-nitude, counting, calculation, and enumeration by asking childrento count objects on a page, to determine greater than and less than,and to correctly identify numbers (Ginsburg & Baroody, 1990). Inprevious research in Taiwan, scores from the TEMA-2 had highinternal consistency (.89–.90) and test–retest reliabilities between.91 and .94 (Hsu, 1998, 2000). Early vocabulary was measuredwith the Peabody Picture Vocabulary Test—Revised (PPVT–R),which asked children to point to pictures that were named by

369MEASURING BEHAVIORAL REGULATION IN FOUR SOCIETIES

research assistants. Scores on the PPVT–R had a split-half reli-ability of .90–.97 in Taiwanese samples (Lu & Liu, 1998).

South Korea. Early mathematics and early vocabulary weremeasured with subtests of the Korean–Wechsler Preschool andPrimary Scale of Intelligence (K–WPPSI; Park, Kwak, & Park,1989). The Mathematics subtest measures relative magnitude,counting, and calculation. In previous research, scores on theKorean mathematics subtest had a split-half reliability of .82–.87and a test–retest reliability of .68 for 4- to 6-year-olds (Park et al.,1989). The Vocabulary subtest requires children to identify pic-tured objects and to define words. Scores from this subtest had asplit-half reliability of .78–.86 and a test–retest reliability of .63for 4- to 6-year-olds (Park et al., 1989). Early literacy skills wereassessed with the Test of Hangul Word Reading, in which childrenare asked to decode two-syllable Korean words and pseudowords(Choi & Yi, 2007). The internal consistency of scores from the testwas .99, split-half reliability was .98–.99, and test–retest reliabilitywas .93–.97 in previous research (Choi & Yi, 2007).

China. Early mathematics and early literacy skills were as-sessed in China with the Zareki-KP task (von Aster, 2001) and theCharacter Recognition task (Chow, McBride-Chang, Cheung, &Chow, 2008), respectively. Using subtests of the Zareki-KP task,we assessed counting and calculation (addition and subtraction)skills separately and added their scores together to make a math-ematics composite score. Counting and calculation scores in thepresent study were significantly correlated (r � .44, p � .001).The task had previously been translated into simple Chinese (Liu,2007). In Liu’s research, scores on the counting and calculationtests had a reliability of .84 and .87, respectively, and werecorrelated with teacher reports and cognitive tasks. For the literacytask, all traditional characters were translated into simplified Chi-nese, and the children were asked to read the characters aloud.

Results

The present study investigated the psychometric properties ofscores on a direct measure of behavioral regulation in four soci-eties. Specifically, we examined (a) the nature and variability ofbehavioral regulation measured with the HTKS in each society,including relations with teacher-rated classroom behavioral regu-lation, and (b) relations to early mathematics, vocabulary, andearly literacy skills.

Missing Data and Multiple Imputation

The HTKS had missing data for just one child out of the 814children that participated across the four societies (see Table 1).The academic outcomes had less than 2% of missing data in eachsociety, except for South Korea, which had 34% missing data foreach academic domain, including early mathematics, vocabulary,and early literacy. Information on the amount of missing data forall variables is presented in Table 1. In all four samples, themajority of participants did not have missing data for more thanone variable.

In order to deal with missing data, we used multiple imputationfor all final models (Acock, 2005). When multiple imputation isused, data are assumed to be missing at random, meaning that thepattern of missingness can be explained by variables that areincluded in the models or by auxiliary variables (variables not

included in our models that are theoretically related to missing dataon the variables of interest) and any remaining missingness israndom (Schafer & Graham, 2002). First, we examined relationsbetween missingness and the variables in our models. Some of thepredictors (such as child age and gender) were significantly relatedto the missingness of variables that were imputed in all foursocieties. Second, we explored possible auxiliary variables. Forvariables with more than 5% missing data, logistic regressionswere conducted between possible auxiliary variables and dummyvariables indicating whether data were missing. Additional cultur-ally relevant auxiliary variables not included in our models but thatmight explain missingness were chosen for each sample and in-cluded the amount of experience in preschool, number of hours inpreschool each week, family income, whether the child spokeSpanish, or whether the child was an ethnic minority.1 Althoughthe missing at random assumption cannot be fully tested, thesefindings suggest that the missing data in the present study may bemissing at random, so we present results using the imputed data(Acock, 2005; Meng, 1995; Rubin, 1996).

To obtain unbiased parameter estimates, multiple imputation(using Stata) was used to create 10 imputed data sets for each ofthe four samples (Acock, 2005; StataCorp, 2007). Descriptivestatistics for each society determined with the original and imputeddata are presented in Table 1. Descriptive statistics were highlysimilar for the original and imputed data, which gave us moreconfidence in using the imputed data. Descriptive statistics, cor-relations, and regression coefficients were also based on analyseson the 10 imputed data sets for each sample. Correlation signifi-cance levels were adjusted to account for the inflated sample size(see Table 2).

Analytic Strategy

We began our analyses by examining descriptive statistics,including intraclass correlation coefficients (ICCs) using multi-level models to account for the nested structure of the data (chil-dren nested in classrooms; Raudenbush & Bryk, 2002). Multilevelanalyses were then used to analyze the data, acknowledging thenonindependent nature of the children due to classroom member-ship, and to ensure that our standard errors were not biased (Peugh,2010).

1 In South Korea, amount of preschool experience significantly pre-dicted missingness for variables with more than 5% missing data (mathe-matics, early literacy, vocabulary, and mother’s education), so it wasincluded as an auxiliary variable in the multiple imputation model. In theUnited States and Taiwan, the auxiliary variables did not significantlypredict missingness on any variables that had more than 5% missing data(mother’s education and teacher-rated classroom behavioral regulation), soonly the predictor variables (child age, gender, mother’s education, andteacher-rated classroom behavioral regulation) were used in the imputationmodel. In China, no variables had more than 5% missing data, so only thepredictor variables (child age, gender, mother’s education, and teacher-rated classroom behavioral regulation) were included in the imputationmodel.

370 WANLESS ET AL.

Nature and Variability of Behavioral RegulationMeasured by the HTKS

Variability in HTKS scores in each society was found withChina showing low and Taiwan showing high variability in scores(see Table 3). In all four societies, children’s HTKS scores rangedfrom 0 to 40, utilizing the entire range of the task. Taiwan (n � 1;0.6%) and South Korea (n � 3; 1.3%) had very few children reachceiling levels of 40 points on the task. China (n � 8; 6.7%) and theUnited States (n � 7, 2.3%), had more children, but overall arelatively small number earned 40 points on the task. Further, lessthan 10% of the children in each society, except for Taiwan (n �36; 23%), scored at floor level. Distributions of task scores withineach society were somewhat skewed, but skewness and kurtosisvalues did not exceed accepted levels for normal distributions(Kline, 2005; see Table 3). China’s HTKS score distribution wasparticularly skewed, with many high-scoring children, and Tai-wan’s distribution had a positive skew, with the majority of chil-dren in Taiwan having low scores. Taiwanese children, however,were the youngest across all the samples.

Relations between child demographic variables and HTKSscores were also analyzed with correlations within each society(see Table 2). Child age and parent education level were positivelyand significantly correlated with HTKS scores. Child gender wasalso significantly related to HTKS scores in the United States butnot in the Asian societies. This relation was negative, indicatingthat being a boy was related to lower HTKS scores in the UnitedStates.

To better understand the psychometric properties of scores onthe HTKS in the four samples, we compared relations betweenHTKS scores and teacher-rated classroom behavioral regulationscores. In the United States and in South Korea, higher scoreson the HTKS were significantly related to higher teacher-ratedclassroom behavioral regulation scores (U.S.: r � .29; SouthKorea: r � .23; see Table 2). In Taiwan and China, however,relations between HTKS and classroom behavioral regulationscores were weak and not significant (Taiwan: r � .09; China: r �.12).

We also compared the amount of variability in HTKS scores andclassroom behavioral regulation within each society by calculatingthe coefficient of variation (standard deviation divided by themean). In all four societies, the HTKS scores reflected greatervariability than the teacher ratings of classroom behavioral regu-lation (see Table 3).

Finally, we calculated intraclass correlation coefficients (ICCs)using multilevel models for the HTKS and CBRS in each societyto determine how much of the variation in these measures was dueto classroom membership (see Table 3). In all four societies, morevariance in CBRS scores than HTKS scores was attributable toclassroom differences, controlling for the predictors used in thepresent regression analyses (see Table 4).

HTKS Scores and Relations to Mathematics,Vocabulary, and Early Literacy

To investigate our second research question, we used correla-tions and multilevel modeling within each sample to determinewhether HTKS scores related to early academic outcomes. In allsamples that measured mathematics (U.S., Taiwan, South Korea,China) and in those that measured vocabulary (U.S., Taiwan,South Korea), higher HTKS scores were significantly correlatedwith higher mathematics skills and higher vocabulary scores (seeTable 2). In addition, higher HTKS scores were significantlyrelated to higher early literacy scores in South Korea, the UnitedStates, and China (early literacy was not assessed in Taiwan).

Table 3Descriptive Statistics of Head–Toes–Knees–Shoulders (HTKS)Task Direct Behavioral Regulation Measure and ChildBehavior Rating Scale (CBRS) Teacher-Rated BehavioralRegulation Measure

Variable United States Taiwan South Korea China

HTKSVariance 121.72 172.26 167.38 76.95Coefficient of variation .42 .83 .54 .28Skewness �0.97 0.18 �0.67 �2.06Kurtosis �0.17 �1.33 �0.94 4.32ICC .06 .00 .09 .03

CBRSVariance .46 .38 .51 .60Coefficient of variation .17 .16 .19 .20Skewness �0.50 0.04 �0.40 �0.03Kurtosis 3.25 2.56 2.77 1.97ICC .15 .38 .04 .60

Note. ICC � intraclass correlation coefficient, which is reported control-ling for age.

Table 2Correlation Matrix

Variable HTKS Vocabulary Mathematics Early literacy

United States (N � 310)

Child age .12� .17�� .19��� .19���

Gender �.16�� .01 .04 �.01Parent education .17�� .37��� .26��� .20���

CBRS .29��� .09 .28��� .23���

HTKS .31��� .47��� .30���

Taiwan (N � 158)

Child age .20� .27��� .37���

Gender �.06 �.13 �.17�

Parent education .18� .19� .10CBRS .09 .12 .33���

HTKS .30��� .34���

South Korea (N � 227)

Child age .53��� .51��� .68��� .55���

Gender �.03 .03 �.06 �.11†

Parent education .22��� .20�� .26��� .11†

CBRS .23��� .13� .22��� .26���

HTKS .36��� .59��� .50���

China (N � 119)

Child age .24�� .42��� .23��

Gender �.09 �.19� .01CBRS .12 .15 .14HTKS .40��� .24��

Note. HTKS � Head–Toes–Knees–Shoulders task; CBRS � Child Be-havior Rating Scale. Gender is coded as 1 for boys and 0 for girls.† p � .10. � p � .05. �� p � .01. ��� p � .001.

371MEASURING BEHAVIORAL REGULATION IN FOUR SOCIETIES

To consider the effect of children being nested in classrooms,ICCs were calculated for all academic outcomes in each sample.Less than 9% of the variance in mathematics, reading, and vocab-ulary was between classrooms. ICCs in South Korea and Taiwanreached statistical significance (ps �.05), but none of the ICCs inthe United States and China were statistically significant (p �.05). Because some of the ICCs were significant, multilevel anal-yses conducted with hierarchical linear modeling (HLM) softwarewere conducted for all models to maintain a consistent analyticalapproach (Raudenbush, Bryk, Cheong, & Congdon, 2004).

Our final models examined the relations between directly mea-sured behavioral regulation (HTKS) scores and early mathematics,vocabulary, and early literacy skills for each sample, controllingfor child age, gender, mother’s education (except in China),teacher perceptions of children’s behavioral regulation, andwhether the data were collected in Oregon or Michigan (only in theU.S. analyses; see Table 4). Overall, HTKS scores were signifi-cantly related to all achievement outcomes with a few differencesbased on the sample. In all four societies, higher HTKS scoreswere significantly related to higher mathematics skills, and pat-terns were similar in the three samples that assessed early literacy(U.S., China, and South Korea) and vocabulary (U.S., Taiwan, andSouth Korea). Specifically, higher scores on the HTKS signifi-cantly predicted higher early literacy in the United States, China,

and South Korea, and higher vocabulary skills in the United Statesand Taiwan, but not in South Korea.

Discussion

Overall, evidence for the psychometric properties of HTKSscores was found across the four samples, but some differencesalso emerged. HTKS scores demonstrated variability in behavioralregulation in each society but were differentially related to teacherperceptions of classroom behavioral regulation, with significantrelations present in the United States and South Korea but not inTaiwan or China. Finally, higher behavioral regulation scores onthe HTKS were significantly related to higher mathematics, vo-cabulary, and early literacy (with the exception of vocabulary inSouth Korea), beyond the influence of control variables andteacher-rated classroom behavioral regulation.

Nature and Variability of Behavioral Regulation

HTKS scores reflected individual variation in behavioral regu-lation in all four samples, which supports research in the UnitedStates finding that HTKS scores capture variability in children’sbehavioral regulation in early childhood (McClelland et al., 2007;Ponitz et al., 2008; Ponitz et al., 2009). Specifically, scores in all

Table 4Coefficients and Standard Errors From Multilevel Models of Early Mathematics, Vocabulary, and Early Literacy Skills on BehavioralRegulation by Society

Variable

United States Taiwan South Korea China

Math VocabularyEarly

literacy Math Vocabulary Math VocabularyEarly

literacy MathEarly

literacy

HTKSUnstand. coeff. .56��� .29��� .63��� .15�� .24�� .08�� .02 .60� .10��� .43�

SE .07 .07 .15 .05 .09 .02 .05 .22 .03 .21Stand. coeff. .40 .24 .23 .23 .20 .26 .04 .24 .32 .19

Child age yearsUnstand. coeff. 4.10† 7.32�� 11.66� 7.39�� 11.01�� 2.31�� 3.09��� 12.82† 1.47��� 6.82�

SE 2.39 2.24 5.00 2.12 4.17 0.53 0.79 6.63 0.37 3.05Stand. coeff. .09 .18 .13 .25 .20 .49 .42 .34 .33 .21

Gendera

Unstand. coeff. 4.31�� 1.22 3.31 �1.43 �4.25† �0.63 0.14 �5.99 �0.39 4.32SE 1.51 1.42 3.27 1.23 2.39 0.41 0.80 5.38 0.46 3.74Stand. coeff. .14 .05 .05 �.08 �.14 �.08 .01 �.09 .07 .11

Mother’s educationUnstand. coeff. 4.20�� 5.84��� 7.21�� 1.00 4.45� 0.65�� 0.86† 1.67SE 1.31 1.02 2.36 1.06 2.02 0.22 0.47 2.13Stand. coeff. .21 .33 .18 .07 .17 .18 .15 .06

CBRSUnstand. coeff. 3.52� �0.11 6.00� 4.20�� 0.98 0.41 0.68 6.20� 0.33 3.83SE 1.40 1.16 2.70 1.25 2.29 0.47 0.90 2.98 0.29 2.42Stand. coeff. .16 �.01 .14 .30 .04 .07 .08 .14 .09 .15

Siteb

Unstand. coeff. �4.70� 2.73 �5.76SE 2.09 1.81 4.50Stand. coeff. �.12 .08 �.08

Note. Unstand. coeff. � unstandardized coefficient; SE � standard error; stand. coeff. � standardized coefficient; HTKS � Head–Toes–Knees–Shoulderstask; CBRS � Child Behavior Rating Scale. Parent education is scored as 1 � 0–8 years; 2 � 9–12 years; 3 � 13–16 years; 4 � �16 years. In the U.S.sample, site was also controlled because data were collected in Michigan and Oregon.a Boys � 1; girls � 0. b Michigan � 1; Oregon � 0.† p � .10. � p � .05. �� p � .01. ��� p � .001.

372 WANLESS ET AL.

four cultures covered the entire range of possible HTKS scores.This suggests that the HTKS differentiated children and was nottoo easy or difficult for the majority of children assessed in eachculture.

In each society, child age and parent education was significantlyand positively related to HTKS behavioral regulation scores. Childgender, however, was not significantly related to HTKS scores inthe Asian samples. It may be that behavioral regulation is pro-moted more strongly, regardless of gender, in the Asian societies.For example, previous research has found that teachers gave moreinstructions regarding behavioral regulation in Chinese classroomscompared with U.S. classrooms (Lan et al., 2009). Moreover,recent research in the United States has documented that boysscored significantly worse on the HTKS than girls in the fall andspring of the kindergarten year (Matthews et al., 2009). Thus, boysmay be more at risk for poor behavioral regulation in the UnitedStates but not in Asian samples, which could be due to a greatertolerance for aggressive, unregulated behavior in boys in theUnited States (Entwisle, Alexander, & Olson, 2007). Taken to-gether, these results extend previous research by finding thatHTKS scores demonstrate variability in behavioral regulationacross four samples in the United States, Taiwan, South Korea, andChina. They also suggest that variability in scores is related tochild demographic factors like age and parent education, butcultural differences exist in how gender is related to HTKS scores.

Relations Between the HTKS and Teacher-RatedClassroom Behavioral Regulation

Results of this study also point to culturally specific relationsbetween HTKS scores and teacher ratings of classroom behavioralregulation. In the United States and South Korea, higher HTKSscores were significantly related to higher classroom behavioralregulation ratings. In Taiwan and China, however, HTKS scoreswere not significantly related to teacher ratings. Previous researchhas also found inconsistent relations between the HTKS scores andteacher-rated classroom behavioral regulation, indicating the needfor further research into this issue (Ponitz et al., 2009; Wanless etal., in press). One explanation for differences in relations betweenclassroom behavioral regulation and HTKS scores may be thefamiliarity of the teachers with the children. In the United Statesand South Korea, where the HTKS was significantly related toteacher-ratings, teacher ratings were collected after the teachershad one semester with children in their classes. In Taiwan, teacherratings were collected at the beginning of the school year, butHTKS was not significant related to teacher ratings. It is possiblethat the Taiwanese teachers had less experience with the childrenin their classes before completing the teacher ratings. However,although teachers in China rated children after one semester,HTKS scores were not significantly related to teacher ratings. Thislack of relation may be a function of the particularly large classsizes in China, which may have limited how familiar the teacherswere with the children, even after one semester (Stevenson &Stigler, 1994). Overall, it is possible that teacher familiarity influ-enced teacher ratings, but we were unable to directly test this in thepresent study, and future research should examine it. On the basisof the differential relations between the HTKS and teacher ratings,our results suggest that in addition to teacher reports, HTKS scores

may independently contribute useful information for differentiat-ing children’s levels of behavioral regulation skills.

Finally, classroom behavioral regulation also differed fromHTKS scores in predictability of academic outcomes. In all foursamples and for all academic outcomes assessed, HTKS scoreswere uniquely predictive of academic skills beyond classroombehavioral regulation. Two possibilities may explain the differentinformation provided by these two assessments. First, HTKSscores may reflect different aspects of regulation than classroombehavioral regulation which are differentially relevant for aca-demic success. An accumulation of classroom experiences with thechildren, with skills other than behavioral regulation, may beincluded in teachers’ ratings, in contrast to a directly administeredmeasure. For example, children may have difficulty with persis-tence or cooperation, which may influence teachers’ ratings of thechildren’s behavioral regulation. These skills, however, are notaspects of behavioral regulation as measured by the HTKS task.Second, teacher ratings may capture different child, classroom,teacher, or cultural characteristics compared with what the HTKSassesses (Mashburn, Hamre, Downer, & Pianta, 2006). One cross-cultural study of Likert scales, for example, suggested that culturalvalues shaped teachers expectations for the child characteristicsrepresented on a Likert scale, which led to the reference-groupeffect, which is the lack of a common reference group. This madecross-cultural comparisons of Likert scale scores problematic(Heine et al., 2002). In the present study, teachers within a societymay also have had difficulty rating children with a similar refer-ence group in mind based on teachers’ different years of experi-ence or different range of ability levels represented in their classes.It is also possible that the regulation measures were not reliable inAsian cultures, although evidence has been found for the measuresin U.S. samples (Matthews et al., 2009; McClelland et al., 2007;McClelland & Morrison, 2003; Ponitz et al., 2009) and for asimplified version of the HTKS in Taiwan (Wanless et al., inpress). Future research is needed to further examine this issue.Together, results comparing HTKS scores and teacher ratingssuggest that in all four samples, HTKS scores provide a uniquewindow on children’s academic achievement, beyond that ofteacher ratings of classroom behavioral regulation.

Relations Between Behavioral Regulation on theHTKS and Early Mathematics, Vocabulary,and Early Literacy

In the present study, HTKS scores predicted children’s aca-demic achievement in all samples. Specifically, HTKS scoressignificantly predicted early mathematics in all four samples, earlyliteracy in all three samples that it was measured (United States,South Korea, and China), and early vocabulary in the samples thatit was measured (United States and Taiwan), with the exception ofvocabulary for South Korean children. These significant relationswere present after controlling for child age, gender, mother’seducation (except in China), teacher-rated classroom behavioralregulation, and whether the data were collected in Oregon orMichigan in the U.S. analyses. It is likely that integrating attention,working memory, and inhibitory control helps children performbetter on achievement tests and in academic settings. Early child-hood classrooms place demands on children to pay attention toinstructions, to remember classroom and activity rules, and to

373MEASURING BEHAVIORAL REGULATION IN FOUR SOCIETIES

inhibit impulses to behave in ways that conflict with these instruc-tions and rules. Children who are able to master these skills, anduse them in tandem in classroom settings, are more likely toacquire new skills and knowledge, which will translate into in-creased mathematics, vocabulary, and reading scores.

This relation was present beyond the influence of classroombehavioral regulation and important background variables. Thesefindings align with previous research relating behavioral regula-tion and academic achievement (Howse et al., 2003; McClelland,Piccinin, & Stallings, 2010) as well as effortful control—whichincludes aspects of behavioral regulation—and academic achieve-ment (Checa et al., 2008; Zhou et al., 2010). The predictive utilityof HTKS scores is noteworthy given that these effects were foundafter taking classroom behavioral regulation into account. Thissuggests that HTKS behavioral regulation is important for earlyacademic success in all four societies, which supports a growingbody of evidence found with U.S. samples (Ponitz et al., 2009).

Of all of the academic domains tested in each sample, the onlydomain not significantly related to HTKS behavioral regulationscores was vocabulary in the South Korean sample. This lack of asignificant relation may also reflect differences in the vocabularymeasures used in each sample. Specifically, the vocabulary mea-sures used in the United States and Taiwan required children toknow the names of words, but the South Korean measure includedan additional component for children to define words. This differ-ence should be examined in future research.

Within all four samples, high behavioral regulation scores on theHTKS were most strongly related to high early mathematics com-pared with vocabulary or literacy. This fits with other findings inthe United States (Blair & Razza, 2007; G. J. Duncan, et al., 2007;Ponitz et al., 2009), and researchers have suggested that behavioralregulation may be particularly important for mathematics skillsbecause the HTKS and math assessments may require similarcognitive processes (Ponitz et al., 2009). Thus, the findings fromthe present study, and from previous research, suggest that theskills children need to be successful on the HTKS (attention,working memory, and inhibitory control) help children achieve inearly mathematics, even more so than in vocabulary or earlyliteracy.

Two issues regarding model specification warrant consideration.First, analyses in the present study examined behavioral regulationas a predictor of academic achievement, even though both mea-sures were collected concurrently. Conversely, it is possible thatchildren with better academic skills subsequently develop strongerbehavioral regulation. Previous research, however, supports a di-rectional relation and has found that early behavioral regulationsignificantly predicts later achievement (even after accounting forchild IQ; McClelland et al., 2006; McClelland et al., 2000). More-over, other research has found that early behavioral regulationsignificantly mediates relations between family risk factors andlater academic achievement (Sektnan, McClelland, Acock, & Mor-rison, 2010) and between emotion regulation and achievement(Howse et al., 2003). Together, this research supports a directionalrelation between behavioral regulation and achievement, includingthe view that children who can pay attention, remember instruc-tions, and inhibit certain behaviors in classroom settings are morelikely to take advantage of learning opportunities and do better onachievement tests. Although it is also possible that better behav-ioral regulation predicts stronger achievement, which then predicts

better behavioral regulation, our data do not allow us to test this inthe present study. This is an important avenue for future research.

Second, it is possible that a spurious variable, such as intelli-gence, accounts for the relation between HTKS scores and aca-demic achievement. In fact, general mental ability, as measured bypsychometric intelligence tests, is the strongest predictor to date ofacademic achievement for children of any age (Gottfredson, 2005).Previous research in the United States, however, has demonstratedthat aspects of behavioral regulation uniquely predicted academicachievement beyond the influence of intelligence (Blair, 2006;McClelland et al., 2006; McClelland et al., 2000). We suggest thatHTKS scores may positively relate to academic achievement overand beyond a measure of intelligence. Although we were not ableto directly test this in the current study, future research conductedwith the HTKS to assess behavioral regulation should include ameasure of intelligence.

Practical Implications

Early childhood professionals in the United States and SouthKorea have called for additional attention to be paid to the impor-tance of social and behavioral skills for school readiness (Kim etal., 2003; Rimm-Kaufman, Pianta, & Cox, 2000). Results from thepresent study support this assertion and suggest that in the UnitedStates, Taiwan, South Korea, and China—three countries that haveparticularly high academic outcomes—strong behavioral regula-tion in preschool predicts children’s early achievement. Thus,promoting the development of behavioral regulation in early child-hood settings in these societies may help children to be moresuccessful in school. This is especially relevant for children whostruggle with self-regulation, including an increasing number ofchildren in the United States (Gilliam & Shahar, 2006). This is alsorelevant in Asian societies, where although compliance is a cul-tural norm, there are variations among children and not all childrenexhibit strong self-regulation (Ahadi et al., 1993).

These results also suggest that promoting early behavioral reg-ulation in the United States, Taiwan, South Korea, and China mayhave significant benefits for young children. A number of behav-ioral regulation interventions used in the United States have dem-onstrated effectiveness for improving children’s skills. For exam-ple, the Tools of the Mind curriculum, which incorporatessociodramatic play, private speech, and drawing strategies thathelp children to pay attention, has increased children’s attention,working memory, and inhibitory control skills, and achievementskills (Barnett et al., 2008; Diamond, Barnett, Thomas, & Munro,2007). In addition, another recent study found that preschoolgames designed to help children practice attention, working mem-ory, and inhibitory control skills significantly improved children’sbehavioral regulation, especially for children low in these skills(Tominey & McClelland, 2008). Early childhood professionals inthe United States, Taiwan, South Korea, and China may be able tohelp children succeed in school by incorporating similar games,especially for children who have low behavioral regulation.

Limitations

The present study revealed a number of findings about the useof the HTKS for measuring behavioral regulation across foursocieties. There are some limitations, however, that should be

374 WANLESS ET AL.

noted. First, all of the participants in this study from Asian soci-eties lived in urban areas. Including children from multiple geo-graphic areas would provide a more accurate picture of the pres-ence and variability of behavioral regulation in these societies.Second, although there was overlap in child age across samples,and age was controlled for in the analyses, the age groups andranges were not the same. This limitation made it difficult tocompare behavioral regulation means, distributions, and floor andceiling effects across samples. In future research, age rangesshould be expanded in order to more clearly define the upper andlower age limits of the HTKS task in each society. Third, somevariables in some samples (mother’s education, early mathematics,vocabulary, and early literacy in South Korea; teacher-rated be-havioral regulation in the United States) had a relatively highamount of missing data, which suggests that results regardinganalyses using these variables may need to be interpreted withcaution. Comparison of means and standard deviations of thesevariables before and after imputation, however, suggests that over-all descriptive scores are relatively similar after missing data havebeen replaced. Fourth, teacher-rated classroom behavioral regula-tion data were collected in the United States, South Korea, andChina in the middle of the school year and in Taiwan at thebeginning of the school year. Future cross-cultural studies shouldalign the time of year that data are collected so that society effectsand timing effects can be disentangled. Finally, future studiesshould use more than one direct measure of behavioral regulationand teacher-rated measure within each society, and use academicmeasures with maximum cross-cultural consistency across societ-ies to better understand how much the assessment accounts fornuances in relations among these variables.

Conclusion

This preliminary study examined the psychometric properties ofscores on a direct measure of behavioral regulation, the HTKS, infour societies. Findings suggest that HTKS scores were reliable,captured variability, and predicted academic achievement in theU.S., Taiwan, South Korea, and China. Cultural differencesemerged in relations between HTKS directly measured behavioralregulation and teacher-rated classroom behavioral regulation.Overall, results suggest that HTKS scores may be reliable andvalidly interpreted as a reflection of behavioral regulation forpreschoolers in the United States, Taiwan, South Korea, andChina. These findings extend previous research on behavioralregulation in the United States and provide the foundation forcontinued research beyond the United States on behavioral regu-lation in high-achieving societies including Taiwan, South Korea,and China to help ensure that all children are successful in school.

References

Acock, A. C. (2005). Working with missing values. Journal of Marriageand Family, 67, 1012–1028. doi:10.1111/j.1741-3737.2005.00191.x

Adams, A.-M., Bourke, L., & Willis, C. (1999). Working memory andspoken language comprehension in young children. International Jour-nal of Psychology, 34, 364–378. doi:10.1080/002075999399701

Ahadi, S. A., Rothbart, M. K., & Ye, R. (1993). Children’s temperament inthe US and China: Similarities and differences. European Journal ofPersonality, 7, 359–378. doi:10.1002/per.2410070506

Archibald, S. J., & Kerns, K. A. (1999). Identification and description of

new tests of executive functioning in children. Child Neuropsychology,5, 115–129. doi:10.1076/chin.5.2.115.3167

Baldi, S., Jin, Y., Skemer, M., Green, P. J., & Herget, D. (2007). Highlightsfrom the PISA 2006: Performance of U.S. 15-year-old students insciences and mathematics literacy in an international context. Washing-ton, DC: U.S. Department of Education, National Center for EducationStatistics.

Bao, L. (2004). China. In J. J. Slater (Ed.), Teen life in Asia (pp. 19–33).Westport, CT: Greenwood Press.

Barnett, W. S., Jung, K., Yarosz, D. J., Thomas, J., Hornbeck, A., Stechuk,R., & Burns, S. (2008). Educational effects of the Tools of the Mindcurriculum: A randomized trial. Early Childhood Research Quarterly,23, 299–313. doi:10.1016/j.ecresq.2008.03.001

Blair, C. (2006). How similar are fluid cognition and general intelligence?A developmental neuroscience perspective on fluid cognition as anaspect of human cognitive ability. Behavioral and Brain Sciences, 29,109–125; discussion 125–160. doi:10.1017/S0140525X06009034

Blair, C., & Razza, R. P. (2007). Relating effortful control, executivefunction, and false belief understanding to emerging math and literacyability in kindergarten. Child Development, 78, 647–663. doi:10.1111/j.1467-8624.2007.01019.x

Bronson, M. B., Tivnan, T., & Seppannen, P. S. (1995). Relations betweenteacher and classroom activity variables and the classroom behaviors ofprekindergarten children in Chapter 1 funded programs. Journal ofApplied Developmental Psychology, 16, 253–282. doi:10.1016/0193-3973(95)90035-7

Carlson, S. M. (2005). Developmentally sensitive measures of executivefunction in preschool children. Developmental Neuropsychology, 28,595–616. doi:10.1207/s15326942dn2802_3

Checa, P., Rodrıguez-Bailon, R., & Rueda, M. (2008). Neurocognitive andtemperamental systems of self-regulation and early adolescents’ socialand academic outcomes. Mind, Brain, and Education, 2, 177–187.doi:10.1111/j.1751-228X.2008.00052.x

Choi, N.-Y., & Yi, S.-H. (2007). The effects of alphabet knowledge onKorean kindergartners’ reading of Hangul words. Journal of KoreanHome Management Association, 25, 151–168.

Chow, B., McBride-Chang, C., Cheung, H., & Chow, C. (2008). Dialogicreading and morphology training in Chinese children: Effects on lan-guage and literacy. Developmental Psychology, 44, 233–244. doi:10.1037/0012-1649.44.1.233

Cole, M. (1996). Cultural psychology: A once and future discipline.Cambridge, MA: Belknap Press of Harvard University Press.

Connor, C. M., Ponitz, C. C., Phillips, B. M., Travis, Q. M., Glasney, S.,& Morrison, F. J. (2010). First graders’ literacy and self-regulationgains: The effect of individualizing student instruction. Journal ofSchool Psychology, 48, 433–455.

Deater-Deckard, K., Mullineaux, P., Petrill, S., & Thompson, L. (2009).Effortful control, surgency, and reading skills in middle childhood.Reading and Writing, 22, 107–116. doi:10.1007/s11145-007-9111-9

Diamond, A., Barnett, W. S., Thomas, J., & Munro, S. (2007). Preschoolprogram improves cognitive control. Science, 318, 1387–1388. doi:10.1126/science.1151148

Dixon, W. E., Jr., & Salley, B. J. (2006). “Shhh! We’re tryin’ to concen-trate”: Attention and environmental distracters in novel word learning.Journal of Genetic Psychology, 167, 393– 414. doi:10.3200/GNTP.167.4.393-414

Dowsett, S. M., & Livesey, D. J. (2000). The development of inhibitorycontrol in preschool children: Effects of “executive skills” training.Developmental Psychobiology, 36, 161–174. doi:10.1002/(SICI)1098-2302(200003)36:2�161::AID-DEV7�3.0.CO;2-0

Duncan, G. J., Dowsett, C. J., Claessens, A., Magnuson, K., Huston, A. C.,Klebanov, P., & Crista, J. (2007). School readiness and later achieve-ment. Developmental Psychology, 43, 1428–1446. doi:10.1037/0012-1649.43.6.1428

375MEASURING BEHAVIORAL REGULATION IN FOUR SOCIETIES

Duncan, R., & McClelland, M. M. (2010). Relations among individualmeasures of self-regulation. Manuscript in preparation.

Dwyer, E. (2004). South Korea. In J. J. Slater (Ed.), Teen life in Asia (pp.205–222). Westport, CT: Greenwood Press.

Eisenberg, N., Sadovsky, A., & Spinrad, T. (2005). Associations ofemotion-related regulation with language skills, emotion knowledge,and academic outcomes. New Directions for Child and AdolescentDevelopment, 109, 109–118. doi:10.1002/cd.143

Entwisle, D. R., Alexander, K. L., & Olson, L. S. (2007). Early schooling:The handicap of being poor and male. Sociology of Education, 80,114–138. doi:10.1177/003804070708000202

Espy, K. A., & Bull, R. (2005). Inhibitory processes in young children andindividual variation in short-term memory. Developmental Neuropsy-chology, 28, 669–688. doi:10.1207/s15326942dn2802_6

Espy, K. A., McDiarmid, M. M., Cwik, M. F., Stalets, M. M., Hamby, A.,& Stern, T. E. (2004). The contribution of executive functions toemergent mathematic skills in preschool children. Developmental Neu-ropsychology, 26, 465–486. doi:10.1207/s15326942dn2601_6

Fahie, C. M., & Symons, D. K. (2003). Executive functioning and theoryof mind in children clinically referred for attention and behavior prob-lems. Journal of Applied Developmental Psychology, 24, 51–73. doi:10.1016/S0193-3973(03)00024-8

Gartstein, M. A., Gonzalez, C., Carranza, J. A., Ahadi, S. A., Ye, R.,Rothbart, M. K., . . . Yang, S. W. (2006). Studying cross-cultural dif-ferences in the development of infant temperament: People’s Republicof China, the United States of America, and Spain. Child Psychiatry andHuman Development, 37, 145–161. doi:10.1007/s10578-006-0025-6

Gathercole, S. E., & Pickering, S. J. (2000). Working memory deficits inchildren with low achievements in the national curriculum at 7 years ofage. British Journal of Educational Psychology, 70, 177–194. doi:10.1348/000709900158047

Gerstadt, C. L., Hong, Y. J., & Diamond, A. (1994). The relationshipbetween cognition and action: Performance of children 3.5–7 years oldon a Stroop-like day–night test. Cognition, 53, 129–153. doi:10.1016/0010-0277(94)90068-X

Gilliam, W. S., & Shahar, G. (2006). Preschool and child care expulsionand suspension rates and predictors in one state. Infants and YoungChildren, 19, 228–245. doi:10.1097/00001163-200607000-00007

Ginsburg, H. P., & Baroody, A. J. (1990). Test of early mathematics ability(2nd ed.). Austin, TX: Pro-Ed.

Gottfredson, L. (2005). Implications of cognitive differences for schoolingwithin diverse societies. In C. Frisby & C. R. Reynolds (Eds.), Com-prehensive handbook of multicultural school psychology (pp. 517–554).New York, NY: Wiley.

Heine, S. J., Lehman, D. R., Peng, K., & Greenholtz, J. (2002). What’swrong with cross-cultural comparisons of subjective Likert scales? Thereference-group effect. Journal of Personality and Social Psychology,82, 903–918. doi:10.1037/0022-3514.82.6.903

Howse, R. B., Calkins, S. D., Anastopoulos, A. D., Keane, S. P., & Shelton,T. L. (2003). Regulatory contributors to children’s kindergarten achieve-ment. Early Education and Development, 14, 101–120. doi:10.1207/s15566935eed1401_7

Hsieh, M.-F. (2004). Teaching practices in Taiwan’s education for youngchildren: Complexity and ambiguity of developmentally appropriatepractices and/or developmentally inappropriate practices. ContemporaryIssues in Early Childhood, 5, 309–329. doi:10.2304/ciec.2004.5.3.5

Hsu, H.-H. (1998). The concept of early learning mathematics seminarcollection. Paper presented at the Life, Games, Mathematics in YoungChildren meeting, Taipei, Taiwan.

Hsu, H.-H. (2000). Child mathematical aptitude test TEMA-2. Taipei,Taiwan: Psychological Publishing Company.

Huntsinger, C. S., Jose, P. E., Liaw, F.-R., & Ching, W.-D. (1997). Culturaldifferences in early mathematics learning: A comparison of Euro-American, Chinese-American, and Taiwan-Chinese families. Interna-

tional Journal of Behavioral Development, 21, 371–388. doi:10.1080/016502597384929

Jose, P. E., Huntsinger, C. S., Huntsinger, P. R., & Liaw, F.-R. (2000).Parental values and practices relevant to young children’s social devel-opment in Taiwan and the United States. Journal of Cross-CulturalPsychology, 31, 677–702. doi:10.1177/0022022100031006002

Kim, J., Lee, Y.-S., Suen, H., & Lee, G. S. (2003). A Delphi study of youngchildren’s readiness in Korea: Challenges and implications for earlychildhood schooling. Educational Research and Evaluation, 9, 345–355.doi:10.1076/edre.9.4.345.17811

Kline, R. B. (2005). Principles and practice of structural equation mod-eling (2nd ed.). New York, NY: Guilford Press.

Kochanska, G., Murray, K. T., & Harlan, E. T. (2000). Effortful control inearly childhood: Continuity and change, antecedents, and implicationsfor social development. Developmental Psychology, 36, 220–232. doi:10.1037/0012-1649.36.2.220

Kunter, M., & Baumert, J. (2006). Who is the expert? Construct andcriteria validity of student and teacher ratings of instruction. LearningEnvironments Research, 9, 231–251. doi:10.1007/s10984-006-9015-7

Lan, X., & Morrison, F. J. (2008, July). Intercorrelations among compo-nents of self-regulation and their relationship with academic outcomes.Paper presented at the meeting of the International Society for the Studyof Behavioral Development, Wurzburg, Germany.

Lan, X., Ponitz, C. C., Miller, K. F., Li, S., Cortina, K., Perry, M., & Fang,G. (2009). Keeping their attention: Classroom practices associated withbehavioral engagement in first grade mathematics classes in China andthe United States. Early Childhood Research Quarterly, 24, 198–211.doi:10.1016/j.ecresq.2009.03.002

Liew, J., McTigue, E., Barrois, L., & Hughes, J. (2008). Adaptive andeffortful control and academic self-efficacy beliefs on achievement: Alongitudinal study of 1st through 3rd graders. Early Childhood ResearchQuarterly, 23, 515–526. doi:10.1016/j.ecresq.2008.07.003

Liu, S. (2007). Early characteristics of number cognition of children withdevelopmental dyscalculia. Beijing Normal University, Beijing, China.

Loo, S. K., & Rapport, M. D. (1998). Ethnic variations in children’sproblem behaviors: A cross-sectional, developmental study of Hawaiischool children. Journal of Child Psychology and Psychiatry, 39, 567–575. doi:10.1017/S0021963098002261

Lu, L., & Liu, H. (1998). Xiu Ding Bi Pao De Tu Hua Ci Hui Ce Yan[Revised Peabody Picture Vocabulary Test: Mandarin Chinese Version].Taipei, Taiwan: Xin Li.

Mahone, E. M., Cirino, P. T., Cutting, L. E., Cerrone, P. M., Hagelthorn,K. M., Hiemenz, J. R., . . . Denckla, M. B. (2002). Validity of theBehavior Rating Inventory of Executive Function in children withADHD and/or Tourette syndrome. Archives of Clinical Neuropsychol-ogy, 17, 643–662.

Mahone, E. M., & Hoffman, J. (2007). Behavior rating of executivefunction among preschoolers with ADHD. The Clinical Neuropsychol-ogist, 21, 569–586.

Martini, M. (1996). ‘What’s new?’ at the dinner table: Family dynamics duringmealtimes in two cultural groups in Hawaii. Early Development and Par-enting, 5, 23–34. doi:10.1002/(SICI)1099-0917(199603)5:1�23::AID-EDP111�3.0.CO;2-D

Mashburn, A. J., Hamre, B. K., Downer, J. T., & Pianta, R. C. (2006).Teacher and classroom characteristics associated with teachers’ ratingsof prekindergartners’ relationships and behaviors. Journal of Psychoe-ducational Assessment, 24, 367–380. doi:10.1177/0734282906290594

Matthews, J. S., Ponitz, C. C., & Morrison, F. J. (2009). Early genderdifferences in self-regulation Journal of Educational Psychology, 101,689–704. doi:10.1037/a0014240

McClelland, M. M. (2009). Interrater reliability for the Head-to-ToesTask. Manuscript in preparation.

McClelland, M. M., Acock, A. C., & Morrison, F. J. (2006). The impact ofkindergarten learning-related skills on academic trajectories at the end of

376 WANLESS ET AL.

elementary school. Early Childhood Research Quarterly, 21, 471–490.doi:10.1016/j.ecresq.2006.09.003

McClelland, M. M., Cameron, C. E., Connor, C. M., Farris, C. L., Jewkes,A. M., & Morrison, F. J. (2007). Links between behavioral regulationand preschoolers’ literacy, vocabulary, and math skills. DevelopmentalPsychology, 43, 947–959. doi:10.1037/0012-1649.43.4.947

McClelland, M. M., & Morrison, F. J. (2003). The emergence of learning-related social skills in preschool children. Early Childhood ResearchQuarterly, 18, 206–224. doi:10.1016/S0885-2006(03)00026-7

McClelland, M. M., Morrison, F. J., & Holmes, D. L. (2000). Children atrisk for early academic problems: The role of learning-related socialskills. Early Childhood Research Quarterly, 15, 307–329. doi:10.1016/S0885-2006(00)00069-7

McClelland, M. M., Piccinin, A. M., & Stallings, M. C. (2010). Relationsbetween preschool attention and sociability and later achievement out-comes. Manuscript submitted for publication.

McClelland, M. M., & Ponitz, C. C. (2010). Self-regulation in earlychildhood: Improving conceptual clarity and developing ecologicallyvalid measures. Manuscript submitted for publication.

McClelland, M. M., Ponitz, C. C., Messersmith, E., & Tominey, S. L.(2010). Self-regulation: The integration of cognition and emotion. InR. M. Lerner (Series Ed.) & W. F. Overton (Vol. Ed.), Handbook oflifespan human development: Vol. 1. Cognition, biology and methods(pp. 509-553). Hoboken, NJ: Wiley.

McClelland, M. M., & Wanless, S. B. (2008, July). Touching your toes infour cultures: Developing a new measure of behavioral regulation foryoung children. Paper presented at the meeting of the InternationalSociety for the Study of Behavioural Development, Wurzburg, Ger-many.

McGrew, K. S., & Woodcock, R. W. (2001). Technical manual,Woodcock–Johnson III. Itasca, IL: Riverside.

McMullen, M., Elicker, J., Wang, J., Erdiller, Z., Lee, S.-M., Lin, C.-H., &Sun, P.-Y. (2005). Comparing beliefs about appropriate practice amongearly childhood education and care professionals from the U.S., China,Taiwan, Korea and Turkey. Early Childhood Research Quarterly, 20,451–464. doi:10.1016/j.ecresq.2005.10.005

Meng, X. L. (1995). Multiple imputation inferences with uncongenialsources of input. Statistical Sciences, 10, 538–573.

Muller, U., Dick, A. S., Gela, K., Overton, W. F., & Zelazo, P. D. (2006).The role of negative priming in preschoolers’ flexible rule use on theDimensional Change Card Sort task. Child Development, 77, 395–412.doi:10.1111/j.1467-8624.2006.00878.x

Mullis, I. V. S., Martin, M. O., Gonzalez, E. J., & Chrostowski, S. J.(2004). TIMSS 2003 international mathematics report. Chestnut Hill,MA: TIMSS & PIRLS International Study Center.

Munoz-Sandoval, A. F., Woodcock, R. W., McGrew, K. S., & Mather, N.(2005). Baterıa III: Woodcock–Munoz: Pruebas de aprovechamiento[Battery III: Woodcock–Munoz: Achievement Tests]. Itasca, IL: River-side.

Oh, S., & Lewis, C. (2008). Korean preschoolers’ advanced inhibitorycontrol and its relation to other executive skills and mental state under-standing. Child Development, 79, 80 –99. doi:10.1111/j.1467-8624.2007.01112.x

Pang, Y., & Richey, D. (2007). Preschool education in China and theUnited States: A personal perspective. Early Child Development andCare, 177, 1–13. doi:10.1080/14797580500252712

Park, H., Kwak, K., & Park, K. (1989). Lorean–Wechsler Preschool andPrimary Scale of Intelligence. Seoul, South Korea: Special Education.

Peugh, J. L. (2010). A practical guide to multilevel modeling. Journal ofSchool Psychology, 48, 85–112. doi:10.1016/j.jsp.2009.09.002

Pickering, S. J., & Gathercole, S. E. (2004). Distinctive working memoryprofiles in children with special educational needs. Educational Psychol-ogy, 24, 393–408. doi:10.1080/0144341042000211715

Ponitz, C. C., McClelland, M. M., Jewkes, A. M., Connor, C. M., Farris,

C. L., & Morrison, F. J. (2008). Touch your toes! Developing a directmeasure of behavioral regulation in early childhood. Early ChildhoodResearch Quarterly, 23, 141–158. doi:10.1016/j.ecresq.2007.01.004

Ponitz, C. C., McClelland, M. M., Matthews, J. S., & Morrison, F. J.(2009). A structured observation of behavioral self-regulation and itscontribution to kindergarten outcomes. Developmental Psychology, 45,605–619. doi:10.1037/a0015365

Raudenbush, S. W. (1994). Random effects models. In H. Cooper & L. V.Hedges (Eds.), The handbook of research synthesis (pp. 301–321). NewYork, NY: Russell Sage Foundation.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models:Applications and data analysis methods (2nd ed.). Thousand Oaks, CA:Sage.

Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., & Congdon, R. (2004).HLM 6.06: Hierarchical linear and nonlinear modeling. Lincolnwood,IL: Scientific Software International.

Rimm-Kaufman, S. E., Pianta, R. C., & Cox, M. J. (2000). Teachers’judgments of problems in the transition to kindergarten. Early ChildhoodResearch Quarterly, 15, 147–166. doi:10.1016/S0885-2006(00)00049-1

Rosenthal, R. (2002). The Pygmalion effect and its mediating mechanisms.In J. Aronson (Ed.), Improving academic achievement: Impact of psy-chological factors on education (pp. 25–36). San Diego, CA: AcademicPress. doi:10.1016/B978-012064455-1/50005-1

Rothbart, M. K., Ahadi, S. A., Hersey, K. L., & Fisher, P. (2001).Investigations of temperament at three to seven years: The Children’sBehavior Questionnaire. Child Development, 72, 1394 –1408. doi:10.1111/1467-8624.00355

Rothbart, M. K., & Posner, M. I. (2005). Genes and experience in thedevelopment of executive attention and effortful control. New Directionsfor Child and Adolescent Development, 109, 101–108. doi:10.1002/cd.142

Rothbart, M. K., & Sheese, B. (2007). Temperament and emotion regula-tion. In J. J. Gross (Ed.), Handbook of emotion regulation (pp. 331–350).New York, NY: Guilford Press.

Rubin, D. B. (1996). Multiple imputation after 18� years. Journal of theAmerican Statistical Association, 91, 473–489. doi:10.2307/2291635.

Rueda, M. R., Posner, M. I., & Rothbart, M. K. (2005). The developmentof executive attention: Contributions to the emergence of self-regulation.Developmental Neuropsychology, 28, 573–594. doi:10.1207/s15326942dn2802_2

Schafer, J. L., & Graham, J. (2002). Missing data: Our view of the state ofthe art. Psychological Methods, 7, 147–177. doi:10.1037/1082-989X.7.2.147

Sektnan, M., McClelland, M. M., Acock, A., & Morrison, F. J. (2010).Relations between early family risk, children’s behavioral regulation,and academic achievement. Early Childhood Research Quarterly, 25,464–479.

Senn, T. E., Espy, K. A., & Kaufmann, P. M. (2004). Using path analysisto understand executive function organization in preschool children.Developmental Neuropsychology, 26, 445– 464. doi:10.1207/s15326942dn2601_5

Shweder, R. A., Goodnow, J. J., Hatano, G., LeVine, R. A., Markus, H. R.,& Miller, P. J. (1998). The cultural psychology of development: Onemind, many mentalities. In R. M. Lerner (Ed.), Handbook of childpsychology: Vol. 1. Theoretical models of human development (5th ed.,pp. 865–937). New York, NY: Wiley.

Simpson, A., & Riggs, K. J. (2006). Conditions under which childrenexperience inhibitory difficulty with a “button-press” go/no-go task.Journal of Experimental Child Psychology, 94, 18–26. doi:10.1016/j.jecp.2005.10.003

StataCorp. (2007). Stata statistical software: Release 10. College Station,TX: Author.

Stevenson, H. W., Chen, C., & Lee, S.-Y. (1993). Mathematics achieve-

377MEASURING BEHAVIORAL REGULATION IN FOUR SOCIETIES

ment of Chinese, Japanese, and American children: Ten years later.Science, 259, 53–58. doi:10.1126/science.8418494

Stevenson, H. W., Lee, S.-Y., Chen, C., Lummis, M., Stigler, J. W., Fan,L., & Ge, F. (1990). Mathematics achievement of children in China andthe United States. Child Development, 61, 1053–1066. doi:10.2307/1130875

Stevenson, H. W., & Stigler, J. W. (1994). The learning gap: Why ourschools are failing and what we can learn from Japanese and Chineseeducation. New York, NY: Simon & Schuster.

Tominey, S., & McClelland, M. M. (2008). Red light, purple light: Initialfindings from an intervention to improve self-regulation over the pre-kindergarten year. Manuscript in preparation.

Tsai, H.-L. A., McClelland, M. M., Pratt, C., & Squires, J. (2006). Adap-tation of the 36-Month Ages and Stages Questionnaire in Taiwan:Results from a preliminary study. Journal of Early Intervention, 28,213–225. doi:10.1177/105381510602800308

van der Schoot, M., Licht, R., Horsley, T. M., Aarts, L. T., van Koert, B.,& Sergeant, J. A. (2004). Inhibitory control during sentence reading indyslexic children. Child Neuropsychology, 10, 173–188.

von Aster, M. (2001). Die Neuropsychologische Testbatterie fur Zahlen-verarbeitung und Rechnen bei Kindern (ZAREKI) [NeuropsychologicalTest Battery for Number Processing and Calculation in Children]. Frank-furt, Germany: Swets & Zeitlinger.

Wall, S. M., & Paradise, L. V. (1981). A comparison of parent and teacher

reports of selected adaptive behaviors of children. Journal of SchoolPsychology, 19, 73–77. doi:10.1016/0022-4405(81)90011-X

Wanless, S. B., McClelland, M. M., Tominey, S. L., & Acock, A. C. (inpress). The influence of demographic risk factors on children’s behav-ioral regulation in prekindergarten and kindergarten. Early Educationand Development.

Woodcock, R. W., & Johnson, M. B. (1989). WJ–R tests of cognitiveability. Itasca, IL: Riverside.

Woodcock, R. W., & Mather, N. (2000). Woodcock Johnson Psycho-Educational Battery–III. Itasca, IL: Riverside.

Woodcock, R. W., & Munoz-Sandoval, A. F. (1996). Baterıa Woodcock–Munoz–R: Pruebas de aprovechamiento–Revisada [Woodcock–Munoz–R Battery: Achievement Tests–Revised] [Supplemental manual].Itasca, IL: Riverside.

Yi, C.-C., & Wu, C.-I. (2004). Taiwan. In J. J. Slater (Ed.), Teen life in Asia(pp. 223–241). Westport, CT: Greenwood Press.

Zhou, Q., Main, A., & Wang, Y. (2010). The relations of temperamentaleffortful control and anger/frustration to Chinese children’s academicachievement and social adjustment: A longitudinal study. Journal ofEducational Psychology, 102, 180–196. doi:10.1037/a0015908

Received August 31, 2009Revision received July 20, 2010

Accepted August 4, 2010 �

378 WANLESS ET AL.