Multilevel Latent Profile Analysis Research Articles

Exploring the moderator role of organisational ambidexterity in the relationship between instructional leadership and teacher instructional practices

This study aimed to explore schools’ ambidexterity profiles and examine how each profile influences the relationship between instructional leadership and teacher instructional practices. Using data from 1150 teachers across 108 schools in Turkey, we employed multilevel latent profile analysis to group schools into clusters based on aggregated teachers’ responses to organisational ambidexterity. We conducted moderation analyses with latent profile variables. Results revealed two distinct profiles: high explorative-exploitative and low explorative-exploitative schools. The study also provided evidence of the significant moderating role that school ambidexterity profiles played in the link between leadership and teaching. We discussed implications for policy and practice.

Exploring Influencing Factors at Student and Teacher/School levels on Science Self-efficacy Using Machine Learning and Multilevel Latent Profile Analysis

The main purpose of this study was to explore major factors at student and teacher/school levels affecting middle-school students’ science self-efficacy among countries with high science academic performance levels in PISA 2015 data by using various regression algorithms of machine learning and multilevel latent profile analysis (MLPA). Out of total of 195 (110 at the student-level and 75 at the teacher/school-level), 88 at the student-level and 38 at the teacher/school-level explanatory variables were finally selected after data pre-processing. Through over 20 regression analysis algorithms of machine learning, 10 variables at the student-level and five variables at the teacher/school-level were found to be significant predictors, consistent with previous studies. Next, MLPA was applied to classify underlying science self-efficacy sub-groups at each level and verify the statistical significance of variables affecting science self-efficacy, chosen from the machine learning. Three classes at the student-level (low, moderate, high) and two at the teacher/school-level (mid-low, mid-high) were selected as the optimal number of latent profiles. At the student-level, students with higher environmental awareness, science activities, interest in broad science topics, instrumental motivation, test anxiety, and achieving motivation were more likely to belong to the moderate or high groups than the low group. At the teacher/school-level, the mid-high group had more considerable science-specific resources, higher instructional leadership, a larger student-teacher ratio, and more science teachers than mid-low group. Finally, the significance of the study was presented, and implications for increasing science self-efficacy in middle school students were suggested.

Examining teacher-child relationship profiles in forecasting children's self-regulation and social-emotional skills: A multilevel latent profile approach

Numerous studies have highlighted the crucial role of teacher-child relationships in children's self-regulation and social-emotional skills. However, whether teacher-child relationship profiles are linked to self-regulation and social-emotional skills remains unclear. The present study employed multilevel latent profile analysis (MLPA) to explore the teacher-child relationship profiles and their associations with self-regulation and social-emotional skills at both the child and class levels. Utilizing data from 584 children nested in 42 preschool classes, single-level LPA unveiled three teacher-child relationship profiles associated with different self-regulation and social-emotional skills. Notably, closeness and conflict emerged as significant factors in profile assignments. Children with low closeness and high conflict tended to experience challenges in self-regulation and social-emotional skills, whereas those with moderate to high closeness reported better outcomes. Furthermore, MLPA identified three class profiles based on the prevalence of children's profiles and their relationships with different class-mean self-regulation and social-emotional skills. Children's age, gender, and teachers' teaching experience were identified as factors related to child- and class-level profiles. These findings emphasize the interplay between teacher-child relationship profiles and classroom dynamics in relation to children's self-regulation and social-emotional skills within a Chinese cultural context.

Profiles of alcohol intoxication and their associated risks in young adults' natural settings: A multilevel latent profile analysis applied to daily transdermal alcohol concentration data.

Transdermal alcohol concentration (TAC) sensors capture aspects of drinking events that self-reports cannot. The multidimensional nature of TAC data allows novel classification of drinking days and identification of associated behavioral and contextual risks. We used multilevel latent profile analysis (MLPA) to create day-level profiles of TAC features and test their associations with (a) daily behaviors and contexts and (b) risk for alcohol use disorders at baseline. Two hundred twenty-two regularly heavy-drinking young adults (Mage = 22.3) completed the Alcohol Use Disorders Identification Test (AUDIT) at baseline and then responded to mobile phone surveys and wore TAC sensors for six consecutive days. MLPA identified day-level profiles using four TAC features (peak, rise rate, fall rate, and duration). TAC profiles were tested as correlates of daily drinking behaviors, contexts, and baseline AUDIT. Four profiles emerged: (a) high-fast (8.5% of days), (b) moderate-fast (12.8%), (c) low-slow (20.4%), and (d) little-to-no drinking days (58.2%). Profiles differed in the odds of risky drinking behaviors and contexts. The highest risk occurred on high-fast days, followed by moderate-fast, low-slow, and little-to-no drinking days. Higher baseline AUDIT predicted higher odds of high-fast and moderate-fast days. Days with high and fast intoxication are reflective of high-risk drinking behaviors and were most frequent among those at risk for alcohol use disorders. TAC research using MLPA may offer novel and important insights to intervention efforts. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Supplemental Material for Profiles of Alcohol Intoxication and Their Associated Risks in Young Adults’ Natural Settings: A Multilevel Latent Profile Analysis Applied to Daily Transdermal Alcohol Concentration Data

Supplemental Material for Profiles of Alcohol Intoxication and Their Associated Risks in Young Adults’ Natural Settings: A Multilevel Latent Profile Analysis Applied to Daily Transdermal Alcohol Concentration Data

Daily profiles of parents' supportive extrinsic emotion regulation of adolescents' negative emotion.

Parental support for adolescent emotion regulation is critical for adolescents' health. Yet, little is known about parents' daily support of adolescents' emotion regulation. This study aimed to typify daily co-parent supportive extrinsic emotion regulation (EER) profiles directed toward adolescents' daily distress and anger. The sample comprised 153 adolescent-parent triads; adolescents' mean age, 15.71 years (SD = 1.53), 51% girls. Over 7 consecutive days, adolescents self-reported their distress and anger, while parents reported their own negative emotions and their perception of the adolescent's negative emotions. Parents also reported daily on their utilization of seven supportive EER strategies, including problem- and emotion-focused strategies. Multilevel latent profile analysis (MLPA) identified four day-level profiles of parental EER; "low" (40% of days), reflecting low EER efforts of both parents across all EER strategies: "high" (12%), reflecting high EER involvement of both parents across all strategies; "mother-high father-low" (26%), reflecting mothers' high and fathers' low use of all strategies; "father-high mother-average" (22%), reflecting fathers' high use of all strategies, and mothers' low to average use of all strategies. The likelihood of specific EER profiles across days did not associate with daily changes in adolescents' anger. However, on days when adolescents felt more distress, the likelihood of a "high" parental EER profile was significantly greater than "low." Findings suggest a dynamic repertoire of co-parent EER profiles, responsive to adolescent heartfelt emotions but not hostility. The lack of parental EER of adolescents' anger might put adolescents at increased risk for anger escalation and the unhealthy discharge of anger. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Who believes in a just world? A multilevel latent profile analysis of Justice Capital using the European Social Survey.

Justice Capital provides a theoretical framework for explaining individual differences in the belief in a just world (BJW). However, this framework has yet to receive empirical validation. Using Round 9 of the European Social Survey, a large (n = 43,209) multi-country (N = 29) sample, I conduct multilevel latent profile analysis and multilevel multinomial logistic regression to determine the latent profiles that emerge at a population level and map the demographic and experiential covariates of these profiles. Incorporating measures of general BJW, distributive and procedural justice, and the belief in equality of opportunity, I find three latent profiles: meritocrats, moderates, and egalitarians. Compared with egalitarians, meritocrats (strong just world believers) are more likely to be male; younger; have a higher income; have attained more years of education; to be politically conservative; and have no recent experience of discrimination or crime. Meritocrats were overrepresented in countries with a higher Human Development Index. This study demonstrates the feasibility of Justice Capital for understanding individual variation in general BJW and related justice beliefs; discussion centers on anomalous findings and extension of this theoretical framework.

Determining the factors affecting the ICT use of Turkish students attending PISA-2018: Multilevel profile analysis approach

This study aims to classify student profiles based on the type and frequency of Information and Communication Technologies (ICT) usage. Each profile exhibits homogeneous characteristics and heterogeneous characteristics compared to other groups. Additionally, the study investigates whether covariates at the school and student levels create differences in these profiles. For this purpose, a mixed-method analysis (Latent Profile Analysis (LPA), Multi-level Latent Profile Analysis (MLPA) and Multi-level Logistic Regression Analysis (MLRA)) were used to identify ICT usage profiles. The study analysed data from 6,890 Turkish students across 186 schools participating in PISA 2018 (Program for International Student Assessment), revealing three distinct ICT usage profiles: Profile-1 (Low-Level ICT Users (19.06%)), Profile-2 (Intermediate-Level ICT Users (71.10%)), and Profile-3 (Advanced-Level ICT Users (9.84%)). The MLRA showed that students with low ICT usage have less access to ICT resources. Students in the Low-Level ICT Users profile also had lower scores on ICT-related motivation factors (interest, competence, social interaction, and autonomy). No significant effect of ICT usage on mathematics, reading, and science scores was observed. Variance differences at student and school levels could not be fully explained by covariates at the school level (type of school, ratio of computers connected to the internet per student, lack of educational materials and personnel). However, the ratio of computers per student effectively explained these differences. According to an MLPA, schools were categorised into two groups based on ICT resources: 'high' and 'low'. In schools with high ICT resources, the proportion of students in Profile-2 was found to be 45.8%, while those in Profile-3 accounted for 9%. In contrast, in schools with low ICT resources, the percentage of students in Profile-2 was 22.45%, and in Profile-3, it was 1.82%. These findings indicate that the level of a school's ICT resources significantly impacts students' technology usage skills. A significant finding of the research is that most Turkish students use ICT at an intermediate level at home for school lessons. The COVID-19 pandemic has further increased the importance of studying at home using ICT resources.

Typologies of secondary school student academic resilience in science with classroom and school context predictors

ABSTRACT The education system strives to help students from low-income families achieve academic success. Academic resilience is related to not only individuals but also classrooms and schools. This study aimed to construct a comprehensive resilience model in science domains that presents the image of resilient students and describes the mechanisms by which they arise. This goal was achieved through a parametric multilevel latent profile analysis with level 1 and level 2 covariates. Latent profile indicators included variables reflecting motivation, subjective well-being, and engagement in science of secondary school students. Variables related to science classroom and school contexts constituted the levels 1 and 2 covariates. We used secondary data from Japan and Macao (China) collected in the 2015 Programme of International Student Assessment (PISA). The sample sizes were 1594 and 1114, respectively. We identified three resilience levels among students. The emergence of resilient students benefited from science classroom discipline, teacher-directed instruction, inquiry-based teaching, adjusting instruction, and teacher support. In Japan, the most robust predictor was inquiry-based teaching, while in Macao, it was teacher-directed instruction. Therefore, science teachers play a crucial role in building students’ resilience. It also suggests that a small class decreases the likelihood of students entering the high-level resilience group.

Students' regulation of anxiety and hope-A multilevel latent profile analysis.

Academic examinations are highly emotional for university students, making ER essential for preventing or reducing the negative consequences of negative emotions on well-being and academic performance. Initial theorizing and research suggest that flexibly applying combinations of strategies can promote successful ER. However, studies using intraindividual approaches to examine ER strategy use in specific contexts across multiple occasions, are lacking. Moreover, the combinations of strategies used by students within different contexts, and the adaptiveness of different strategies for regulating different emotions, remain unexplored. To address these gaps, we conducted an experience sampling study to identify patterns of students' momentary ER and to examine how context (achievement-related vs. nonachievement-related), emotions (anxiety vs. hope), and academic performance function as potential covariates. Over 200 university students rated their anxiety, hope, and use of eight ER strategies over a 7-day period, six times a day, prior to an important exam in 2016 and 2017. Results of a two-level latent profile analysis revealed distinct profiles of ER that differed on both levels. Intraindividually, ER patterns differed as a function of type of emotion and context experienced. More specifically, momentary use of multiple strategies tended to be associated with greater anxiety, while in the achievement context this association was reduced. Interindividually, students' tendencies to use different ER patterns were not related to test performance. Our findings suggest that ER strategy selection depends on both context and emotions, and advance ER research by considering intraindividual strategy use in concrete achievement situations. Yet the sample is not sufficiently representative. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Adolescents’ self-regulated and affective learning, teacher support and digital reading literacy: A multilevel latent profile approach

A considerable body of research has demonstrated the pivotal roles played in digital reading by self-regulated learning (SRL) and affective learning. Yet, whether teacher support facilitates SRL and affective learning has hitherto remained unclear. This cross-cultural study adopted multilevel latent profile analysis (LPA) to establish latent profiles of adolescents’ SRL and positive affect, explored how such profiles were differentially correlated to digital-reading performance, and examined how perceived teacher support predicted profile memberships, at both student and school levels. Using data collected from 34,000 students nested in 1,100 schools from three Asian and three Western educational systems by the Programme for International Student Assessment 2018, single-level LPA revealed three or four student profiles, in associations with disparate digital reading. While self-regulatory strategies were decisive in profile assignment, emotions did not vary significantly across profiles or predict performance. The least self-regulated learners often fell into a vicious circle of poor emotional well-being and low performance, but highly self-regulated and capable digital readers did not necessarily report the most positive emotions. Multilevel LPA identified two or three school profiles based on relative frequency of student profiles, in relations to different school-mean digital reading. Students’ perceived teacher support significantly predicted student-profile memberships; greater support increased likelihoods of being stronger self-regulated learners. But such support did not predict school-profile memberships. East-West cultural variations existed in all the influential patterns.

How do working students fare? A person‐centric approach to understanding patterns of work–school conflict and facilitation

AbstractStudent workers are largely understudied in organizational research, yet they represent an important part of the workforce. Their numbers are expected to rise as tuition continues to increase, and many adult workers are returning to school. The current study utilizes a person‐centric approach to investigate latent patterns of work–school conflict (WSC) and work–school facilitation (WSF) as well as the antecedent (i.e., supervisor work–school support) and outcomes (i.e., job satisfaction, well‐being, and school outcomes) tied to these latent patterns across school weeks. In a pilot study (n = 178 student workers), we first confirmed whether classes of WSC and WSF could be identified using latent profile analysis. Then, using 347 weekly WSC and WSF experiences, multilevel latent profile analysis identified four distinct profiles of WSC and WSF at the week level. Also, this modeling demonstrated that individuals tend to fluctuate between at least two profiles across weeks. Supervisor work–school support was found to predict weekly profile membership. Furthermore, the week‐level profiles exhibited distinct relations with weekly outcomes (i.e., job satisfaction, well‐being, and school preparedness). The results provide important theoretical and practical insights into the joint experience of interrole conflict and facilitation among student workers.

Multilevel Latent Profile Analysis of Global Competence in PISA 2018: Small Country-Level Differences

We examined the global competence of 15-year-old students based on PISA 2018 data ( N = 355,579 from 57 countries). Our multilevel latent profile analysis identified two groups at both individual and country levels, resulting in a total of four groups (2 x 2; “competent” and “less competent”). While individual-level differences were pronounced and consistent across all seven global competence indicators, the country-level differences were small overall, with only “attitude towards immigrants” and “respect for people from other cultures” showing somewhat salient group differences. There were more globally “less competent” students in 52 of 57 countries, suggesting the needs to improve students’ global competence in virtually all participating countries. Global competence was related to family socioeconomic status, academic achievement, and countries’ standing on the conservatism/liberalism dimension. We considered different ways to enhance global competence among students and society as a whole.

Profiles of teaching practices and teacher co-operation: A person-centered approach in multilevel context

In this study, teachers’ profiles in terms of their instructional and professional practices were defined by using multilevel latent profile analysis. For this study, Turkish national data from TALIS 2018 (N = 6861 teachers from 443 schools) were utilized. Five teacher profiles that differed especially in terms of “clarity of instruction” emerged. Members of Profile E are the teachers using instructional practices most, and the members of Profile A are the ones using these practices least frequently. Results indicated the important role of both the “school climate for innovativeness” and school location as contextual factors in predicting profile memberships.

Working and working out: Decision-making inputs connect daily work demands to physical exercise.

Work demands can undermine engagement in physical exercise, posing a threat to employee health and well-being. Integrating resource theories and a novel decision-making theory called the decision triangle, we propose that this effect may emerge because work stress changes the energetic and emotional processes people engage in when making decisions about exercise after work. Using diary-style data across two workweeks (N = 83 workers, 783 days), we used multilevel latent profile analysis to extract common decision input profiles, or daily configurations of energy and affect as key decision-making resources. Consistent with the decision triangle, three profiles emerged: visceral inputs (low energy/high negative affect), automatic inputs (low energy/low negative affect), and logical inputs (high energy/low negative affect). Daily job demands were highest among the visceral profile. In turn, the daily visceral profile related to the lowest likelihood of and intensity of physical exercise after work, especially relative to the daily logical profile. Whether or not those in the daily automatic profile exercised depended on their health orientation, or trait-level value of maintaining personal health. Our results support decision-making as a promising mechanism explaining the link between work demands and healthy leisure choices. Organizational interventions can target work stress, health orientation, or logical decision-making to promote frequent and vigorous employee physical exercise. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

'Prosociality' in Daily School Life and Early Adolescents' Peer Aggression: A Multilevel Latent Profile Analysis Approach.

We aimed to typify prosocial characteristics of aggressive youth. We classified early adolescents based on daily configurations of prosocial behavior and autonomous prosocial motivations (performing prosocial behavior for identified and intrinsic reasons) and controlled prosocial motivations (performing prosocial behavior for external and introjected reasons) and explored the links between the obtained sub-groups and peer aggression. The sample included 242 Israeli six-graders [Mage = 11.96 (SD = 0.18), 50% girls] and their teachers. At the daily level, adolescents self-reported on prosocial behaviors and their autonomous and controlled prosocial motivations for ten consecutive days. At the trait level, adolescents reported on global, reactive, and proactive peer aggression. Teachers reported on adolescents' global peer aggression. Using multilevel latent profile analysis, we identified four day-level profiles of prosociality: 'high prosocial autonomous' (39% of days), 'low prosocial' (35%), 'average prosocial controlled' (14%), and 'high prosocial bi-motivation' (13%). At the adolescent level, we identified four sub-groups, each characterized by one dominant daily profile: 'stable high autonomy' (33% of adolescents); 'stable high bi-motivation' (12%); 'often average controlled' (16%); 'often low' (39%). Higher self-reported aggressive adolescents, particularly proactive aggressive, had the least chance of being in the 'stable high autonomy' sub-group of all sub-groups. Teacher-reported aggressive adolescents had the least likelihood of being in the 'stable high autonomy' sub-group and the most likelihood of being in the 'often low' sub-group. In sum, peer aggression is a function of the configured phenomenology of prosocial behavior and motivations, with high prosocial autonomously motivated youth being the least aggressive.

Daily receiving and providing of social support at work: identifying support exchange patterns in hierarchical data

ABSTRACT The norm of reciprocity proposes that one who receives support feels obliged to return obtained benefits. Yet how employees regulate the mutual flow of social support with one another does not always follow a reciprocal dynamic, which may have different implications for employee outcomes based on whether social exchanges gain, drain or conserve resources. To better understand support exchange patterns and their relationship with basic need satisfaction (for autonomy, competence and relatedness) and emotional exhaustion in employees’ daily lives, we performed a multilevel latent profile analysis (N = 124 employees, 894 days). At the day level, we identified four support exchange profiles differing in low vs high received and provided social support (i.e., low-reciprocating, under-reciprocating, over-reciprocating and high-reciprocating days). At the person level, we identified three support exchange classes of employees, differing in the degree to which employees showed certain support exchange profiles over time (i.e., over-reciprocators, low-reciprocators and high-reciprocators). Pairwise comparisons with multinomial logistic regression revealed that over-reciprocating days were associated with the highest levels of need satisfaction for autonomy and relatedness and the lowest levels of emotional exhaustion. Moreover, over-reciprocators were most likely to satisfy their needs for autonomy. We discuss theoretical and practical contributions of our findings.

Exploring Teachers’ Instructional Practice Profiles: Do Distributed Leadership and Teacher Collaboration Make a Difference?

Purpose: While the literature includes multiple studies on the relationship between school leadership and instructional quality, they often use instructional practice as a continuous variable, assuming that a teacher would perform all sub-dimensions of instructional practice at a similar rate and failing to link distributed leadership to classroom teaching. Addressing these gaps in the literature, this study aims to identify teacher- and school-level latent profiles of teachers’ instructional practices and to investigate how distributed leadership predicts teachers’ membership in different instructional practice profiles, with the mediating role of teacher collaboration. Research Methods/Approach: The study employed a cross-sectional survey design using Türkiye's TALIS data for lower secondary education. Multilevel latent profile analysis with mediation modeling was conducted on data from 3,223 teachers in 192 schools. Findings: This analysis yielded four teacher profiles: laissez-faire, typical, controlling, and versatile; and two school profiles, high controlling and high laissez-faire. Findings indicate that distributed leadership promotes professional collaboration in lessons among teachers, which could, in turn, play a critical role in determining both individual teacher- and school-level profiles. Implications: This study provides practical contributions to understanding the nature of classroom teaching, suggesting that future studies should use instructional practice profiles instead of a single construct of teaching.

Profiles of Chinese preschoolers' academic and social-emotional development in relation to classroom quality: A multilevel latent profile approach.

Linking classroom quality to separate domains of child development might neglect the transactional interactions across developmental domains. This research utilized latent profiles across academic and social-emotional development to explore which aspects of classroom quality can predict children's profiles at the classroom level. Data were drawn from 96 preschool classrooms and 547 children (3-5 years old) in China in 2020. Multilevel latent profile analysis identified three profiles (entitled low-, average- and high-level development at the individual level), and two classes (entitled average and below-, average and above) at the classroom level. Multinominal logistic regression analyses revealed that instructional quality in math, science, and diversity, and the interactional quality in supporting children's learning and critical thinking, predicted children's profiles.

Does classroom matter? - A longitudinal multilevel perspective on students' achievement goal orientation profiles during lower secondary school

The present study adds to earlier person-oriented research by investigating differences in students' achievement goal orientation (AGO) profiles and their development using a simultaneous consideration of classroom patterns with longitudinal multilevel methods. The sample of almost 10,000 lower secondary school students, representing over 600 classrooms, was surveyed on their AGOs in the 7th and 9th grade. Multilevel latent profile analyses (MLPAs) and transition analysis (MLTA) revealed similar student profiles in AGOs in both grades: success-oriented, moderate multiple goals and avoidance-oriented, as well as two classroom types: success-oriented and mixed orientation classrooms with varied relative proportions of different student-level profiles and patterns of likely transitions. Stability of profiles was more typical than change. Maladaptive transitions were related to lower, and stable and adaptive transitions to higher GPA in the end of 9th grade. In success-oriented classrooms, it was more common to maintain or adopt the success-orientation across lower secondary school compared to the other classroom type.