Three-level Hierarchical Linear Model Research Articles

Exploring the predictors of teachers’ teaching autonomy: A three-level international study

Concerning the critical function of teachers' teaching autonomy in instructional practice and professional development, this study explores the relative importance of three predictive mechanisms: a. professionalism of teachers, b. complexity of teaching practice, and c. culture of individualism. Using the sample of 118,347 teachers from 8262 schools in 27 countries derived from the 2018 Teaching and Learning International Survey (TALIS), we applied a three-level hierarchical linear model (HLM) for analysis. We found that teachers’ teaching autonomy was mainly predicted by the professionalism of teachers and the culture of individualism rather than the complexity of teaching practice. Implications for individual teachers, school leaders, and policymakers are discussed.

Do Municipal Contexts Matter for Adolescent Mental Health? A Within-Municipality Analysis of Nationwide Norwegian Survey Data Across Six Years.

Despite growing concerns about substantial socio-economic differences between districts in many developed nations, limited attention has been paid to how adolescent mental health may be shaped by district characteristics. A few studies have shown that adolescent mental health is related to contextual factors such as district socio-economic status, neighborhood disorder, and quality of infrastructure. However, prior estimates may be an artifact of unmeasured differences between districts. To address these concerns, we used data from the nationwide Norwegian Ungdata surveys (N = 278,764), conducted across the years 2014 to 2019. We applied three-level hierarchical linear models to examine within-municipality associations between municipal factors and adolescent mental health in the domains of internalizing problems (i.e., depressive symptoms), externalizing problems (i.e., behavioral problems), and well-being (i.e., self-esteem), thereby accounting for all time-invariant municipality-level confounders. Our results showed that municipal-level safety, infrastructure, and youth culture are associated with adolescent mental health problems. Further, cross-level interaction models indicated gender-specific associations, with stronger associations of municipality infrastructure and community belongingness with increased self-esteem and reduced delinquent behaviors among girls than boys. Our findings highlight that municipality-level interventions may be a feasible strategy for adolescent mental health, even in a society characterized by low inequality and high redistribution.

Implementation and effectiveness of a sustained depression prevention program for high school students in Japan

ABSTRACT In Japan, many high school students have severe depression. Preventive interventions for school-based cognitive-behavioral depression are effective countermeasures against depressive problems among high school students. However, since high school curriculums are time-consuming, such interventions are less common among high school students. Therefore, we conducted a long-term but low-frequency intervention that divides the program content through several grades. We examined the effectiveness of our intervention based on the data obtained by a survey during the two-year intervention period and one-year follow-up period for 225 students who had enrolled in 2014–2019. Three-level hierarchical linear model analysis showed that students who participated in the program had improved depressive symptoms and related factors over these three years. The Welch test and chi-square test showed fewer depressive symptoms and a lower percentage of students with depression than the 608 Japanese students who did not receive the intervention; data were collected through an online survey at the end of the program and one year after the intervention ended. The results suggest that a school-based cognitive-behavioral depression prevention program using a long-term but low-frequency approach may effectively reduce the implementation burden into the school curriculum, promote sustained implementation, and prevent depression among high school students.

Collaboration and socio-economic inequality: Estimating the effects of intra-school and inter-school inequality on collaborative problem-solving skills

The Programme for International Student Assessment (PISA), the world’s most extensive assessment of student scholastic abilities, included in its 2015 cycle a human-agent measure of collaborative problem-solving skills (CPS). To better understand the development of collaboration proficiency in youth, we assess the effects on CPS of several individual, mezzo, and national level variables, focusing on intra-school and inter-school wealth inequality. The results of the three-level hierarchical linear models on 46 countries show that both types of inequality have negative effects on CPS individual scores. In the case of intra-school inequality, the negative effect holds after controlling for individual differences in science performance. Generalized trust and political equality, measured at the national level, counteract these effects independently, reaffirming the importance of education to democratic citizenship and institutions.

Race/Ethnicity, Attachment, and the Cortisol Awakening Response among Young Sexual Minority Men

Young sexual minority men (YSMM), particularly YSMM of color, are more likely to have dysregulated physiological stress responses compared to heterosexuals. Moreover, the quality of social relationships impacts physiological stress patterns among young adults. One key indicator of social relationships is that of adult attachment, which refers to the types of socioemotional bonds that individuals form with peers, family, and romantic partners. However, the association between adult attachment and physiological stress functioning among diverse samples of YSMM has been underexplored. Thus, the current study sought to bridge this gap in the literature by exploring how race/ethnicity moderates the association between adult attachment and physiological stress patterns among Black versus White YSMM. N = 63 YSMM participated in a 5-day daily diary study in which they completed a baseline survey and provided saliva samples over the 5-day period in order to measure diurnal cortisol. Three-level hierarchical linear modeling was used in order to examine the association between race/ethnicity, adult attachment, and diurnal cortisol over the 5-day period. Results suggest that adult attachment avoidance was associated with the cortisol awakening response (CAR) among YSMM. In addition, race/ethnicity moderated the association between adult attachment avoidance and the CAR such that YSMM who identified as Black had evidence of a more dysregulated physiological stress response as compared to YSMM who identified as White. Future research should focus on further unpacking the drivers of physiological stress patterns among White and Black YSMM.

Investigating the Effects and Social Validity of an Informational Text Structure Intervention for Reading and Writing in Grades Four and Five

The purpose of this study was to investigate the effects and social validity of Read STOP Write, a text structure intervention for reading and writing, in grades four and five. In a cluster randomized trial, 11 teachers in three elementary schools were randomly assigned to deliver Read STOP Write or RARE Reading & Writing, an alternative treatment focused on question answering. In total, 160 students received Read STOP Write, and 191 students received RARE Reading & Writing. Data included pretest and posttest measures of students’ informational text structure identification, reading comprehension, and writing quality, and interviews to understand teachers’ perceptions of the social validity of the interventions. Student measures were analyzed using three-level hierarchical linear modeling. Teacher interviews were analyzed qualitatively using typological analysis. Results indicated that students who received Read STOP Write outperformed students who received RARE Reading & Writing on a researcher-developed measure of informational writing quality with promising effects on measures of informational text structure identification and reading comprehension skills but no effects on standardized measures of general reading comprehension or writing quality. Teacher interviews revealed that teachers perceived the goals, procedures, and effects of both interventions as socially valid. Implications for research and practice are discussed.

Teaching quality and student learning achievements in Ethiopian primary education: How effective is instructional quality in closing socioeconomic learning achievement inequalities?

Teaching quality affects student learning achievements, and high socioeconomic status (SES) students usually outperform low SES ones. However, teaching quality may affect students differently depending on their SES backgrounds. This study employs three-level hierarchical linear modeling with Ethiopian Young Lives data to examine the potential differential effects of the three basic dimensions of teaching quality (teacher support, classroom management, and cognitive activation) across student SES groups. Our findings reveal that, in general, the three dimensions of teaching quality are positively associated with learning achievements. However, compared to low and middle SES background students, high SES background students benefit from classroom management and cognitive activation. This suggests that teaching quality (classroom management and cognitive activation) may contribute to widening inequalities in learning achievements. Implications of the findings for policy and practice are discussed.

Association of organizational culture and climate with variation in the clinical outcomes of collaborative care for maternal depression in community health centers.

Organizational factors may help explain variation in the effectiveness of evidence-based clinical innovations through implementation and sustainment. This study tested the relationship between organizational culture and climate and variation in clinical outcomes of the Collaborative Care Model (CoCM) for treatment of maternal depression implemented in community health centers. Organizational cultures and climates of 10 community health centers providing CoCM for depression among low-income women pregnant or parenting were assessed using the organizational social context (OSC) measure. Three-level hierarchical linear models tested whether variation in culture and climate predicted variation in improvement in depression symptoms from baseline to 6.5-month post-baseline for N = 468 women with care ±1 year of OSC assessment. Depression symptomology was measured using the Patient Health Questionnaire (PHQ-9). After controlling for patient characteristics, case mix, center size, and implementation support, patients served by centers with more proficient cultures improved significantly more from baseline to 6.5-month post-baseline than patients in centers with less proficient cultures (mean improvement = 5.08 vs. 0.14, respectively, p = .020), resulting in a large adjusted effect size of dadj = 0.78. A similar effect was observed for patients served by centers with more functional climates (mean improvement = 5.25 vs. 1.12, p < .044, dadj = 0.65). Growth models indicated that patients from all centers recovered on average after 4 months of care. However, those with more proficient cultures remained stabilized whereas patients served by centers with less proficient cultures deteriorated by 6.5-month post-baseline. A similar pattern was observed for functional climate. Variation in clinical outcomes for women from historically underserved populations receiving Collaborative Care for maternal depression was associated with the organizational cultures and climates of community health centers. Implementation strategies targeting culture and climate may improve the implementation and effectiveness of integrated behavioral health care for depression.

Sense of School Belonging: The Illusion of a School-Size Effect

Over the past thirty years, the sense of school belonging received increasing attention in the educational field because of the many developmental and educational advantages associated with it. Previous research suggested that small schools are in a better position to create a strong sense of belonging among students. The aim of our study was to test the effect of school size on students’ sense of school belonging and to test whether this effect holds after considering the effect of the classroom after controlling for individual differences. The study was conducted on a sample of 1698 sixth- and seventh-grade students distributed among 20 school institutes in Switzerland. Multilevel analyses have been conducted. While first results showed an effect of the school size on the sense of school belonging, a comparison of the two-level and three-level hierarchical linear models showed that this effect is no longer significant when considering the intermediate level of the classroom in the analysis, indicating that previous research may have overestimated the effect of school size on the sense of school belonging. These evidence raises the question of which levels of the school environment are more relevant to better understanding individual differences in the sense of school belonging.

Patients’ symptoms and strengths as predictors of long-term outcomes of CBT for generalized anxiety disorder – A three-level, multi-predictor analysis

Although cognitive behavioral therapy (CBT) is an effective treatment for generalized anxiety disorder (GAD), GAD often shows a chronic clinical course and common deterioration after treatment. Many trials have examined the efficacy of treatments in GAD, but little is known about intake predictors of long-term treatment outcomes. This study examined potential predictors of long-term treatment outcomes based on the individual’s symptom severity and strengths (behavioral, cognitive, interpersonal) at intake. Long-term outcomes were defined as worry at six-month follow-up (six-m FU) and worry decrease from intake and post-treatment to six-m FU. Data from 137 CBT outpatients with a GAD diagnosis from two randomized clinical trials were analyzed using three-level hierarchical linear modeling. Results revealed that worrying decreased up to the six-m FU. In single-predictor models, intake symptom severity and strength measures predicted worry at the six-m FU. In multi-predictor models, only behavioral strengths remained a significant predictor. Worry decrease from intake to the six-m FU was only predicted by behavioral strengths. These findings provide relevant information about intake predictors of long-term outcomes after CBT for GAD and underscore the potential relevance of assessing patients’ strengths for clinical practice.

Analyzing the Relationship between the Major League Baseball (MLB) Attendance and the Length of Baseball Games: A Hierarchical Linear Modeling of the 2010-19 MLB Seasons

In order to investigate the relationship between Major League Baseball (MLB) attendance and game duration, this study analyzed the number of spectators throughout the 2010-19 MLB regular seasons (N=24,297) with attendance determinants identified in earlier studies. Also, a three-level hierarchical linear modeling (HLM) was used in consideration of the characteristics of attendance data for several seasons, teams, and individual games. A total of 27 independent variables were used in the current research, and 17 attendance determinants, including game duration (γ =4.45, p<.05), were identified as having a significant impact on the number of spectators. Furthermore, by confirming the hierarchical structure of the attendance data, this study highlights how critical it is to utilize appropriate analytical methods. Based on these findings and statistical analysis that confirmed MLB''s current attendance decline, these results will help in the development of a number of strategies to increase or sustain team’s attendance to practitioners.

Race moderates the association between adult attachment avoidance and the cortisol awakening response among young sexual minority men

Adult attachment, or the ways in which individuals develop and maintain trusting and loving relationships with peers and romantic partners, has been demonstrated to influence hypothalamic-pituitary-adrenal (HPA) axis functioning among young sexual minority men (YSMM). Theory and preliminary evidence suggest that differences in the lived experiences of White and Black sexual minority young adults may influence the way in which adult attachment influences the HPA-axis functioning. We sought to further this field of inquiry by examining if race moderates the association between adult attachment insecurity (i.e., avoidance and anxiety) and HPA-axis functioning among YSMM. Sixty-three YSMM participated in a 5-day daily diary study in which they completed a baseline survey to assess adult attachment orientation and provided saliva samples 4 times per day over a 5-day period in order to measure cortisol across the day. Three-level hierarchical linear modeling was used to examine the association between adult attachment, the cortisol awakening response (CAR), diurnal cortisol, and the moderating effect of race. We found no main effect of attachment anxiety or avoidance on HPA-axis functioning. However, we found that race moderated the association between adult attachment avoidance and the CAR such that Black YSMM with higher attachment avoidance had a lower CAR as compared to White YSMM. Results suggest that there may be key differences between White and Black YSMM men’s lived experiences that should be explored in future research.

The influence of ICT use and related attitudes on students’ math and science performance: multilevel analyses of the last decade’s PISA surveys

This study analyzed the latest four PISA surveys, 2009, 2012, 2015, and 2018, to explore the association between students’ ICT-related use and math and science performance. Using ICT Engagement Theory as a theoretical framework and a three-level hierarchical linear modeling approach, while controlling for confounding effects, ICT-related independent variables of interest were added to the models at the student, school, and country levels. The series of models revealed that, in general, an increase in ICT availability and ICT use both inside and outside school had a negative association with learning outcomes, while students’ positive attitude toward ICT demonstrated a strong positive relationship. However, students’ perceived autonomy related to ICT use had the strongest association with academic performance, which is consistent with the changing nature of the modern learning environments. Findings revealed that virtually all forms of student ICT use, both inside and outside of school and whether subject related or not, had no substantive positive relationship with student performance in math or science. Conversely, higher student attitude toward, confidence in, belief in utility of, and autonomous use of ICT was associated with higher math and science performance for each of the four years of the study. Incidentally, we also found that while country GDP per capita had no consistent association with student performance, a school’s provision of extra-curricula activities did. Recommendations for educational leaders, teachers, and parents are offered.

Influence of students’ perceptions of instruction quality on their digital reading performance in 29 OECD countries: A multilevel analysis

Students' perceptions of instruction quality have long been considered a core element of the effectiveness of education. In particular, perceived high-quality instruction in classrooms might motivate students to gain better digital reading performance, which helps them become well-prepared in the era of information and communication technology (ICT). Thus, this study investigated the correlation between students' perceptions of instruction quality features and their digital reading performance from the aspects of perceived instructional approaches, classroom management, and supportive climate. Data on 223,807 15-year-old students in 29 Organization for Economic Co-operation and Development (OECD) countries were extracted from the latest Programme for International Student Assessment (PISA) 2018 database and the study adopted three-level hierarchical linear modeling (HLM). The findings demonstrated the positive influence of students' perceived adaptation of instruction, stimulation of reading engagement, disciplinary climate, teacher interest, and teacher support. However, students’ perceived teacher-directed instruction, reading skills exercises, teaching of digital skills, language instruction time, and teacher feedback were unexpectedly negatively correlated with their digital reading performance. Finally, the study concludes by discussing its implications for improving reading instruction quality.

Therapist and counseling center effects on international students' counseling outcome: A mixed methods study.

We employed a convergent mixed methods design to examine therapist and counseling center effects on international student clients' (ISCs) counseling outcomes. Using the Center for Collegiate Mental Health (CCMH) data set (2015-2017), we conducted a three-level hierarchical linear model with clients (N = 85,110) nested in therapists (N = 1,267), and therapists nested in counseling centers (N = 111), with clients' international status predicting distress (DI) in their last sessions while controlling for initial DI. Compared to domestic students, the average last session DI was significantly higher among ISCs. Random effects were significant, suggesting that some therapists and centers were more effective in their work with ISCs than others. When the proportion of ISCs seen was accounted into the model, we found a cross-level interaction in which the last session DI differences between ISCs and domestic students were significant for centers seeing a small percentage of ISCs but not for centers with large percentages of ISCs in the caseload. Grounded theory analysis of qualitative data from 11 therapists with international backgrounds revealed therapist and center factors that converged with our quantitative findings. Participants reported adhering to general clinical frameworks when working with ISCs given the lack of training on international competence (which may help explain the effectiveness gap), but also noted nuanced culturally-informed components that likely contribute to more effectively working with ISCs. Findings around center effects were complemented by qualitative results emphasizing systemic representation and engagement with diversity, creative outreach efforts, and administrative/leadership support. Implications for practice and research are discussed. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

The Relationships Among ICT-Related Psychological Factors, School Contextual Factors and Secondary Students’ Reading Performance: A Multilevel Analysis Across 47 Economies

Evolving information and communication technology (ICT) reshapes people’s reading activities by popularizing digital reading. Past studies focused on the relationships between students’ frequency of ICT use and reading performance but neglected the possible interrelationships between students’ ICT-related psychological factors, school contextual factors, and reading performance. This study applied the Self-determination Theory (SDT) to explore the relationships between students’ ICT-related psychological factors (perceived interest, autonomy, competence, and social interaction in using ICT) and reading performance. The possible moderation effects of schools’ contextual variables (schools’ support in ICT devices and schools’ support in teachers’ capacity to integrate technology in teaching) were also investigated. Data from 222,293 secondary students from 10,103 schools in 47 economies who participated in the Program for International Student Assessment (PISA) 2018 were analyzed using three-level hierarchical linear modeling. The results provided support for the SDT by confirming the positive correlations between students’ ICT-related interest, autonomy and competence and reading performance, and the accentuating moderation effect of schools’ contextual factors. However, students’ perceived social interaction in ICT use was negatively correlated with reading performance, and this negative correlation worsened with increasing schools’ support in ICT devices while buffered with increasing schools’ support in teachers’ capacity to use technology.

Patterns of cognitive returns to Information and Communication Technology (ICT) use of 15-year-olds: Global evidence from a Hierarchical Linear Modeling approach using PISA 2018

Existing literature shows varying impacts of Information and Communication Technology (ICT) on learning outcomes depending on the type and quality of technology used. However, current research on the optimal level of ICT use for the cognitive development of students is scarce and has remained country-specific, primarily focusing on developed economies. This paper undertakes a cross-country comparison across 79 nations, investigating ICT use and its association with cognitive gain patterns as determined by reading, mathematics, and science test scores of 15-year-olds. The Programme for International Student Assessment (PISA) 2018 dataset collected by the Organisation of Economic Cooperation and Development (OECD) has been used by applying two-stage regression analysis. The first stage involves a three-level Hierarchical Linear Modeling (HLM) factoring for data nested at the student-, school-, and country-level. The second stage of the empirical model involves a heterogeneity analysis to evaluate variance in ICT use patterns across different groups of countries, clustered on the basis of their level of ICT development. The results show a positive impact of ICT engagement on the test scores of students across all the subjects. However, returns to ICT use tend to start diminishing after the engagement level of students crosses a medium threshold of using ICT several times within a week. Furthermore, the heterogeneity analysis supports conspicuous diminishing patterns in ICT use irrespective of the economic status of the students. Cross-country comparisons show that diminishing returns to ICT use are more prominent in countries with well-developed ICT infrastructure than in less-developed ones. Where diminishing returns hold, excessive use of ICT in education is not an optimal choice, and significant cognitive gains can be achieved by using the complementarity between traditional learning techniques with ICT-based learning in different blended settings.

For whom does a match matter most? Patient-level moderators of evidence-based patient-therapist matching.

A double-blind, randomized controlled trial tested the effectiveness of a personalized Match System in which patients are assigned to therapists with a "track record" of effectively treating a given patient's primary concern(s) (e.g., anxiety). Matched patients demonstrated significantly better outcomes than those assigned through usual pragmatic means. The present study examined patient-level moderators of this match effect. We hypothesized that the match benefits would be especially pronounced for patients who presented with (a) greater overall problem severity, and (b) greater problem complexity (i.e., number of elevated problem domains). We also explored if patient racial/ethnic minority status moderated the condition effect. Patients were 218 adults randomized to the Match or as-usual assignment condition, and then treated naturalistically by 48 therapists. The primary outcome was the Treatment Outcome Package (TOP), a multidimensional assessment tool that also primed the Match algorithm (based on historical, therapist-level effectiveness data), and assessed trial patients' symptoms/functioning and demographic information at baseline. Moderator effects were tested as patient-level interactions in three-level hierarchical linear models. The beneficial match effect was significantly more pronounced for patients with higher initial severity (-0.03, 95% CI -0.05, -0.01) and problem complexity (-0.01, 95% CI -0.02, -0.004), yet the high correlation between severity and complexity called into question the uniqueness of the complexity moderator effect. Moreover, the match effect was more pronounced for racial/ethnic minority patients (i.e., nonwhite; -0.05, 95% CI -0.09, -0.01). Measurement-based matching is especially effective for patients with certain characteristics, which further informs mental health treatment personalization. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Working alliance, therapist expressive skills, and client outcome in psychodynamic therapy.

Bordin's (1979) theory suggests that therapist techniques that call for client introspection and self-observation will be more effective when the working alliance (WA) is stronger. Psychodynamic therapists use expressive techniques to elicit this introspection and self-observation. We examined whether therapists' use of expressive skills (e.g., encouraging expression of thoughts and feelings; helping clients understand the reasons behind their thoughts, feelings, and behaviors) when the WA is high, versus low, was related to client outcome in open-ended, psychodynamic treatment. Ten therapists rated the WA with their 47 clients, who rated their perceptions of helping skills, after 2,284 counseling sessions. Clients also completed the Outcome Rating Scale (ORS) in reference to the week following each session. We examined time-ordered relationships by creating lagged variables for WA (T-2) and therapist expressive skills (TES; T-1) and used these scores to predict ORS ratings (T) in a three-level Hierarchical Linear Modeling (HLM) analysis (sessions nested within clients, nested within therapists). Lagged WA (T-2) and expressive skills (T-1) were decomposed into between-therapist, within-therapist, and within-clients components. We found a 3-way interaction for within-therapist working alliance, within-client WA at T-2, and expressive skills at T-1 predicting ORS at T. When clients who had a weaker WA with their therapist than the therapists' other clients, had a session with a stronger WA than the client's usual WA, then more expressive skills in the next session was related to better psychological well-being in the subsequent session. Implications and future directions are discussed. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Differential return on investment: Academic growth in mathematics and reading based on initial performance.

Students vary in their initial achievement when they enter school and their rate of academic growth as they move through school. These differences have implications for classroom instruction and educational policy. Although previous research has examined initial achievement and growth differences, a gap remains in understanding how initial level of achievement interacts with subsequent growth as children move through school. Using Vygotsky's zone of proximal development (ZPD) and return on investment as theoretical grounding, this registered report examined how students' initial academic performance relative to their school predicts their subsequent academic achievement. The stage 1 accepted registered report is available at https://osf.io/9zmak/. Specifically, we tracked the achievement of a cohort of students who started at or above their school's mean at the beginning of third grade and tested a range of hypotheses regarding their achievement and growth as well as which students showed the greatest gains from their time in school. Using a large database of student academic achievement in the United States, this registered report included de-identified data from all students from fall 2014 to spring 2017 in grades three through five from the ten US states with the highest participation for the Northwest Evaluation Association's Measures of Academic Progress (MAP®) - a computer adaptive test of academic achievement in mathematics and reading. Because the MAP is taken at least twice per school year, up to six scores were included on mathematics and reading achievement for effective samples of approximately 220,000 students. We built separate reading and mathematics three-level piecewise longitudinal hierarchical linear models (student repeated measures, nested within students, nested within schools) to model student growth from the beginning of third grade to the end of fifth grade (i.e., three academic years and two summers). For both mathematics and reading, average student achievement growth slowed as they progressed from third through fifth grade. From there, the findings diverged. In mathematics, student growth was mostly similar across achievement levels and grades from third through fifth. However, in reading, above-average students demonstrated slower growth than average students during the school year but faster growth during the summer. Also of note, at the beginning of third grade, the highest achieving students outscored average students in their school by more than 2 years in mathematics and 3 years in reading. Our results may be able to be explained via a ZPD model, which posits development only occurs when students are placed in appropriately challenging environments. In mathematics, the observed pattern of relatively consistent growth across achievement levels suggests average students were just as likely to be in their ZPD as higher achieving students. In reading, as initial achievement increased, student reading growth slowed, which suggests the higher the initial achievement, the less likely students were to be in their ZPD. If a goal of education is for students to learn new things, our results suggest existing school offerings in reading are not meeting that goal equitably for students across the performance spectrum. Differential growth patterns should be considered when designing learning experiences for students who enter with a wide range of prior mastery.