Multilevel Latent Class Analysis Research Articles

Profiling Teachers’ Motivation for Professional Development: A Nationwide Study

Situated in the context of advanced placement (AP) reform in the United States, we investigated profiles of teachers’ motivation for participating in professional development (PD) courses in a two-cohort sample of nt1 = 2,369 and nt2 = 2,170 chemistry teachers via multilevel latent class analysis. In addition, the study investigated to what extent profile membership was related to factors at the teacher, school, and PD levels. Participation in PD courses was associated with one of three profiles, labeled “reform-motivated,” “convenience-motivated,” and “interaction-motivated.” Participation in PD courses was more likely to be reform-motivated if a teacher had a major in chemistry, more experience teaching AP, more positive attitudes toward PD, or higher enactment of AP redesign in the classroom, or if the PD course was formal and face-to-face. The results show that teachers have different motivations for participating in PD courses and provide insight into how to engage teachers in professional learning.

Multilevel Factor Mixture Modeling: A Tutorial for Multilevel Constructs

Multilevel factor mixture modeling (FMM) is a hybrid of multilevel confirmatory factor analysis (CFA) and multilevel latent class analysis (LCA). It allows researchers to examine population heterogeneity at the within level, between level, or both levels. This tutorial focuses on explicating the model specification of multilevel FMM that considers the conceptualization of multilevel constructs. Empirical data sets are used to demonstrate the applications of multilevel FMM for within-level constructs, between-level constructs, and within- and between-level constructs. Detailed model specifications of integrating latent classes into multilevel constructs are provided. For modeling the heterogeneity at the between level, parametric and nonparametric approaches are compared both conceptually and substantively using demonstration data. The interpretations of results using multilevel FMM are also provided. The tutorial is concluded with a discussion of some important aspects of applying multilevel FMM.

Trust repertoires and the reception of institutional responses to the COVID-19 crisis in Europe: A latent class analysis

This article examines the role of institutional trust in current European societies. Based on a secondary data analysis of Eurobarometer data (response rate 39.6%), it maps institutional trust repertoires and analyzes their consequences for a crisis that disturbed public life immensely in 2020 and 2021—the COVID-19 pandemic and the measures to fight this. Methodologically, it applies a multilevel latent class analysis of 18 institutions. Taking inspiration from “cultural backlash” theory, the explanatory analyses incorporate socio-political values and geographical identifications. The results show that there are seven different trust repertoires in the European Union (EU) countries, ranging from 24 percent mostly trustful to 11 percent mostly distrustful. EU Countries can be clustered into four classes, each with specific repertoire distributions. Particularly satisfaction with one’s own life and world developments is associated with higher trust. Compliance with COVID-19 policies is most likely when citizens trust both national political institutions and media institutions; other institutions matter less. Country health expenditure has a limited effect on the reception of COVID-19 policies but does influence membership of trust repertoires.

Towards a typology of European migration attitudes across time and space: a person-centred approach

ABSTRACT Citizens’ attitudes towards migration have received ample scholarly and public attention in recent years. While there is a growing consensus on what is driving such migration attitudes, we know little about the heterogeneity of these attitudes. Several scholars have used and operationalised terms such as ‘migration mood’, ‘policy preferences’, or ‘anti-foreigner sentiment’ interchangeably, while others have argued that citizens have distinct preferences, depending on the type of immigrant, and on what policies, or consequences of migration are studied. We shed new light on this debate, by suggesting a person-centred approach. We investigate if a typology of attitudes towards migrants and migration at the individual and country level is present, by applying multilevel latent class analysis to European Social Survey data (2002–2020). We uncover three distinct profiles of migration preferences at the individual level, and three specific country level clusters. Altogether, Europeans make a distinction between policy preferences, and their threat perceptions of out-groups, though most Europeans are either generally in favour or against immigration. Further, we demonstrate that there has been a shift in attitudinal profiles at the individual and at the country level: the group of Europeans that is generally positive about migration has grown in size.

Revisiting the Structural (In)Variances of Homicide: Examining the Differential Effects of Context Across Homicide Types

Objectives Extend foundational work on the structural covariates of homicide (concentrated disadvantage in particular) by examining the differential effects of context across distinct types of incidents. Methods Using data on 31,513 incidents nested within 4,598 places from the National Violent Death Reporting System (NVDRS) from 2003 to 2018, multilevel latent class analysis (MLCA) classified homicides into types. Two-level negative binomial regression models subsequently examined the place-level correlates of homicide counts disaggregated by type. Results Four empirically distinct types of incidents were identified, including felony, argument, intimate partner violence, and child abuse homicides. While concentrated disadvantage has a positive relationship with all four types, its effect is significantly stronger for felony and argument homicides relative to incidents involving intimate partner violence and child abuse. Conclusions The results indicate that context exhibits differential effects across homicide types, in turn suggesting that their underlying social processes and mechanisms are distinct.

A two-step estimator for multilevel latent class analysis with covariates

We propose a two-step estimator for multilevel latent class analysis (LCA) with covariates. The measurement model for observed items is estimated in its first step, and in the second step covariates are added in the model, keeping the measurement model parameters fixed. We discuss model identification, and derive an Expectation Maximization algorithm for efficient implementation of the estimator. By means of an extensive simulation study we show that (1) this approach performs similarly to existing stepwise estimators for multilevel LCA but with much reduced computing time, and (2) it yields approximately unbiased parameter estimates with a negligible loss of efficiency compared to the one-step estimator. The proposal is illustrated with a cross-national analysis of predictors of citizenship norms.

Families with young children in homeless shelters: Developmental contexts of multisystem risks and resources.

We apply a multisystem perspective to three aims relevant to resilience for young children in emergency and transitional homeless shelters. We consider profiles of risks and resources before shelter, early childhood program enrollment during shelter, and the likelihood of returning to shelter or having a subsequent child welfare placement. We used longitudinal, city-wide data from multiple sources integrated at the individual level across the lifespan for 8 birth cohorts. Young children (N = 1,281) stayed in family shelters during an 18-month period during a multisystem intervention. Risk factor rates were high as were rates of early childhood program enrollment (66.1% in any program; 42.3% in a high-quality program), which may suggest positive effects of the multisystem intervention. Multilevel latent class analysis revealed four profiles, considering prior shelter stays, prior child welfare placements, prior elevated lead levels, perinatal factors (teenage mother, prenatal care, low maternal education, and poor birth outcomes), demographics, and early childhood program enrollment and quality. One profile with higher rates of child welfare placement before the shelter stay and considerable enrollment in high-quality early childhood programs corresponded to lower rates of subsequent child welfare placement. Profiles did not differ on the likelihood of returning to shelter.

Reasons for not drinking among young adults with simultaneous alcohol and cannabis use: A latent class analysis applied to daily diary data

Simultaneous alcohol and cannabis use (i.e., simultaneous use) is prevalent among young adults and often associated with negative consequences. Understanding reasons for not drinking (RND) may provide insight into a key intervention target for reducing negative consequences associated with simultaneous use. RND may vary on a day-to-day level, and multiple RND may be endorsed on a given day. Latent class analysis (LCA) of daily diary data is a nuanced approach that can identify complex patterns of daily RND as well as its day- and person-level covariates. The current study was a secondary data analysis of daily diary data from young adults who engaged in heavy drinking and recent simultaneous use (n = 154). We aimed to: (1) characterize daily RND, (2) use LCA to classify day-level patterns of RND, and (3) compare latent classes on same-day variables (i.e., positive and negative affect, day of the week), previous-day variables (i.e., substance use, intoxication level, consequences), and person-level characteristics (i.e., age, sex, baseline substance use frequency, simultaneous use motives). Participants completed up to 14 consecutive diaries. Multilevel LCA identified four classes of heterogeneous daily RND profiles. Daily RND classes significantly differed in terms of day of the week, previous day quantity of cannabis use, and several baseline variables (age, typical substance use, simultaneous use motives). Study findings offer preliminary support for heterogeneous RND classes among young adults engaging in simultaneous use and suggest multiple avenues for future research.

Modeling careless responding in ambulatory assessment studies using multilevel latent class analysis: Factors influencing careless responding.

As the number of studies using ambulatory assessment (AA) has been increasing across diverse fields of research, so has the necessity to identify potential threats to AA data quality such as careless responding. To date, careless responding has primarily been studied in cross-sectional surveys. The goal of the present research was to identify latent profiles of momentary careless responding on the occasion level and latent classes of individuals (who differ in the distribution of careless responding profiles across occasions) on the person level using multilevel latent class analysis (ML-LCA). We discuss which of the previously proposed indices seem promising for investigating careless responding in AA studies, and we show how ML-LCA can be applied to model careless responding in intensive longitudinal data. We used data from an AA study in which the sampling frequency (3 vs. 9 occasions per day, 7 days, n = 310 participants) was experimentally manipulated. We tested the effect of sampling frequency on careless responding using multigroup ML-LCA and investigated situational and respondent-level covariates. The results showed that four Level 1 profiles ("careful," "slow," and two types of "careless" responding) and four Level 2 classes ("careful," "frequently careless," and two types of "infrequently careless" respondents) could be identified. Sampling frequency did not have an effect on careless responding. On the person (but not the occasion) level, motivational variables were associated with careless responding. We hope that researchers might find the application of an ML-LCA approach useful to shed more light on factors influencing careless responding in AA studies. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Supplemental Material for Modeling Careless Responding in Ambulatory Assessment Studies Using Multilevel Latent Class Analysis: Factors Influencing Careless Responding

Supplemental Material for Modeling Careless Responding in Ambulatory Assessment Studies Using Multilevel Latent Class Analysis: Factors Influencing Careless Responding

The avant-garde consumers: A new perspective on quality evaluations of performing arts

This article further investigates a well-established issue in consumer research concerning cultural offerings, namely the relationship between expert (critics) and audience quality evaluations. We provide a new perspective by arguing that such a relationship must consider the heterogeneity of consumers and experts alike. This article addresses the issue of quality evaluation of performing arts, and the purpose is to determine the relationship between, on the one hand, the criteria adopted for quality assessment, and on the other hand, the preferences for particular theater performances. This is done by simultaneously considering heterogeneity among the consumers and among the critics, which is a new approach in the literature. Using data from newspaper critics’ evaluations and from surveys completed by a sample of operagoers at the Royal Danish National Theater, we perform a multilevel latent class analysis, and we find three distinctive segments: The the unexperienced audience, the conventional connoisseurs and the unconventional avant-gardes. The avant-garde consumers and experts are localized as a distinct segment of audience, with different preferences and different evaluation criteria than the other audience segments (the conventional connoisseurs and the unexperienced audience). Newness and innovation are the most important criteria for the avant-garde consumer segment. The goal of an opera house or a theater is not just to increase attendance and revenue, but equally to innovate the artform. In this perspective, the avant-garde consumers are an important segment for the opera houses and producers of other cultural products, and a better understanding of the early adopters of artistic innovation (the avant-garde consumers), their preferences, evaluation criteria, and behavior, would seem to be crucial information. We suggest future empirical research could focus on avant-garde consumers as an important segment of the audience.

The Impact of Celebrity Endorsements on the Cryptocurrency Industry Through a Multi-Level Latent Class Analysis and Software-Based Approach. (Multidisciplinary Study)

The Impact of Celebrity Endorsements on the Cryptocurrency Industry Through a Multi-Level Latent Class Analysis and Software-Based Approach. (Multidisciplinary Study)

A multilevel person-centered examination of students' learning anxiety and its relationship with student background and school factors

A person-centered approach was adopted to identify distinct classes of learning anxiety (LA) using 158,578 Grade 8 students in 1097 schools in China. Through a nonparametric multilevel latent class analysis, it identified a best-fitting model consisting of three student-level latent classes (low, moderate, and high LA) and three school-level latent classes (rare, moderate, and intense LA schools). To further examine the effects of potential covariates on the student and school latent classes of LA, a multinomial logistic regression was conducted. The main results were, first, girls exhibited a higher odds ratio in moderate and high student LA classes. What's more, students' total academic scores were positively related to their odds of being in the moderate LA class while negatively related to their odds of being in the high LA class. Lastly, school leadership, school belongingness, and teacher-student relationship were all negatively correlated with the odds that schools belong to the intense and moderate LA schools.

Social inequalities in multimorbidity patterns in Europe: A multilevel latent class analysis using the European Social Survey (ESS)

Multimorbidity is associated with lower quality of life, greater disability and higher use of health services and is one of the main challenges facing governments in Europe. There is a need to identify and characterize patterns of chronic conditions and analyse their association with social determinants not only from an individual point of view but also from a collective point of view. This paper aims to respond to this knowledge gap by detecting patterns of chronic conditions and their social determinants in 19 European countries from a multilevel perspective. We used data from the ESS round 7. The final sample consisted of 18,933 individuals over 18 years of age, and patterns of multimorbidity from 14 chronic conditions were detected through Multilevel Latent Class Analysis, which also allows detecting similarities between countries. Gender, Age, Housing Location, Income Level and Educational Level were used as individual covariates to determine possible associations with social inequalities. The goodness-of-fit indices derived in a model with six multimorbidity patterns and five countries clusters. The six patterns were “Back, Digestive and Headaches”, “Allergies and Respiratory”, “Complex Multimorbidity”, “Cancer and Cardiovascular”, “Musculoskeletal” and “Cardiovascular”; the five clusters could be associated with some geographical areas or welfare states. Patterns showed significant differences in the covariates of interest, with differences in education and income being of particular interest. Some significant differences were found among patterns and the country groupings. Our findings show that chronic diseases tend to appear in a combined and interactive way, and socioeconomic differences in the occurrence of patterns are not only of the individual but also of group importance, emphasising how the welfare states in each country can influence in the health of their inhabitants.

Latent classes of men's intimate partner violence perpetration and attitudes towards gender norms: A UN multi-country, cross-sectional study in Asia and the Pacific.

ObjectiveTo examine distinct patterns of IPV perpetration and examined gender equitable attitudes as a correlate of these patterns among men from six countries in Asia and the Pacific.Design2011–12 UN Multi-country Study on Men and Violence cross-sectional study.SettingHouseholds in Bangladesh, Cambodia, China, Indonesia, Sri Lanka and Papua New Guinea.Participants10,178 men aged 18–49 years residing in Bangladesh, Cambodia, China, Indonesia, Sri Lanka and Papua New Guinea.Primary outcomes measureOur primary outcome was distinct patterns of IPV perpetration which were derived from multilevel latent class analyses.ResultsThe odds of being assigned to the Low All Forms of IPV Perpetration class than the High All Forms of IPV Perpetration class was lower for men in the middle tertile group than men in the high tertile group for gender equitable attitudes. The odds of being assigned to the High Emotional IPV Perpetration class than the High All Forms of IPV Perpetration class was greater for men in the low tertile group than men in the high tertile group for gender equitable attitudes. The odds of being assigned to the High Physical/Emotional/Economic IPV Perpetration class than the High All Forms of IPV Perpetration class was lower for men in the low tertile group than men in the high tertile group for gender equitable attitudes.ConclusionsGender transformative interventions that use an adaptive, personalized approach to men’s typology of IPV perpetration may be beneficial to reduce violence against for women in the Asia-Pacific region.

Typology of Cybercrime Victimization in Europe: A Multilevel Latent Class Analysis

The present study reveals hidden patterns of group membership across cybercrime victims in European countries. We used a multilevel latent class analysis of data from the 2019 Eurobarometer, a regionally representative sample of 21,908 individuals from 28 countries, to identify such subgroups and patterns of cybercrime victimization. The findings suggest there are two overarching cybercrime victim profiles in Europe based on individuals’ levels of “online activity” and “cybersecurity guardianship”: the “at-risk class” (19%, higher risk) and the “cautious class” (81%, lower risk). Ten different types of cybercrime victimization were compared, and while individual-level predictors were primarily used to produce different groups of cybercrime victims, our findings suggest that researchers consider both individual- and country-level predictors to understand cybercrime victimization patterns in greater depth.

Who Benefits from School-Based Teen Pregnancy Prevention Programs? Examining Multidimensional Moderators of Program Effectiveness Across Four Studies.

Recent research has suggested the importance of understanding for whom programs are most effective (Supplee et al., 2013) and that multidimensional profiles of risk and protective factors may moderate the effectiveness of programs (Lanza & Rhoades, 2012). For school-based prevention programs, moderators of program effectiveness may occur at both the individual and school levels. However, due to the relatively small number of schools in most individual trials, integrative data analysis across multiple studies may be necessary to fully understand the multidimensional individual and school factors that may influence program effectiveness. In this study, we applied multilevel latent class analysis to integrated data across four studies of a middle school pregnancy prevention program to examine moderators of program effectiveness on initiation of vaginal sex. Findings suggest that the program may be particularly effective for schools with USA-born students who speak another language at home. In addition, findings suggest potential positive outcomes of the program for individuals who are lower risk and engaging in normative dating or individuals with family risk. Findings suggest potential mechanisms by which teen pregnancy prevention programs may be effective.

Validity of Comparative Judgment Scores: How Assessors Evaluate Aspects of Text Quality When Comparing Argumentative Texts

The advantage of comparative judgment is that it is particularly suited to assess multidimensional and complex constructs as text quality. This is because assessors are asked to compare texts holistically and to make a quality judgment for each text in a pairwise comparison based upon on the most salient and critical differences. Also, the resulted rank order is based on the judgment of all assessors, representing the shared consensus. In order to be able to select the right number of assessors, the question is to what extent the conceptualization of assessors prevails in the aspects they base their judgment on, or whether comparative judgment minimizes the differences between assessors. In other words, can we detect types of assessors who tend to consider certain aspects of text quality more often than others? A total of 64 assessors compared argumentative texts, after which they provided decision statements on what aspects of text quality had informed their judgment. These decision statements were coded on six overarching themes of text quality: argumentation, organization, language use, language conventions, source use, references, and layout. Using a multilevel-latent class analysis, four different types of assessors could be distinguished: narrowly focused, broadly focused, source-focused, and language-focused. However, the analysis also showed that all assessor types mainly focused on argumentation and organization, and that assessor types only partly explained whether the aspect of text quality was mentioned in a decision statement. We conclude that comparative judgment is a strong method for comparing complex constructs like text quality. First, because the rank order combines different views on text quality, but foremost because the method of comparative judgment minimizes differences between assessors.

Profiles of instructional quality in primary and secondary education: Patterns, predictors, and relations to student achievement and motivation in science

Teachers implement different types of instruction, and the quality of their instruction is crucial for enhancing student outcomes. However, studies examining various patterns of teachers’ instructional quality are scarce, particularly in science teaching. Even fewer studies have investigated the nature of instructional quality in primary and secondary education. This study analysed the Norwegian data from the Trends in Mathematics and Science Study (TIMSS) 2015, using Grades 5 and 9 from primary and secondary education. Using multi-level latent class analysis, we identified distinct profiles of instructional quality that focused on four aspects of instructional clarity: clarity of instruction and teacher support (i.e., engaging teaching, social and emotional support, and subject domain support) at the student and classroom levels. The findings showed that the patterns of these profiles varied across different aspects of instructional quality in both grades. Further analyses revealed that student characteristics, particularly language at home and socioeconomic status, predicted the profile memberships at the student level, whereas teacher competence (i.e., self-efficacy in science teaching) predicted the profile memberships at the classroom level. In addition, different profiles of instructional quality were significantly related to motivation and, to a certain extent, achievement in science. We discuss the implications of these findings for research on instructional quality and for the design of personalized professional development programmes that aim to improve teacher instruction in primary and secondary science classrooms.

A five-factor model of perseverative thought.

Like diagnostic status, clinically relevant thought remains overwhelmingly conceptualized in terms of discrete categories (e.g., worry, rumination, obsessions). However, definitions can vary widely. The area of perseverative thought (or clinically relevant thought more broadly) would benefit substantially from a consensus-based, empirically grounded taxonomy similar to the Hierarchical Taxonomy of Psychopathology (Kotov et al., 2017) or the Big Five for personality. This article addresses three major barriers to establishing such a taxonomy: (a) a lack of research explicitly comparing categorical (subtype) versus dimensional models, (b) primary reliance on between-person measures rather than modeling at the level of the thought (within person), and (c) insufficient emphasis on replication and refinement. Participants included an unselected crowdsourced sample (790 observations from 286 participants) and an independent anxious-depressed replication sample (808 observations from 277 participants). Participants made dimensional ratings for three idiographic clinically relevant thoughts on a range of features. Multilevel latent class analysis and multilevel exploratory factor analysis were applied to identify and extract natural patterns of covariation among features at the level of the thought, controlling for person-level tendencies. A consistent five-dimension solution emerged across both samples and reliably outperformed the best-fitting categorical solution in terms of fit, replicability, and explanatory power. Identified dimensions were dyscontrol, self-focus, valence, interpersonal, and uncertainty. Findings support a five-factor latent structure of perseverative thought. Theoretical, empirical, and clinical implications and future directions are discussed. (PsycInfo Database Record (c) 2022 APA, all rights reserved).