Factor Mixture Modeling Research Articles

Symptom Dimensions and Cognitive Impairments in Individuals at Clinical High Risk for Psychosis

BackgroundUnderstanding the intricate relationships between symptom dimensions, clusters, and cognitive impairments is crucial for early detection and intervention in individuals at clinical high risk for psychosis. This study delves into this complex interplay in a clinical high risk sample with the aim of predicting the conversion to psychosis. MethodsA comprehensive cognitive assessment was performed in 744 clinical high risk individuals. The study included a 3-year follow-up period to allow assessment of conversion to psychosis. Symptom profiles were determined using the Structured Interview for Prodromal Syndromes. By applying factor analysis, symptom dimensions were categorized as dominant negative symptoms (NSs), positive symptoms-stressful, and positive symptoms-odd. The factor scores were used to define 3 dominant symptom groups. Latent class analysis (LCA) and factor mixture model (FMM) were employed to identify discrete clusters based on symptom patterns. The 3-class solution was chosen for the LCA and FMM analysis. ResultsIndividuals in the dominant NS group exhibited significantly higher conversion rates to psychosis than those in the other groups. Specific cognitive variables, including performance on the Brief Visuospatial Memory Test–Revised (odds ratio = 0.702, p = .001) and Neuropsychological Assessment Battery Mazes Test (odds ratio = 0.776, p = .024), significantly predicted conversion to psychosis. Notably, cognitive impairments associated with NSs and positive symptoms-stressful affected different cognitive domains. LCA and FMM cluster 1, which was characterized by severe NSs and positive symptoms-odd, exhibited more impairments in cognitive domains than other clusters. No significant difference in the conversion rate was observed among the LCA and FMM clusters. ConclusionsThese findings highlight the importance of NSs in the development of psychosis and suggest specific cognitive domains that are affected by symptom dimensions.

A novel CFA + EFA model to detect aberrant respondents

Abstract Aberrant respondents are common but yet extremely detrimental to the quality of social surveys or questionnaires. Recently, factor mixture models (FMMs) have been employed to identify individuals providing deceptive or careless responses. We propose a comprehensive FMM for continuous outcomes that combines confirmatory and exploratory factor models to classify both the nonaberrant and aberrant respondents. The flexibility of the proposed classification model allows for the identification of two of the most common aberrant response styles, namely faking and careless responding. We validated our approach by means of two simulations and two case studies. The results indicate the effectiveness of the proposed model in dealing with aberrant responses in social and behavioural surveys.

A latent profiling approach to foreign language enjoyment and L2 achievement

ABSTRACT Both Broaden-and-Build theory and Control-Value theory (CVT) posit that enjoyment contributes to academic achievement. However, empirical studies show mixed results concerning the predictive effects between these constructs in Second Language Acquisition (SLA). This study examines the link between second language (L2) achievement and Foreign Language Enjoyment (FLE) among 1,360 Chinese tertiary-level students learning English as a foreign language (EFL). Participants completed a national standardised English language test and an FLE questionnaire. Data analysis involved the use of Mplus to perform factor mixture modelling (FMM), a technique that identifies latent subgroups within the sample based on their response patterns. Results revealed four distinct profiles, each displaying different FLE-L2 achievement profile characteristics. The importance of this study lies in its analytic approach, which recognises the non-uniform connections between the three dimensions of FLE and L2 achievement across learners. This finding challenges the notion of straightforward correlations and sheds light on the complexities of these relationships, thereby addressing inconsistencies observed in previous studies and paving the way for more nuanced future research.

A Systematic Review of Factor Mixture Model Applications

In this study, a systematic review was conducted on peer-reviewed articles of factor mixture model (FMM) applications. A total of 304 studies were included with 334 applications published from 2003–2022. FMM was mostly used in these studies to detect latent classes and model heterogeneity. Most of the studies were conducted in the U.S. with different populations, including students, adults, and the general population. The average sample size was 3,562, and the average number of items was 17.34. Measurement tools containing mostly Likert type items and measuring structures in the field of psychology were used in these FMM analyses. Most FMM studies that were reviewed were applied with maximum likelihood estimation methods as implemented in Mplus software. Most studies reported multiple fit indices, the most common of which were AIC, BIC, and entropy. The mean numbers of classes and factors across the 334 applications were 2.96 and 2.17, respectively. Psychological and behavioral disorders, gender, and age variables were mostly the focus in the FMM studies and included use of covariates these analyses. As a result of this systematic review, the trends in FMM analyses were better understood, and some wrong practices in reporting were revealed.

Measuring psychotherapeutic processes in the context of psychedelic experiences: Validation of the General Change Mechanisms Questionnaire (GCMQ).

Therapeutic and salutogenic effects of psychedelic drugs have been attributed to psychotherapeutic or psychotherapy-like processes that can unfold during the acute psychedelic experience and beyond. Currently, there are no psychometric instruments available to comprehensively assess psychotherapeutic processes (as conceptualized by empirical psychotherapy research) in the context of psychedelic experiences. We report the initial validation of the General Change Mechanisms Questionnaire (GCMQ), a self-report instrument designed to measure five empirically established general change mechanisms (GCMs) of psychotherapy-(1) resource activation, (2) therapeutic relationship, (3) problem actuation, (4) clarification, and (5) mastery-in the context of psychedelic experiences. An online survey in a sample of 1153 English-speaking and 714 German-speaking psychedelic users was conducted to evaluate simultaneously developed English- and German-language versions of the GCMQ. The theory-based factor structure was confirmed. The five GCMQ scales showed good internal consistency. Evidence for convergent validity with external measures was obtained. Significant associations with different settings and with therapeutic, hedonic, and escapist use motives confirmed the hypothesized context dependence of GCM-related psychedelic experiences. Indicating potential therapeutic effects, the association between cumulative stressful life events and well-being was significantly moderated by resource activation, clarification, and mastery. Factor mixture modeling revealed five distinct profiles of GCM-related psychedelic experiences. Initial testing indicates that the GCMQ is a valid and reliable instrument that can be used in future clinical and nonclinical psychedelic research. The five identified profiles of GCM-related experiences may be relevant to clinical uses of psychedelics and psychedelic harm reduction.

Heterogeneity of Alexithymia Subgroups

Abstract: Alexithymia is a subclinical experience in which individuals struggle to identify, distinguish, and describe their own emotions. It is most commonly measured with the self-reported Toronto Alexithymia Scale-20 (TAS-20). However, scholars hold different views on its structure, resulting in challenges in classifying individuals with alexithymia, which is detrimental to clinical diagnosis, counseling, and intervention. The present study aimed to investigate the types (or subgroups) of alexithymia within a sample of college students ( n = 707) from four Chinese universities. Two latent classes of three-factor two-class model solution were effectively identified by the Factor Mixture Model (FMM) approach: a “High-EOT alexithymia” class (18.2%) and a “Non-alexithymia” class (81.8%). The two subgroups exhibited similar performance in difficulty in identifying feelings (DIF) and difficulty in describing feelings (DDF), but they differed significantly in externally oriented thinking (EOT). This suggests that EOT might be a diagnostic criterion for alexithymia.

Clinical profiles of adolescent personality pathology: a latent structure examination of the Semi-Structured Interview for Personality Functioning DSM-5 (STiP-5.1) in a help-seeking sample

BackgroundDespite the introduction of dimensional conceptualisations of personality functioning in the latest classification systems, such as Criterion A of the Alternative Model of Personality Disorders in the DSM-5, heterogeneous clinical presentation of personality pathology remains a challenge. Relatedly, the latent structure of personality pathology as assessed by the Semi-Structured Interview for Personality Functioning DSM-5 (STiP-5.1) has not yet been comprehensively examined in adolescents. Therefore, this study aimed to examine the latent structure of the STiP-5.1, and, based on those findings, to describe any unique clinical profiles that might emerge.MethodsThe final sample comprised 502 participants aged 11–18 years consecutively recruited from a specialised personality disorder outpatient service, as well as general day clinic and inpatient wards at the University Hospital University Hospital of Child and Adolescent Psychiatry and Psychotherapy, University Hospital Bern, Switzerland. Participants were assessed using the STiP-5.1, as well as a battery of other psychological measures by clinical psychologists or trained doctoral students. Variations of Factor Analysis, Latent Class Analysis and Factor Mixture Models (FMM) were applied to the STiP-5.1 to determine the most appropriate structure.ResultsThe best fitting model was an FMM comprising four-classes and two factors (corresponding to self- and interpersonal-functioning). The classes differed in both overall severity of personality functioning impairment, and in their scores and clinical relevance on each element of the STiP-5.1. When compared to the overall sample, classes differed in their unique clinical presentation: class 1 had low impairment, class 2 had impairments primarily in self-functioning with high depressivity, class 3 had mixed levels of impairment with emerging problems in identity and empathy, and class 4 had severe overall personality functioning impairment.ConclusionsA complex model incorporating both dimensional and categorical components most adequately describes the latent structure of the STiP-5.1 in our adolescent sample. We conclude that Criterion A provides clinically useful information beyond severity (as a dimensional continuum) alone, and that the hybrid model found for personality functioning in our sample warrants further attention. Findings can help to parse out clinical heterogeneity in personality pathology in adolescents, and help to inform early identification and intervention efforts.

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.

The distribution of parent-reported attention-deficit/hyperactivity disorder and subclinical autistic traits in children with and without an ADHD diagnosis.

Autistic traits are often reported to be elevated in children diagnosed with attention-deficit/hyperactivity disorder (ADHD). However, the distribution of subclinical autistic traits in children with ADHD has not yet been established; knowing this may have important implications for diagnostic and intervention processes. The present study proposes a preliminary model of the distribution of parent-reported ADHD and subclinical autistic traits in two independent samples of Australian children with and without an ADHD diagnosis. Factor mixture modelling was applied to Autism Quotient and Conners' Parent Rating Scale - Revised responses from parents of Australian children aged 6-15years who participated in one of two independent studies. A 2-factor, 2-class factor mixture model with class varying factor variances and intercepts demonstrated the best fit to the data in both discovery and replication samples. The factors corresponded to the latent constructs of 'autism' and 'ADHD', respectively. Class 1 was characterised by low levels of both ADHD and autistic traits. Class 2 was characterised by high levels of ADHD traits and low-to-moderate levels of autistic traits. The classes were largely separated along diagnostic boundaries. The largest effect size for differences between classes on the Autism Quotient was on the Social Communication subscale. Our findings support the conceptualisation of ADHD as a continuum, whilst confirming the utility of current categorical diagnostic criteria. Results suggest that subclinical autistic traits, particularly in the social communication domain, are unevenly distributed across children with clinically significant levels of ADHD traits. These traits might be profitably screened for in assessments of children with high ADHD symptoms and may also represent useful targets for intervention.

Developmental profiles of arithmetic fluency skills from grades 1 to 9 and their early identification.

The aim of the present study was to examine the kinds of developmental profiles of arithmetic fluency skills that can be identified across Grades 1-9 (ages 7-16) in a large Finnish sample (n = 2,518). The study also examined whether membership in the developmental profiles could be predicted using a comprehensive set of kindergarten-age factors, including information on cognitive skills; motivational, parental, and home environment factors; and gender. Four profiles of arithmetic fluency skills development were identified using a factor mixture model: persistent arithmetic difficulties (12.23%), precocious onset (50.24%), delayed onset (36.96%), and precocious onset with a Grade 7 drop (.06%). The Cholesky models predicting membership in the three largest profiles suggested that overall, the strongest kindergarten-age predictors were cognitive skills (especially counting, number concepts, spatial relations, rapid automatized naming [RAN], phonological awareness, and letter knowledge), but motivational, parental, and home environment factors were also significant. Membership in the profile with precocious onset was predicted by most of the kindergarten-age measures, suggesting that the strengths in early skills, as well as motivational, parental, and home environment factors, are reflected in the advanced start in arithmetic development at school. The profiles with delayed onset and persistent difficulties were similar in most kindergarten-age measures but differed in task avoidance and four cognitive skills (letter knowledge, counting, number concepts, and RAN), suggesting that these factors predict differential development over the longer term. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Symptoms and prevalence of common mental disorders in a heterogenous outpatient sample: an investigation of clinical characteristics and latent subgroups

BackgroundPatient-reported outcome measures (PROM) provide clinicians with information about patients’ perceptions of distress. When linked with treatment and diagnostic registers, new information on common mental health disorders (CMHD) and service use, may be obtained, which might be useful clinically and for policy decision-making. This study reports the prevalence of CMHD and their association with PROM severity. Further, subgroups of self-reported symptoms of depression and anxiety were examined, and their association with clinician-assessed mental disorders, functional impairment, and service use.MethodsIn a cohort study of 2473 (63% female) outpatients, CMHD was examined with pre-treatment scores of self-reported depression and anxiety, and the number of assessments and psychotherapy appointments one year after treatment start. Factor mixture modelling (FMM) of anxiety and depression was used to examine latent subgroups.ResultsOverall, 22% of patients with a CMHD had an additional comorbid mood/anxiety disorder, making the prevalence lower than expected. This comorbid group reported higher symptoms of anxiety and depression compared to patients with non-comorbid disorders. FMM revealed three classes: “anxiety and somatic depression” (33%), “mixed depression and anxiety” (40%), and “cognitive depression” (27%). The anxiety and somatic depression class was associated with older age, being single and on sick leave, higher probability of depressive-, anxiety-, and comorbid disorders, having more appointments and higher functional impairment. Although the cognitive depression class had less somatic distress than the mixed depression and anxiety class, they reported more functional impairment and had higher service use.ConclusionThe results show that higher levels of somatic symptoms of depression could both indicate higher and lower levels of functional impairment and service use. A group of patients with high somatic depression and anxiety was identified, with severe impairment and high service needs. By gaining insights into CMHD factors’ relation with clinical covariates, self-reported risk factors of depression and anxiety could be identified for groups with different levels of aggravating life circumstances, with corresponding service needs. These could be important symptom targets in different groups of patients.

Incorporating machine learning into factor mixture modeling: Identification of covariate interactions to explain population heterogeneity

Factor mixture modeling (FMM) has been widely adopted in health and behavioral sciences to examine unobserved population heterogeneity. Covariates are often included in FMM as predictors of the latent class membership via multinomial logistic regression to help understand the formation and characterization of population heterogeneity. However, interaction effects among covariates have received considerably less attention, which might be attributable to the fact that interaction effects cannot be identified in a straightforward fashion. This study demonstrated the utility of structural equation model or SEM trees as an exploratory method to automatically search for covariate interactions that might explain heterogeneity in FMM. That is, following FMM analyses, SEM trees are conducted to identify covariate interactions. Next, latent class membership is regressed on the covariate interactions as well as all main effects of covariates. This approach was demonstrated using the Traumatic Brain Injury Model System National Database.

A Constrained Factor Mixture Model for Detecting Careless Responses that is Simple to Implement

Using constrained factor mixture models (FMM) for careless response identification is still in its infancy. Existing models have overly restrictive statistical assumptions that do not identify all types of careless respondents. The current paper presents a novel constrained FMM model with more reasonable assumptions that capture both longstring and random careless respondents. We provide a comprehensive comparison of the statistical assumptions between the proposed model and two previous constrained models. The proposed model was evaluated using both real data ( N = 1,455) and statistical simulation. The results showed that the model had a superior fit, stronger convergent validity with other indicators of careless responding, more accurate parameter recovery and more accurate identification of careless respondents when compared to its predecessors. The proposed model does not require additional data collection effort, and thus researchers can routinely use it to control careless responses. We provide user-friendly syntax with detailed explanations online to facilitate its use.

Mechanisms of cognitive disinhibition explain individual differences in adult attention deficit hyperactivity disorder traits

BackgroundAttention deficit hyperactivity disorder (ADHD) in adults is strongly associated with psychiatric comorbidity and functional impairment. Here, we aimed to use a newly developed online cognitive battery with strong psychometric properties for measuring individual differences in three cognitive mechanisms proposed to underlie ADHD traits in adults: 1) attentional control – the ability to mobilize cognitive resources to stop a prepotent motor response; 2) information sampling/gathering – adequate sampling of information in a stimulus detection task prior to making a decision; and 3) shifting - the ability to adapt behavior in response to positive and negative contingencies. MethodsThis cross-sectional and correlational study recruited 650 adults (330 males) aged 18–69 years (M = 33.06; MD = 31.00; SD = 10.50), with previously diagnosed ADHD (n = 329) and those from the general community without a history of ADHD (n = 321). Self-report measures of ADHD traits (i.e., inattention/disorganization, impulsivity, hyperactivity) and the cognitive battery were completed online. ResultsLatent class analysis, exploratory structural equation modeling and factor mixture modeling revealed self-reported ADHD traits formed a unidimensional and approximately normally distributed phenotype. Bayesian structural equation modeling demonstrated that all three mechanisms measured by the cognitive battery, explained unique, incremental variance in ADHD traits, with a total of 15.9% explained in the ADHD trait factor. ConclusionsAttentional control and shifting, as well as the less researched cognitive process of information gathering, explain individual difference variance in self-reported ADHD traits with potential to yield genetic and neurobiological insights into adult ADHD.

Differences in mental health and alcohol use across profiles of COVID-19 disruptions.

This study aimed to examine differences in mental health and alcohol use outcomes across distinct patterns of work, home, and social life disruptions associated with the COVID-19 pandemic. Data from 2093 adult participants were collected from September 2020 to April 2021 as a part of a larger study examining the impacts of the COVID-19 pandemic on substance use. Participants provided data on COVID-19 pandemic experiences, mental health outcomes, media consumption, and alcohol use at baseline. Alcohol use difficulties, including problems related to the use, desire to use alcohol, failure to cut down on alcohol use, and family/friend concern with alcohol use, were measured at 60-day follow-up. Factor mixture modeling followed by group comparisons, multiple linear regressions, and multiple logistic regressions was conducted. A four-profile model was selected. Results indicated that profile membership predicted differences in mental health and alcohol use outcomes above and beyond demographics. Individuals experiencing the most disruption reported the strongest daily impact of COVID-19 and significantly high levels of depression, anxiety, loneliness, overwhelm, alcohol use at baseline, and alcohol use difficulties measured at 60-day follow-up. The findings highlight the need for integrated mental health and/or alcohol services and social services targeting work, home, and social life during public health emergencies in order to respond effectively and comprehensively to the needs of those requiring different types of support.

Misclassification Error, Binary Regression Bias, and Reliability in Multidimensional Poverty Measurement: An Estimation Approach Based on Bayesian Modelling

ABSTRACT Measurement error affects the quality of population orderings of an index and, hence, increases the misclassification of the poor and the non-poor groups and affects statistical inferences from binary regression models. Hence, the conclusions about the extent, profile, and distribution of poverty are likely to be misleading. However, the size and type (false positive/negatives) of classification error have remained untraceable in poverty research. This paper draws upon previous theoretical literature to develop a Bayesian-based estimator of population misclassification and binary-regression coefficient bias. The study uses the reliability values of existing poverty indices to set up a Monte Carlo study based on factor mixture models to illustrate the connections between measurement error, misclassification, and bias and evaluate the procedure and discusses its importance for real-world applications.

Heterogeneity in children at risk of math learning difficulties.

This study recruited 428 Singaporean children at risk of math learning difficulties (MLD; Mage = 83.9 months, SDage = 4.35 months; 41% female). Using a factor mixture model that considered both quantitative and qualitative differences in math ability, two qualitatively different groups were identified: one with generalized difficulties across different math skills and the other with more focal difficulties in arithmetic fluency. Reading, working memory capacity, and numeracy (number line estimation skills and numerical discrimination) uniquely explained group membership. Children within each group differed in the extent of difficulties they exhibited, with numeracy variables differentially contributing to math ability in each group. Findings speak against a dimensional view of MLD and underscore the conceptual limitations of using basic numeracy performance to profile learning difficulties.

Procedures for Analyzing Multidimensional Mixture Data

The multidimensional mixture data structure exists in many test (or inventory) conditions. Heterogeneity also relatively exists in populations. Still, some researchers are interested in deciding to which subpopulation a participant belongs according to the participant’s factor pattern. Thus, in this study, we proposed three analysis procedures based on the factor mixture model to analyze data in the multidimensional mixture context. Simulations were manipulated with different levels of factor numbers, factor correlations, numbers of latent classes, and class separation. Issues with regard to model selection were discussed at first. The results showed that in the two-class situations the procedures of “factor structure first then class number” (Procedure 1) and “factor structure and class number considered simultaneously” (Procedure 3) performed better than the “class number first then factor structure” (Procedure 2) and yielded precise parameter estimation and classification accuracy. It would be appropriate to choose Procedures 1 and 3 when strong measurement invariance is assumed while using an information criterion, but Procedure 1 saved more time than Procedure 3. In the three-class situations, the performance of all three procedures was limited. Implementations and suggestions have been addressed in this research.

Value-based decision-making network functional connectivity correlates with substance use and delay discounting behaviour among young adults

Substance use disorders are characterized by reduced control over the quantity and frequency of psychoactive substance use and impairments in social and occupational functioning. They are associated with poor treatment compliance and high rates of relapse. Identification of neural susceptibility biomarkers that index risk for developing a substance use disorder can facilitate earlier identification and treatment. Here, we aimed to identify the neurobiological correlates of substance use frequency and severity amongst a sample of 1,200 (652 females) participants aged 22–37 years from the Human Connectome Project. Substance use behaviour across eight classes (alcohol, tobacco, marijuana, sedatives, hallucinogens, cocaine, stimulants, opiates) was measured using the Semi-Structured Assessment for the Genetics of Alcoholism. We explored the latent organization of substance use behaviour using a combination of exploratory structural equation modelling, latent class analysis, and factor mixture modelling to reveal a unidimensional continuum of substance use behaviour. Participants could be rank ordered along a unitary severity spectrum encompassing frequency of use of all eight substance classes, with factor score estimates generated to represent each participant’s substance use severity. Factor score estimates and delay discounting scores were compared with functional connectivity in 650 participants with imaging data using the Network-based Statistic. This neuroimaging cohort excludes participants aged 31 and over. We identified brain regions and connections correlated with impulsive decision-making and poly-substance use, with the medial orbitofrontal, lateral prefrontal and posterior parietal cortices emerging as key hubs. Functional connectivity of these networks could serve as susceptibility biomarkers for substance use disorders, informing earlier identification and treatment.

A Factor Mixture Model for Item Responses and Certainty of Response Indices to Identify Student Knowledge Profiles

AbstractThe certainty of response index (CRI) measures respondents' confidence level when answering an item. In conjunction with the answers to the items, previous studies have used descriptive statistics and arbitrary thresholds to identify student knowledge profiles with the CRIs. Whereas this approach overlooked the measurement error of the observed item responses and indices, we address this by proposing a factor mixture model that integrates a latent class model to detect student subgroups and a measurement model to control for student ability and confidence level. Applying the model to 773 seventh graders' responses to an algebra test, where some items were related to new material that had not been taught in class, we found two subgroups: (1) students who had high confidence in answering items involving the new material; and (2) students who had low confidence in answering items involving the new material but higher general self‐confidence than the first group. We regressed the posterior probability of the group membership on gender, prior achievement, and preview behavior and found preview behavior a significant factor associated with the membership. Finally, we discussed the implications of the current study for teaching practices and future research.