Understanding the development of healthy lifestyle behaviors as well as its early predictors and associated outcomes is paramount, given the importance of healthy lifestyle behaviors in academic success, health, and well-being. In this dissertation, I conducted two studies to investigate the development of three modifiable lifestyle behaviors (i.e., physical activity, sedentary behavior, and sleep duration) in middle childhood, using a nationally representative sample of US children from the Early Childhood Longitudinal Study, Kindergarten Class of 2010-11. Study 1 was designed to unpack the developmental associations among the three lifestyle behaviors by examining the transactional within-person associations and co-development. Moderate-to-vigorous physical activity (MVPA), video game use (as a measure of mentally active sedentary behavior), and sleep duration were reported by parents in 3rd through 5th grades. Results from the random intercept cross-lagged panel model showed that MVPA, video game use, and sleep duration were not transactionally related over time at the within-person level. Analysis with the multivariate latent growth curve model showed that changes in these lifestyle behaviors over time were not correlated either. These main findings were replicated across social contexts of socioeconomic status, gender, and race/ethnicity groups. Study 2 focused on investigating predictors and outcomes associated with developmental trajectories of the three lifestyle behaviors. Specifically, kindergarten self-regulation (rated by teachers) that reflects individual agency was examined as a predictor. Fifth grade externalizing and internalizing problems (rated by teachers), body mass index (assessed with weight and height), and academic achievement (assessed by reading, math, and science achievement) were included as mental health, physical health, and academic outcomes, respectively. Results showed that kindergarten self-regulation predicted initial levels (i.e., 3rd grade) of video game use and sleep duration. After controlling for the stability and other lifestyle behaviors, 5th grade externalizing problems and academic achievement were predicted by initial levels of sleep duration, and body mass index was predicted by initial levels of sleep duration and MVPA and changes in MVPA. Most of these findings were replicated in subsamples across social contexts. Overall, this dissertation provides important insight into the development of healthy lifestyle behaviors in US children.

The overarching goal of this dissertation was to contribute to the field’s understanding of young children’s development of ethnic-racial identification. In particular, Study 1 presented the adaptation of three measures that are developmentally appropriate for assessing young children’s ethnic-racial attitudes, ethnic-racial centrality, and ethnic-racial knowledge, and tested the psychometric properties of each measure. Findings from Study 1 provided limited initial support for the construct validity and reliability of the measures; importantly, there were many differences in the descriptives and measurement properties based on the language in which children completed the measures. In addition to measurement of ethnic-racial identification, Study 2 used the measures developed in Study 1 and tested whether Mexican-origin mothers’ adaptive cultural characteristics (i.e., ERI affirmation, ethnic-racial centrality, and involvement in Mexican culture) when children were 3 years of age predicted greater cultural socialization efforts with children at 4 years of age and, in turn, children’s ethnic-racial identification (i.e., children’s ethnic-racial attitudes, ethnic-racial centrality, ethnic-racial knowledge, and identification as Mexican) at 5 years of age. Furthermore, children’s characteristics (i.e., gender and skin tone) were tested as moderators of these processes. Findings supported expected processes from mothers’ adaptive cultural characteristics to children’s ethnic-racial identification via mothers’ cultural socialization across boys and girls, however, relations varied by children’s skin tone. Findings highlight the important role of children’s individual characteristics in cultural socialization and young children’s developing ethnic-racial identification over time. Overall, given the paucity of studies that have examined ethnic-racial identification among young children, the results from Study 1 and Study 2 have the potential to stimulate growth of knowledge in this area.

Dimensionality assessment is an important component of evaluating item response data. Existing approaches to evaluating common assumptions of unidimensionality, such as DIMTEST (Nandakumar & Stout, 1993; Stout, 1987; Stout, Froelich, & Gao, 2001), have been shown to work well under large-scale assessment conditions (e.g., large sample sizes and item pools; see e.g., Froelich & Habing, 2007). It remains to be seen how such procedures perform in the context of small-scale assessments characterized by relatively small sample sizes and/or short tests. The fact that some procedures come with minimum allowable values for characteristics of the data, such as the number of items, may even render them unusable for some small-scale assessments. Other measures designed to assess dimensionality do not come with such limitations and, as such, may perform better under conditions that do not lend themselves to evaluation via statistics that rely on asymptotic theory. The current work aimed to evaluate the performance of one such metric, the standardized generalized dimensionality discrepancy measure (SGDDM; Levy & Svetina, 2011; Levy, Xu, Yel, & Svetina, 2012), under both large- and small-scale testing conditions. A Monte Carlo study was conducted to compare the performance of DIMTEST and the SGDDM statistic in terms of evaluating assumptions of unidimensionality in item response data under a variety of conditions, with an emphasis on the examination of these procedures in small-scale assessments. Similar to previous research, increases in either test length or sample size resulted in increased power. The DIMTEST procedure appeared to be a conservative test of the null hypothesis of unidimensionality. The SGDDM statistic exhibited rejection rates near the nominal rate of .05 under unidimensional conditions, though the reliability of these results may have been less than optimal due to high sampling variability resulting from a relatively limited number of replications. Power values were at or near 1.0 for many of the multidimensional conditions. It was only when the sample size was reduced to N = 100 that the two approaches diverged in performance. Results suggested that both procedures may be appropriate for sample sizes as low as N = 250 and tests as short as J = 12 (SGDDM) or J = 19 (DIMTEST). When used as a diagnostic tool, SGDDM may be appropriate with as few as N = 100 cases combined with J = 12 items. The study was somewhat limited in that it did not include any complex factorial designs, nor were the strength of item discrimination parameters or correlation between factors manipulated. It is recommended that further research be conducted with the inclusion of these factors, as well as an increase in the number of replications when using the SGDDM procedure.

Differentiating bilingual children with primary language impairment (PLI) from those with typical development in the process of learning a second language has been a challenge. Studies have focused on improving the diagnostic accuracy of language measures for bilinguals. However, researchers are faced with two main challenges when estimating the diagnostic accuracy of new measures: (a) using an a priori diagnosis of children (children with and without PLI), as a reference may introduce error given there is no gold standard for the a priori classification; and (b) classifying children into only two groups may be another source of error given evidence that there may be more than two language ability groups with different strengths and weaknesses or, alternatively, a single group characterized by a continuum of language performance. The current study tested for the number of distinct language ability groups and their characteristics in predominately Spanish-speaking children in the U.S. without using an a priori classification as a reference. In addition, the study examined to what extent the latent groups differed on each measure, and the stability of language ability groups across three assessment methods in Spanish (standardized tests, language sample analyses, and comprehensive assessment), taking in to account English and non-verbal cognitive skills. The study included 431 bilingual children attending English-only education. Three latent profile analyses were conducted, one for each method of assessment. Results suggested more than two distinct language ability groups in the population with the method of assessment influencing the number and characteristics of the groups. Specifically, four groups were estimated based on the comprehensive assessment, and three based on standardized assessment or language sample analysis in Spanish. The stability of the groups was high on average, particularly between the comprehensive assessment and the standardized measures. Results indicate that an a priori classification of children into two groups, those with and without PLI, could lead to misclassification, depending on the measures used.

The purpose of this mixed-methods study was to understand the key constructs and processes underlying the mentoring relationships between doctoral students and their mentors. First, exploratory and confirmatory factor analyses were used to evaluate the measurement structure underlying the 34-item Ideal Mentor Scale (IMS; Rose, 2003), followed by an examination of factorial invariance and differences in latent means between graduate students differing by gender, age, and Master's vs. Doctoral status. The IMS was administered to 1,187 graduate students from various departments across the university at Arizona State University (ASU); this sample was split into two independent samples. Exploratory factory analysis on Sample 1 (N = 607) suggested a new four-factor mentoring model consisting of Affective Advocacy, Academic Guidance, Scholarly Example, and Personal Relationship. Subsequent confirmatory factor analysis on Sample 2 (N = 580) found that this four-factor solution was superior to the fit of a previously hypothesized three-factor model including Integrity, Guidance, and Relationship factors (Rose, 2003). Latent mean differences were evaluated for the four-factor model using structured means modeling. Results showed that females placed more value on factors relating to Affective Advocacy, Academic Guidance, and Scholarly Example, and less value on Personal Relationship than males. Students 30 and older placed less value on Scholarly Example and Personal Relationship than students under 30. There were no significant differences in means for graduate students pursuing a Master's versus a Doctoral degree. iii Further study qualitatively examined mentoring relationships between doctoral students and their faculty mentor using the Questionnaire on Supervisor Doctoral Student Interaction (QSDI) coupled with semi-structured interviews. Graduate support staff were interviewed to gather data on program characteristics and to provide additional context. Data were analyzed using Erickson's Modified Analytical Inductive method (Erickson, 1986). Findings showed that the doctoral students valued guidance, advocacy and constructive, timely feedback but realized the need to practice self-reliance to complete. Peer mentoring was important. Most of the participants valued a mentor's advocacy and longed to co-publish with their advisor. All students valued intellectual freedom, but wished for more direction to facilitate timelier completion of the degree. Development of the scholarly identity received little or no overt attention.

Designing studies that use latent growth modeling to investigate change over time calls for optimal approaches for conducting power analysis for a priori determination of required sample size. This investigation (1) studied the impacts of variations in specified parameters, design features, and model misspecification in simulation-based power analyses and (2) compared power estimates across three common power analysis techniques: the Monte Carlo method; the Satorra-Saris method; and the method developed by MacCallum, Browne, and Cai (MBC). Choice of sample size, effect size, and slope variance parameters markedly influenced power estimates; however, level-1 error variance and number of repeated measures (3 vs. 6) when study length was held constant had little impact on resulting power. Under some conditions, having a moderate versus small effect size or using a sample size of 800 versus 200 increased power by approximately .40, and a slope variance of 10 versus 20 increased power by up to .24. Decreasing error variance from 100 to 50, however, increased power by no more than .09 and increasing measurement occasions from 3 to 6 increased power by no more than .04. Misspecification in level-1 error structure had little influence on power, whereas misspecifying the form of the growth model as linear rather than quadratic dramatically reduced power for detecting differences in slopes. Additionally, power estimates based on the Monte Carlo and Satorra-Saris techniques never differed by more than .03, even with small sample sizes, whereas power estimates for the MBC technique appeared quite discrepant from the other two techniques. Results suggest the choice between using the Satorra-Saris or Monte Carlo technique in a priori power analyses for slope differences in latent growth models is a matter of preference, although features such as missing data can only be considered within the Monte Carlo approach. Further, researchers conducting power analyses for slope differences in latent growth models should pay greatest attention to estimating slope difference, slope variance, and sample size. Arguments are also made for examining model-implied covariance matrices based on estimated parameters and graphic depictions of slope variance to help ensure parameter estimates are reasonable in a priori power analysis.