Latent Modeling Approach Research Articles

A single-step firm-level identification of classification shifting

PurposeThis study proposes a new model to measure unexpected core earnings, using Bayesian dynamic latent method.Design/methodology/approachThe Bayesian dynamic latent modeling approach identifies the effects that stem from complex, multidimensional variables related to culture and legal framework, on unexpected core earnings. It also allows testing whether there is persistence over time in unexpected core earnings. We use sequential Bayesian Monte Carlo methods, also known as particle filtering, that simplify estimations.FindingsIn an international empirical application, we find evidence of persistence in unexpected core earnings as well as classification shifting. The impact of the legal framework on classification shifting shows variability across samples. Religion reduces classification shifting, whereas the cultural variables of power distance, masculinity and uncertainty avoidance enhances it. Interestingly, the persistence in unexpected core earnings is strong and moderates the ability of legal framework and religion in abating classification shifting.Research limitations/implicationsIn terms of policy implications, we show that strengthening legal framework would improve financial reporting and reduce the scope for manipulation. This could involve stricter enforcement mechanisms, increased penalties for non-compliance and regular audits to detect and deter classification shifting practices. Given that religion plays a role in moderating classification shifting, policymakers may explore partnerships or collaborations with religious institutions to promote ethical financial practices. Engaging religious leaders and organizations can help emphasize the importance of integrity and ethical behavior in financial reporting, potentially influencing the behavior of individuals and organizations.Originality/valueTo the best of our knowledge this is the first study that opts for Bayesian dynamic latent model for an international sample.

Do latent profiles of self-reported anxiety, depression, and PTSD map onto clinician ratings? An examination with intimate partner violence survivors

ObjectiveStudies exploring latent profiles of mental health in trauma survivors have largely relied on self-report, making it unclear whether these patterns correspond with clinician-assessed psychopathology. The purpose of the current study was to examine latent profiles of self-reported PTSD, depression, and anxiety in a sample of 387 women who had experienced intimate partner violence (IPV) and investigate whether profiles mapped onto clinician-rated measures of the same outcomes. MethodParticipants completed a series of semi-structured interviews and self-report measures assessing PTSD, depression, and anxiety. ResultsLatent profile analyses revealed a 3-profile solution characterized by Low (22.48 %), Moderate (37.98 %), and High (39.53 %) self-reported symptomology. Clinician ratings were significant predictors of membership in the low vs. moderate vs. high symptomology profiles. However, normalized means showed discrepancies between self-report and clinician assessment regarding which issue was rated most severe. ConclusionsResults suggest that while latent modeling approaches relying on self-report may adequately approximate common underlying patterns of psychopathology, they have limitations in identifying which disorders are most salient for clinical intervention.

Improved estimates of foetal growth are associated with perinatal outcomes: A latent modelling approach in a population-based birth cohort.

We aimed to estimate latent foetal growth conditions and explore their determinants among maternal characteristics and ultrasound biometric parameters. We additionally investigated the influence of foetal growth conditions on perinatal variables. We used data from live-born singletons in the Maternal and Child Health and Nutrition in Acre, Brazil (MINA-Brazil Study) population-based birth cohort. Maternal and perinatal characteristics were assessed in medical records from the maternity hospital and interviews with participants from July 2015 to June 2016. A sub-sample went through ultrasound examinations during the antenatal period, with assessment of foetal head and abdominal circumferences, and femur length. We estimated latent foetal growth conditions with a structural equation modelling framework, informed by the child's birth weight z-scores (BWZ) and birth length z-scores (BLZ) according to gestational age. Odds ratios and 95% confidence intervals (CIs) for the occurrence of perinatal events were estimated according to linear predictions of the latent variable. We included 1253 participants. Latent foetal growth conditions explained 88.3% of BWZ and 53.7% of BLZ variation. Maternal elevated blood pressure, primiparity, smoking, malaria, and insufficient gestational weight gain negatively impacted foetal growth conditions. In the subsample (n = 499), ultrasound biometric parameters assessed at 28 weeks were positively associated with the latent variable, with the largest contribution from foetal abdominal circumference. Each standardised unit of predicted foetal growth conditions halved the chance for preterm birth (95% CI = 0.26, 0.74) and longer hospital stay (>3 days) (95% CI = 0.28, 0.88). Conversely, BWZ and BLZ were not independently associated with these perinatal variables in separate logistic regression models. Latent foetal growth conditions jointly encompassing weight gain and linear growth during gestation were negatively influenced by a scenario of dual burden of maternal morbidities, with perinatal implications.

A plausible role of imagination in pretend play, counterfactual reasoning, and executive functions.

A notable observation is the similarities in the cognitive processes of pretend play (PP) and counterfactual reasoning (CFR) as both involve thinking about alternatives to reality. It is argued by Weisberg and Gopnik (Cogn. Sci., 37, 2013, 1368) that alternative thinking in PP and CFR is underpinned by an imaginary representational capacity but few studies have empirically investigated this link. We use a variable latent modelling approach to test a hypothetical model of the structural relationship of PP and CFR predicting that if PP and CFR are cognitively similar; they should have similar patterns of associations with Executive Functions (EFs). Data were collected on PP, CFR, EFs and Language from 189 children (M=4.8 years, males=101, females=88). Confirmatory factor analyses showed that measures of PP and CFR loaded onto single latent constructs and were significantly correlated (r=.51, p=.001) with each other. Hierarchical multiple regression analyses revealed that EF accounted for unique significant variance in both PP (β=21) and CFR (β=22). The results of the structural equation modelling revealed that the data were a good fit for the hypothetical model. We discuss the plausible role of a general underlying imaginative representational capacity to explain similarities in the cognitive mechanisms of different states of alternative thinking like PP and CFR.

Bayesian joint spatial modelling of anemia and malaria in Guinea

<abstract><p>In sub-Saharan Africa, anemia and malaria are the leading causes of morbidity and mortality among children under the age of five years. Guinea is one of the countries where the two diseases have devastating effects. Both of these diseases have been studied separately, but the two diseases exhibit inherent dependence between them, therefore, modelling them in isolation negates practical reality. This study aims at jointly estimating the spatial linear correlation between anemia and malaria, as well as to investigate the differences in contextual, socioeconomic and demographic factors affecting morbidity among children under five years in Guinea. Statistical approaches are used to handle modelling of binary outcomes with allowance for spatial components and joint responses. In particular, a latent model approach is proposed in the methodology to investigate the linear correlation between anemia and malaria allowing for spatial and non-spatial effects. All the parameters are estimated using Bayesian approach based on Markov chain Monte Carlo (MCMC) technique. According to the findings, 76.15% of children under the age of five years in Guinea were anemic, and 14.31% had malaria. Furthermore, the results showed that the child's malaria status is significantly associated with the place of residence, his/her age and ownership of television as an indicator of well being. In terms of anemia in children, there was a significant association with age, mother's education level and ethnicity group of the household head. The Nzerekore region, had both high malaria and anemia prevalences in children under five years. The latent model results showed that there was weak positive correlation between anemia and malaria in Nzerekore and Boke regions. Based on the shared component model, there was a significant unobserved risk factor that both diseases share.</p></abstract>

The Italian Translation of the Supports Intensity Scale-Children (SIS-C Italian): Measurement Invariance and Differences

Abstract The Supports Intensity Scale – Children's Version (SIS-C) was translated into Italian using a committee approach to translation. Latent modeling approaches enabled the leveraging of the large standardization sample from the U.S. (n = 4,015) to generate translation-specific norms from data collected in Italy (n = 435) for children and youth ranging ages 5-16 years placed in six evenly distributed age groups by country. Findings indicated the structure of the SIS-C (i.e., seven support need domains organized under an overall support needs construct), was supported in the Italian context. However, there were age-related differences in the U.S. and Italian samples. In the Italian sample, norms were established for the 5-8 years, 9-10 years, and 11-16 years age groups. Moreover, the Italian sample also differed from other European samples and SIS-C translations. The importance of understanding cultural contexts in interpreting findings from the SIS-C is discussed, along with ways in which SIS-C findings can be used to inform policy and practice in the Italian context.

Modeling of Wire Electro-Spark Machining of Inconel 690 Superalloy Using Support Vector Machine and Random Forest Regression Approaches

Machine learning approaches are disseminating very rapidly in manufacturing industry. They emerge out as latent modeling approaches showing a great potential for solving manufacturing-related problems. This study aims to analyze and predict the micro-irregularities on the surface of Inconel 690 superalloy processed by wire electro-spark machining using machine learning approaches. Experimental runs are completed by using the box-Behnken design of response surface methodology. Spark-on time, spark-off time, current and voltage with varying constraints are used as the input process variables, whereas surface integrity is used as the output variable. Support vector machines with two different functions and random forest regression are exercised to predict the machining outcomes with the help of existing models and validate the experimental outcomes. Support vector machine with radial basis function emerged out as the best approach in predicting the process outcomes. It is also perceived that spark-on time is the most dominating input process variable.

Surveying Ethics: a Measurement Model of Preference for Precepts Implied in Moral Theories (PPIMT)

Recent research in empirical moral psychology attempts to understand (rather than place judgment on) the salient normative differences that laypeople have when making moral decisions by using survey methodology that is based on the operationalized principles from moral theories. The PPIMT is the first measure designed to assess respondents’ preference for the precepts implied in the three dominant moral theories: virtue ethics, deontology, and consequentialism. The current study used a latent modeling approach to determine the most theoretically and psychometrically-sound model for the PPIMT using a combined sample of college students from a southeastern university in U.S. and MTurk respondents. The PPIMT model fit was acceptable (χ2 = 84.125, df = 40, p = 0.001; RMSEA = 0.052, 90%CI = 0.037 to 0.068; CFI = 0.980; SRMR = 0.035) with four items for Virtue, four items for Deontology, and three items for Consequentialism.

The Supports Intensity Scale Children's Version-Icelandic Translation: Examining Measurement Properties.

An innovation in developing the Supports Intensity Scale-Children's Version (SIS-C) was the adoption of latent variable modeling approaches to norm development. In regard to translated versions of the SIS-C, the latent modeling approaches provided opportunities to leverage the large standardization sample generated in the United States (n = 4,015) to generate translation-specific norms from data collected on smaller samples in other countries and enable future cross-cultural analyses. In this study, data were collected on children in Iceland who received special education services (as defined and delivered in Iceland), a more diverse group of children with disabilities than the U.S. sample. This provided a unique context to explore cross-cultural differences. Findings indicated the structure of the SIS-C (i.e., seven support need domains organized under an overall support needs construct), was supported in the Icelandic context. However, findings also suggested that supports planning teams in Iceland must consider specific age-related factors that differ from other cultural contexts.

Adaptive estimation in structured factor models with applications to overlapping clustering

This work introduces a novel estimation method, called LOVE, of the entries and structure of a loading matrix $A$ in a latent factor model $X=AZ+E$, for an observable random vector $X\in \mathbb{R}^{p}$, with correlated unobservable factors $Z\in \mathbb{R}^{K}$, with $K$ unknown, and uncorrelated noise $E$. Each row of $A$ is scaled, and allowed to be sparse. In order to identify the loading matrix $A$, we require the existence of pure variables, which are components of $X$ that are associated, via $A$, with one and only one latent factor. Despite the fact that the number of factors $K$, the number of the pure variables and their location are all unknown, we only require a mild condition on the covariance matrix of $Z$, and a minimum of only two pure variables per latent factor to show that $A$ is uniquely defined, up to signed permutations. Our proofs for model identifiability are constructive, and lead to our novel estimation method of the number of factors and of the set of pure variables, from a sample of size $n$ of observations on $X$. This is the first step of our LOVE algorithm, which is optimization-free, and has low computational complexity of order $p^{2}$. The second step of LOVE is an easily implementable linear program that estimates $A$. We prove that the resulting estimator is near minimax rate optimal for $A$, with respect to the $\|\ \|_{\infty ,q}$ loss, for $q\geq 1$, up to logarithmic factors in $p$, and that it can be minimax-rate optimal in many cases of interest. The model structure is motivated by the problem of overlapping variable clustering, ubiquitous in data science. We define the population level clusters as groups of those components of $X$ that are associated, via the matrix $A$, with the same unobservable latent factor, and multifactor association is allowed. Clusters are respectively anchored by the pure variables, and form overlapping subgroups of the $p$-dimensional random vector $X$. The Latent model approach to OVErlapping clustering is reflected in the name of our algorithm, LOVE. The third step of LOVE estimates the clusters from the support of the columns of the estimated $A$. We guarantee cluster recovery with zero false positive proportion, and with false negative proportion control. The practical relevance of LOVE is illustrated through the analysis of a RNA-seq data set, devoted to determining the functional annotation of genes with unknown function.

Configuring perceived fit to mitigate consumer animosity in the context of cross-border sport sponsorships

ABSTRACT Research question: While cross-border sport sponsorships are widespread, such partnerships introduce a notable complication – consumers in one country may dislike the sponsor’s country of origin (COO). This raises the question as to whether animosity towards a sponsor’s COO negatively affects sponsorship outcomes, and if so, how it can be addressed. For the latter, we examine holistic sponsor-object fit as well as a set of its constituent elements. Research methods: Data collection pertained to a brand engaged in a hypothetical sponsorship. Study 1 involves a Serbian brand sponsoring the Croatia national football team and for Study 2 German sponsors of the England national football team. Survey data are analyzed using a latent modeling approach. Results and findings: Study 1 shows that animosity reduces consumers’ attitude towards the sponsorship. However, higher perceived sponsor-object fit weakens this effect. Study 2 replicates this finding, and on a more granular level establishes the moderating properties of several sub-dimensions of fit. Congruence in color, personality and status ameliorate animosity. Implications: We outline implications for sponsors operating in environments where their COO invokes animosity and how sponsor-object fit may mitigate this.

A Network Approach to Depressive Disorders

Treatments to depression often take a latent modeling approach whereby depressive symptoms are a common cause of the disorder. However, such an approach is not without problems. For example, covariance amongst indicators are deemed spurious. Network modeling has been proposed as an alternative way of explaining depression. The purpose of this article is to introduce the network approach and outline how it can be used to foster the development of effective treatment packages. Viewing depression as a network may result in a reduction in the number of patients experiencing depressive episodes.

Bullying Perpetration and Victimization: A Test of Traditional and Cyber-Behaviors as Latent Constructs.

The current investigation hypothesized and tested latent bully/victim traits for physical, verbal, or relational bullying/victimization, both cyber and traditional behaviors. Data were collected from 1,356 German students who attended Grades 5 to 10: 48.4% males, 49.3% females from eight different schools in Northern Germany. Based on two samples for cross-validation (Ntraining set = 525, Nvalidation set = 525), study findings provided strong evidence of adequate model fit, both for traditional and cyber behaviors. Consistent with the current state of knowledge, bullying and victimization latent traits highly associated, more so for cyber behaviors than traditional ones. Thus, both the theoretical plausibility as well as statistical evidence support the application of latent modeling to these behaviors. Further research is needed to replicate the applied measurement models proposed in this work and to reveal moderators or measurement invariance across diverse populations. Nevertheless, the current evidence substantiates the importance of the application of a latent modeling approach to overcome known psychometric challenges of reliability and validity in bullying research.

Investigating the genetic architecture of general and specific psychopathology in adolescence

Whilst associations between polygenic risk scores (PRSs) for schizophrenia and various phenotypic outcomes have been reported, an understanding of developmental pathways can only be gained by modelling comorbidity across psychopathology. We examine how genetic risk for schizophrenia relates to adolescent psychosis-related and internalizing psychopathology using a latent modelling approach, and compare this to genetic risk for other psychiatric disorders, to gain a more comprehensive understanding of the developmental pathways at this age. PRSs for schizophrenia, major depressive disorder, neuroticism and bipolar disorder were generated for individuals in the Avon Longitudinal Study of Parents and Children (ALSPAC) birth cohort. Multivariate linear regression was used to examine the relationships of these PRSs with psychopathology factors modelled within (i) a correlated factors structure and (ii) a bifactor structure. The schizophrenia PRS was associated with an increase in factors describing psychotic experiences, negative dimension, depression and anxiety, but, when modelling a general psychopathology factor based on these measures, specific effects above this persisted only for the negative dimension. Similar factor relationships were observed for the neuroticism PRS, with a (weak) specific effect only for anxiety once modelling general psychopathology. Psychopathology during adolescence can be described by a general psychopathology construct that captures common variance as well as by specific constructs capturing remaining non-shared variance. Schizophrenia risk genetic variants identified through genome-wide association studies mainly index negative rather than positive symptom psychopathology during adolescence. This has potentially important implications both for research and risk prediction in high-risk samples.

Neighbourhood change and neighbour complaints: How gentrification and densification influence the prevalence of problems between neighbours

As people who live in closest proximity to us, the conduct of neighbours can have an impact upon our lives, even if they are relative strangers. While previous research has generally examined the positive effects of good neighbour interactions, neighbours can also be a source of nuisance, conflict and distress. In the advent of socio-structural processes of urban policy and change – such as gentrification and densification – the taken-for-granted conventions that once regulated neighbour interactions are being eroded, potentially leading to greater levels of neighbour problems and complaints. In this paper, we apply a latent modelling approach to identify subgroups of neighbourhoods based on their profiles of neighbour problems and to assess whether these subgroups are characterised by the degree of social change in the neighbourhood towards the dual processes of gentrification and densification. The findings show that high intensity problems are associated with both processes, but that class factors of gentrification are more influential than density in accounting for neighbour tensions.

S139. INVESTIGATING THE GENETIC ARCHITECTURE OF GENERAL AND SPECIFIC PSYCHOPATHOLOGY IN ADOLESCENCE USING SCHIZOPHRENIA POLYGENIC SCORES

BackgroundWhilst associations between polygenic risk scores (PRSs) for schizophrenia and various phenotypic outcomes have been reported, an understanding of developmental pathways can only be gained by modelling comorbidity across psychopathology, something no studies have done to date. We examine how genetic risk for schizophrenia relates to a broad range of adolescent psychopathology using a latent modelling approach, and compare this to genetic risk for other psychiatric disorders, to gain a more comprehensive understanding of development pathways at this age.MethodsPRSs for schizophrenia, major depressive disorder, neuroticism and bipolar disorder were generated for individuals in the Avon Longitudinal Study of Parents and Children (ALSPAC) birth cohort. Multivariate linear regression was used to examine relationships of these PRSs with psychopathology factors modelled within i) a correlated factors structure, and ii) a bifactor structure.ResultsThe schizophrenia PRS was associated with an increase in factors describing psychotic experiences, negative dimension, depression, and anxiety, but once modelling a general psychopathology factor specific effects above this persisted only for the negative dimension. Similar factor relationships were observed for the neuroticism PRS, with a (weak) specific effect only for anxiety once modelling general psychopathology.DiscussionPsychopathology during adolescence can be described by a general psychopathology construct that captures common variance as well as by specific constructs capturing remaining non-shared variance. Schizophrenia risk genetic variants identified through genome-wide association studies mainly index negative rather than positive symptom psychopathology during adolescence. This has potentially important implications both for research and risk prediction in high-risk samples.

The Relationship Between Sleep and Cognition in Children Referred for Neuropsychological Evaluation: A Latent Modeling Approach

Children with conditions affecting cognitive processes experience high levels of sleep disturbance, which may further compound the cognitive ramifications of their disorders. Despite this, existing studies in this area have been primarily confined to only particular diagnostic groups and/or a limited scope of sleep and cognitive parameters. The current study characterized the nature of sleep problems and examined the relationship between a wide range of sleep-related problems and cognitive functioning in a large (N = 103) diagnostically heterogeneous sample of youth (aged 6–16) referred for neuropsychological assessment. Structural equation modeling was used to examine the relationship between sleep-related problems (i.e., daytime sleepiness, sleep onset latency, sleep fragmentation, sleep time variability, sleep debt) and cognitive performance (i.e., executive functioning, sustained attention, memory, processing speed). Sleep fragmentation emerged as the most prominent sleep-related problem in the present sample. Structural equation modeling demonstrated a negative association between sleep-related problems and cognition that did not reach statistical significance (β = −0.084, p = 0.629). The current statistical approach may be used as a conceptual framework for future work examining these multi-dimensional constructs in a parsimonious fashion.

A Spatiotemporal Multispecies Model of a Semicontinuous Response

SummaryAs accessible and potentially vulnerable species high up in the food chain, birds are often used as indicator species to highlight changes in ecosystems. This study focuses on multiple spatially dependent relationships between a raptor (sparrowhawk), a potential prey species (house sparrow) and a sympatric species (collared doves) in space and time. We construct a complex spatiotemporal latent Gaussian model to incorporate both predator–prey and sympatric relationships, which is novel in two ways. First, different types of species interactions are represented by a shared spatiotemporal random effect, which extends existing approaches to multivariate spatial modelling through the use of a joint latent modelling approach. Second, we use a delta–gamma model to capture the semicontinuous nature of the data to model the binary and continuous sections of the response jointly. The results indicate that sparrowhawks have a localized effect on the presence of house sparrows, which could indicate that house sparrows avoid sites where sparrowhawks are present.

Transdiagnostic psychiatric disorder risk associated with early age of menarche: A latent modeling approach

BackgroundEarlier age of pubertal maturation in females is associated with increased risk for mental health problems in adolescence, compared with on-time or later maturation. However, most investigations of pubertal timing and mental health consider risk for individual disorders and fail to account for comorbidity. A latent-modeling approach using a large, nationally representative sample could better explain the transdiagnostic nature of the consequences of early-onset puberty. MethodsData on age of menarche and mental disorders were drawn from a population-representative sample of adolescents (n=4925), ages 13–17. Confirmatory factor analysis was used to fit four latent disorder categories: distress, eating, and externalizing, and fear disorders. Timing of menarche included those with earlier (age≤10, age 11) and later age of onset (age 13, 14+), relative to those with average timing of menarche (age 12). Associations between timing of menarche and latent disorders were estimated in a structural equation model (SEM), adjusted for age, income, race, parent marital status, BMI, and childhood adversity. ResultsThe measurement model evidenced acceptable fit (CFI=0.91; RMSEA=0.02). Onset of menarche before age 11 was significantly associated with distress disorders (coefficient=0.096; p<0.0001), fear disorders (coefficient=0.09; p<0.0001), and externalizing disorders (coefficient=0.039; p=0.049) as compared to on-time or late menarche. No residual associations of early menarche with individual disorders over and above the latent disorders were observed. ConclusionThe latent modeling approach illuminated meaningful transdiagnostic psychiatric associations with early timing of menarche. Biological processes initiated at puberty can influence cognitive and affective processes as well as social relationships for adolescents. Under developmentally normative conditions, these changes may be adaptive. However, for those out of sync with their peers, researchers and clinicians should recognize the potential for these processes to influence liability to a broad array of psychopathological consequences in adolescence.

A latent modeling approach to genotype–phenotype relationships: maternal problem behavior clusters, prenatal smoking, and MAOA genotype

This study illustrates the application of a latent modeling approach to genotype-phenotype relationships and gene × environment interactions, using a novel, multidimensional model of adult female problem behavior, including maternal prenatal smoking. The gene of interest is the monoamine oxidase A (MAOA) gene which has been well studied in relation to antisocial behavior. Participants were adult women (N = 192) who were sampled from a prospective pregnancy cohort of non-Hispanic, white individuals recruited from a neighborhood health clinic. Structural equation modeling was used to model a female problem behavior phenotype, which included conduct problems, substance use, impulsive-sensation seeking, interpersonal aggression, and prenatal smoking. All of the female problem behavior dimensions clustered together strongly, with the exception of prenatal smoking. A main effect of MAOA genotype and a MAOA × physical maltreatment interaction were detected with the Conduct Problems factor. Our phenotypic model showed that prenatal smoking is not simply a marker of other maternal problem behaviors. The risk variant in the MAOA main effect and interaction analyses was the high activity MAOA genotype, which is discrepant from consensus findings in male samples. This result contributes to an emerging literature on sex-specific interaction effects for MAOA.