James C. DiPerna

Kindergarten Predictors of Mathematical Growth in the Primary Grades: An Investigation Using the Early Childhood Longitudinal Study--Kindergarten Cohort.

This study examined longitudinal predictive relationships between young childrens classroom behaviors and their growth in mathematics skills during the primary grades. Using the Early Childhood Longitudinal Study--Kindergarten cohort, the authors tested a longitudinal model featuring positive (interpersonal skills and approaches to learning) and negative (internalizing and externalizing) behaviors at kindergarten entry as predictors of growth in mathematics achievement through 3rd grade. Results indicated that negative behaviors demonstrate negligible relationships with early development of mathematics skills. Interpersonal skills demonstrated a small negative relationship with mathematics growth when other positive behaviors were included as predictors in the model. In contrast, approaches to learning (goal-directed behavior, persistence, organization) demonstrated small positive relationships with growth in mathematical skills and may represent a skill domain for educators to consider in designing their instructional practices.

Developing Short Forms of the EARLI Numeracy Measures: Comparison of Item Selection Methods

Currently, few measures are available to monitor young children’s progress in acquiring key early academic skills. In response to this need, the authors have begun developing measures (i.e., the Early Arithmetic, Reading and Learning Indicators, or EARLI) of preschoolers’ numeracy skills. To accurately and efficiently monitor acquisition of early skills, users require multiple short forms that are appropriate in difficulty level for young children at different points in time. In the current study, the authors used three item selection methods to create multiple short forms for two EARLI numeracy measures. The authors then compared these item selection methods on projected internal consistency and concurrent validity estimates for the resulting forms. The short forms selected by these methods did not differ significantly on either criterion and appeared to be sufficiently sensitive to measure initial and acquisition of numeracy skills over time by preschool children enrolled in Head Start.

Confirmatory factor analysis of the Early Arithmetic, Reading, and Learning Indicators (EARLI).

Despite growing interest in early intervention, there are few measures available to monitor the progress of early academic skills in preschoolers. The Early Arithmetic, Reading, and Learning Indicators (EARLI; DiPerna, Morgan, & Lei, 2007) were developed as brief assessments of critical early literacy and numeracy skills. The purpose of the current study was to examine the factor structure of the EARLI probes via confirmatory factor analysis (CFA) in a sample of Head Start preschoolers (N=289). A two-factor model with correlated error terms and a bifactor model provided comparable fit to the data, although there were some structural problems with the latter model. The utility of the bifactor model for explaining the structure of early academic skills as well as the utility of the EARLI probes as measures of literacy and numeracy skills in preschool are discussed.

Examining early literacy skill differences among children in Head Start via latent profile analysis.

The purpose of the present study was to determine whether there are systematic differences in literacy skills among children from less-advantaged households, using latent profile analysis. Early reading skills were measured using the Early Arithmetic, Reading, and Learning Indicators (EARLI; DiPerna, Morgan, & Lei, 2007) literacy tasks. Participants (N = 166) were 4-year-old children enrolled in Head Start. Results revealed three classifications of children: Profile 1 had the lowest overall literacy skill levels, with relative strength in expressive vocabulary and weaknesses in skills related to letter knowledge (i.e., Alphabet Recitation, Letter Naming); Profile 2 had the highest overall level of literacy skills and a relative strength on the Segmenting task; and Profile 3 showed the greatest variability across the EARLI probe scores, with a relative strength in Alphabet Recitation and weaknesses on tasks measuring phonemic and phonological awareness (i.e., Letter Sounds, Sound Deletion, and Segmenting). Repeated measures ANOVA revealed that, although students in each profile demonstrated skill growth over a 6-month period, significant skill differences still remained between profiles at the end of the preschool year.

Measuring Quality in Kindergarten Classrooms: Structural Analysis of the Classroom Assessment Scoring System (CLASS K-3).

Research Findings: The purpose of the current study was to evaluate the structural validity of scores on a measure of global classroom quality, the Classroom Assessment Scoring System (CLASS K–3; Pianta, La Paro, & Hamre, 2008). Using observational data from a sample of 417 kindergarten classrooms from the southern and mid-Atlantic regions of the United States, we used confirmatory factor analysis to examine the structural validity of the CLASS K–3. Factor analytic findings supported a 3-factor and 10-dimension structure for the CLASS K–3; however, some modifications were made to the original CLASS model. Practice or Policy: Although the overall structure of the CLASS has been generally consistent across validation studies, some facets of the model may be less stable than others. Additional examination of alternative factor structures is needed to further clarify the relationships among the CLASS dimensions and domains. Current psychometric evidence provides support for continued use of the CLASS to guide intervention, instruction, and professional development.

A Cluster Randomized Trial of the Social Skills Improvement System-Classwide Intervention Program (SSIS-CIP) in First Grade.

The purpose of this study was to evaluate the efficacy of a universal social skills program, the Social Skills Improvement System Classwide Intervention Program (SSIS-CIP; Elliott & Gresham, 2007), for students in first grade. Classrooms from 6 elementary schools were randomly assigned to treatment or business-as-usual control conditions. Teachers assigned to the treatment condition implemented the SSIS-CIP over a 12-week period. Students’ social skills, problem behaviors, and approaches to learning were assessed via teacher ratings and direct observations of classroom behavior. In addition, their early literacy and numeracy skills were measured via computer-adaptive standardized tests. SSIS-CIP participation yielded small positive effects in students’ social skills (particularly empathy and social engagement) and approaches to learning (academic motivation and engagement). Students’ problem behaviors and academic skills, however, were unaffected by SSIS-CIP exposure.

Maximizing measurement efficiency of behavior rating scales using Item Response Theory: An example with the Social Skills Improvement System — Teacher Rating Scale

Measurement efficiency is an important consideration when developing behavior rating scales for use in research and practice. Although most published scales have been developed within a Classical Test Theory (CTT) framework, Item Response Theory (IRT) offers several advantages for developing scales that maximize measurement efficiency. The current study provides an example of using IRT to maximize rating scale efficiency with the Social Skills Improvement System - Teacher Rating Scale (SSIS - TRS), a measure of student social skills frequently used in practice and research. Based on IRT analyses, 27 items from the Social Skills subscales and 14 items from the Problem Behavior subscales of the SSIS - TRS were identified as maximally efficient. In addition to maintaining similar content coverage to the published version, these sets of maximally efficient items demonstrated similar psychometric properties to the published SSIS - TRS.

Childhood Social Skills as Predictors of Middle School Academic Adjustment

The present study used multiple regression analyses to examine the relationships between fifth-grade social skills and eighth-grade academic achievement. Data were drawn from the Early Childhood Longitudinal Study–Kindergarten Class of 1998-1999 (ECLS-K). Results indicated no relationship between positive or negative social behavior in fifth grade and academic achievement or teacher-rated academic skills in eighth grade. However, consistent with previous studies, fifth-grade approaches to learning were found to be positive predictors of both academic achievement and teacher-rated academic skills in eighth grade. In addition, these results suggest that socioeconomic status plays a significant and potentially unexplored avenue for understanding these outcomes. These results further illuminate the way behaviors in elementary school relate to academic adjustment to middle school.

Structural Validity of CLASS K-3 in Primary Grades: Testing Alternative Models.

This study examined the internal structure of the Classroom Assessment Scoring System (CLASS; K–3 version). The original CLASS K–3 model (Pianta, La Paro, & Hamre, 2008) and 5 alternative models were tested using confirmatory factor analysis with a sample of first- and second-grade classrooms (N = 141). Findings indicated that a slightly modified version of the original CLASS K–3 3-factor model best fit the current data. Although stable findings emerged across the current and previous studies, particularly in relation to the presence of 3 latent domains, there is also some variability across structures at different grade levels with regard to the bifactor and 3-factor models.

Student and Teacher Ratings of Academic Competence: An Examination of Cross-Informant Agreement

Two studies were conducted with samples of middle and high school teachers and students to examine cross-informant agreement on the Academic Competence Evaluation Scales. Cross-informant agreement was examined using Pearson correlations and conditional probability indices. Results of Study 1 (N = 65) and Study 2 (N = 66) indicated that teacher and student ratings shared more variance than typically reported in prior cross-informant studies of problem or social behavior. Although teacher and student ratings shared strong correlations, there was significant disagreement when scores were dichotomized for decision making. Implications of the results for school psychology practice and future research are discussed.