Monte Carlo local likelihood approximation.

We propose the Monte Carlo local likelihood (MCLL) method for approximating maximum likelihood estimation (MLE). MCLL initially treats model parameters as random variables, sampling them from the posterior distribution as in a Bayesian model. The likelihood function is then approximated up to a constant by fitting a density to the posterior samples and dividing the approximate posterior density by the prior. In the MCLL algorithm, the posterior density is estimated using local likelihood density estimation, in which the log-density is locally approximated by a polynomial function. We also develop a new method that allows users to efficiently compute standard errors and the Bayes factor. Two empirical and three simulation studies are provided to demonstrate the performance of the MCLL method.

Trifactor Models for Multiple-Ratings Data.

In this study we extend and assess the trifactor model for multiple-ratings data in which two different raters give independent scores for the same responses (e.g., in the GRE essay or to subset of PISA constructed-responses). The trifactor model was extended to incorporate a cross-classified data structure (e.g., items and raters) instead of a strictly hierarchical structure. we present a set of simulations to reflect the incompleteness and imbalance in real-world assessments. The effects of the rate of missingness in the data and of ignoring differences among raters are investigated using two sets of simulations. The use of the trifactor model is also illustrated with empirical data analysis using a well-known international large-scale assessment.

Pairwise residuals and diagnostic tests for misspecified dependence structures in models for binary longitudinal data.

Maximum likelihood estimation of models for binary longitudinal data is typically inconsistent if the dependence structure is misspecified. Unfortunately, diagnostics specifically designed for detecting such misspecifications are scant. We develop residuals and diagnostic tests based on comparing observed and expected frequencies of response patterns over time in the presence of arbitrary time-varying and time-invariant covariates. To overcome the sparseness problem, we use lower-order marginal tables, such as two-way tables for pairs of time-points, aggregated over covariate patterns. Our proposed pairwise concordance residuals are valuable for exploratory diagnostics and for constructing both generic tests for misspecified dependence structure as well as targeted adjacent pair concordance tests for excess serial dependence. The proposed methods are straightforward to implement and work well for general situations, regardless of the number of time-points and the number and types of covariates.

The spatial relationship between neurons and astrocytes in HIV-associated dementia.

Specific neuronal spatial distributional patterns have previously been correlated with increasing severity of HIV-associated dementia (HAD). As astrocytes are also a putative site of neurotoxicity, we investigated the spatial relationships of astrocytes with pyramidal and interneurons in the superior frontal gyrus from 29 patients who died from acquired immunodeficiency syndrome. Frontal cortical brain tissue was taken from diseased HIV patients who had been assessed for the presence and severity of HAD using the Memorial Sloan-Kettering Scale. No correlation was found between neuronal density and severity of dementia. However, the pattern of astrocytes became more clustered as dementia progressed. Bivariate spatial pattern analysis of neuronal populations with astrocytes revealed that, with increasing dementia severity, astrocytes and large pyramidal neurons increasingly "repelled" each other, while astrocytes and interneurons evidenced increasing "attraction." This implies that astrocytes may be more likely to be situated in the vicinity of surviving interneurons but less likely to be situated near surviving large pyramidal neurons in the setting of progressing HAD.

Avoiding zero between-study variance estimates in random-effects meta-analysis.

Fixed-effects meta-analysis has been criticized because the assumption of homogeneity is often unrealistic and can result in underestimation of parameter uncertainty. Random-effects meta-analysis and meta-regression are therefore typically used to accommodate explained and unexplained between-study variability. However, it is not unusual to obtain a boundary estimate of zero for the (residual) between-study standard deviation, resulting in fixed-effects estimates of the other parameters and their standard errors. To avoid such boundary estimates, we suggest using Bayes modal (BM) estimation with a gamma prior on the between-study standard deviation. When no prior information is available regarding the magnitude of the between-study standard deviation, a weakly informative default prior can be used (with shape parameter 2 and rate parameter close to 0) that produces positive estimates but does not overrule the data, leading to only a small decrease in the log likelihood from its maximum. We review the most commonly used estimation methods for meta-analysis and meta-regression including classical and Bayesian methods and apply these methods, as well as our BM estimator, to real datasets. We then perform simulations to compare BM estimation with the other methods and find that BM estimation performs well by (i) avoiding boundary estimates; (ii) having smaller root mean squared error for the between-study standard deviation; and (iii) better coverage for the overall effects than the other methods when the true model has at least a small or moderate amount of unexplained heterogeneity.

Ignoring Non-ignorable Missingness.

The classical missing at random (MAR) assumption, as defined by Rubin (Biometrika 63:581-592, 1976), is often not required for valid inference ignoring the missingness process. Neither are other assumptions sometimes believed to be necessary that result from misunderstandings of MAR. We discuss three strategies that allow us to use standard estimators (i.e., ignore missingness) in cases where missingness is usually considered to be non-ignorable: (1) conditioning on variables, (2) discarding more data, and (3) being protective of parameters.

Role of marriage, motherhood, son preference on adolescent girls' and young women's empowerment: Evidence from a panel study in India.

BACKGROUND: Marriage is a key determinant of health and well-being of adolescent girls and young women (AGYW) in India. It is a key life event in which girls move to their marital households, often co-residing with their in-laws and begin childbearing. The change in the normative environment in conjunction with cultural norms surrounding son preference influences women's overall life course. However, there is scant research about the association between these life transitions and changes in empowerment among AGYW in India. METHODS: Using two waves of data from prospective cohort panel dataset that followed unmarried (6,065 observations in each wave) and married AGYW (3,941 observations from each wave) over a three-year period from Uttar Pradesh and Bihar, we examined how marriage, childbearing, and having a son is associated with changes in AGYW's empowerment, especially considering whether AGYW marry into patrilocal households (household with in-laws) as an effect modifier. Empowerment indicators included freedom of movement or mobility, decision-making power, access to economic using Kabeer's framework as our theoretical approach. RESULTS: Marriage was associated with lower freedom of movement with a pronounced effect on those who co-resided with their in-laws. Marriage was associated with greater decision-making power for AGYW who did not co-reside with the in-laws. Motherhood was positively correlated with greater freedom of movement, marginally higher intrahousehold decision-making power, and better access to economic resources. No statistically significant evidence that having at least one son compared to having daughters only (or no daughters) conferred additional changes in girls' freedom of movement, intrahousehold decision-making power, and access to economic resources. CONCLUSION: Findings highlight the importance of understanding the vulnerabilities of being newly married in adolescence and emphasize the need for having interventions that target newly married AGYW along with mothers-in-law to empower them.

Diagnosing and Handling Common Violations of Missing at Random.

Ignorable likelihood (IL) approaches are often used to handle missing data when estimating a multivariate model, such as a structural equation model. In this case, the likelihood is based on all available data, and no model is specified for the missing data mechanism. Inference proceeds via maximum likelihood or Bayesian methods, including multiple imputation without auxiliary variables. Such IL approaches are valid under a missing at random (MAR) assumption. Rabe-Hesketh and Skrondal (Ignoring non-ignorable missingness. Presidential Address at the International Meeting of the Psychometric Society, Beijing, China, 2015; Psychometrika, 2023) consider a violation of MAR where a variable A can affect missingness of another variable B also when A is not observed. They show that this case can be handled by discarding more data before proceeding with IL approaches. This data-deletion approach is similar to the sequential estimation of Mohan et al. (in: Advances in neural information processing systems, 2013) based on their ordered factorization theorem but is preferable for parametric models. Which kind of data-deletion or ordered factorization to employ depends on the nature of the MAR violation. In this article, we therefore propose two diagnostic tests, a likelihood-ratio test for a heteroscedastic regression model and a kernel conditional independence test. We also develop a test-based estimator that first uses diagnostic tests to determine which MAR violation appears to be present and then proceeds with the corresponding data-deletion estimator. Simulations show that the test-based estimator outperforms IL when the missing data problem is severe and performs similarly otherwise.

The Role of Gender Norms on Intimate Partner Violence Among Newly Married Adolescent Girls and Young Women in India: A Longitudinal Multilevel Analysis.

Gender norms have been posited to impact intimate partner violence (IPV), but there is scant evidence of the longitudinal association between community-level gender norms and IPV. Using longitudinal data on 3,965 married girls surveyed in India, we fitted mixed-effects ordinal and binary logistic regression models for physical IPV intensity and occurrence of sexual IPV. We found a 26% increase in the odds that women experience frequent physical IPV per one unit increase in greater community-level equitable gender norms. We did not find an association between community-level equitable gender norms and sexual IPV. Findings suggest that the relationship between gender norms and physical and sexual IPV differs, indicating the need for tailored interventions for different types of IPV.

The Role of Conditional Likelihoods in Latent Variable Modeling.

In psychometrics, the canonical use of conditional likelihoods is for the Rasch model in measurement. Whilst not disputing the utility of conditional likelihoods in measurement, we examine a broader class of problems in psychometrics that can be addressed via conditional likelihoods. Specifically, we consider cluster-level endogeneity where the standard assumption that observed explanatory variables are independent from latent variables is violated. Here, "cluster" refers to the entity characterized by latent variables or random effects, such as individuals in measurement models or schools in multilevel models and "unit" refers to the elementary entity such as an item in measurement. Cluster-level endogeneity problems can arise in a number of settings, including unobserved confounding of causal effects, measurement error, retrospective sampling, informative cluster sizes, missing data, and heteroskedasticity. Severely inconsistent estimation can result if these challenges are ignored.

A randomized controlled trial of the Transdiagnostic Intervention for Sleep and Circadian Dysfunction (TranS-C) to improve serious mental illness outcomes in a community setting.

OBJECTIVE: To determine if the Transdiagnostic Intervention for Sleep and Circadian Dysfunction (TranS-C) improves functional impairment, psychiatric symptoms, and sleep and circadian functioning. METHOD: Adults diagnosed with serious mental illness (SMI) and sleep and circadian dysfunction (N = 121) were randomly allocated to TranS-C plus usual care (TranS-C + UC; n = 61; 8 individual weekly sessions) or 6 months of Usual Care followed by Delayed Treatment with TranS-C (UC-DT; n = 60). Schizophrenia (45%) and anxiety disorders (47%) were common. Blind assessments were conducted pre-treatment, post-treatment, and 6 months later (6FU). The latter two were the post-randomization points of interest. The location was Alameda County Behavioral Health Care Services (ACBHCS), a Community Mental Health Center (CMHC) in California. RESULTS: For the primary outcomes, relative to UC-DT, TranS-C + UC was associated with reduction in functional impairment (b = -3.18, p = 0.025, d = -0.58), general psychiatric symptoms (b = -5.88, p = 0.001, d = -0.64), sleep disturbance (b = -5.55, p < .0001, d = -0.96), and sleep-related impairment (b = -9.14, p < .0001, d = -0.81) from pre-treatment to post-treatment. These effects were maintained to 6-month follow-up (6FU; d = -0.42 to -0.82), except functional impairment (d = -0.37). For the secondary outcomes, relative to UC-DT, TranS-C + UC was associated with improvement in sleep efficiency and on the Sleep Health Composite score from pre-treatment to 6FU. TranS-C + UC was also associated with reduced total wake time and wake time variability from pre-treatment to post-treatment, as well as reduced hallucinations and delusions, bedtime variability, and actigraphy measured waking activity count variability from pre-treatment to 6FU. CONCLUSIONS: A novel transdiagnostic treatment, delivered within a CMHC setting, improves selected measures of functioning, symptoms of comorbid disorders, and sleep and circadian outcomes. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

A randomized controlled two-stage trial in the treatment of bulimia nervosa, comparing CBT versus motivational enhancement in Phase 1 followed by group versus individual CBT in Phase 2.

OBJECTIVE: To conduct a randomized, controlled, two-stage trial in the treatment of bulimia nervosa, comparing cognitive-behavioral therapy (CBT) versus motivational enhancement in Phase 1 followed by group versus individual CBT in Phase 2. METHODS: A total of 225 patients with bulimia nervosa or eating disorder not otherwise specified (EDNOS) were recruited into a randomized controlled trial lasting 12 weeks with follow-ups at 1 year and 2.5 years. RESULTS: Patients improved significantly across all of the interventions with no differences in outcome or treatment adherence. Including motivational enhancement therapy rather than a CBT first phase of treatment did not affect outcome. CONCLUSIONS: Outcome differences between individual and group CBT were minor, suggesting that group treatment prefaced by a short individual intervention may be a cost-effective alternative to purely individual treatment.

A Dyadic IRT Model.

We propose a dyadic Item Response Theory (dIRT) model for measuring interactions of pairs of individuals when the responses to items represent the actions (or behaviors, perceptions, etc.) of each individual (actor) made within the context of a dyad formed with another individual (partner). Examples of its use include the assessment of collaborative problem solving or the evaluation of intra-team dynamics. The dIRT model generalizes both Item Response Theory models for measurement and the Social Relations Model for dyadic data. The responses of an actor when paired with a partner are modeled as a function of not only the actor's inclination to act and the partner's tendency to elicit that action, but also the unique relationship of the pair, represented by two directional, possibly correlated, interaction latent variables. Generalizations are discussed, such as accommodating triads or larger groups. Estimation is performed using Markov-chain Monte Carlo implemented in Stan, making it straightforward to extend the dIRT model in various ways. Specifically, we show how the basic dIRT model can be extended to accommodate latent regressions, multilevel settings with cluster-level random effects, as well as joint modeling of dyadic data and a distal outcome. A simulation study demonstrates that estimation performs well. We apply our proposed approach to speed-dating data and find new evidence of pairwise interactions between participants, describing a mutual attraction that is inadequately characterized by individual properties alone.

How far are associations between child, family and community factors and child psychopathology informant-specific and informant-general?

BACKGROUND: Assessments of child psychopathology commonly rely on multiple informants, e.g., parents, teachers and children. Informants often disagree about the presence or absence of symptoms, reflecting reporter bias, situation-specific behaviour, or random variation in measurement. However, few studies have systematically tested how far correlates of child psychopathology differ between informants. METHODS: Parents, teachers and children in the 1999 British Child and Adolescent Mental Health Survey (n = 4,525, ages 11-15 years) completed the Strengths and Difficulties Questionnaire. Multiple source regression models tested the extent to which child, family, school and neighbourhood characteristics were differentially associated with the three informants' reports. The 2004 British Child and Adolescent Mental Health Survey (n = 3,438, ages 11-15 years) was used for replication. RESULTS: Almost all significant correlates of child mental health were differentially related to parent, teacher and child ratings of adjustment. Parental distress, parent-rated family functioning, and child physical health problems were most strongly associated with parent ratings. Child ability and attainment, socio-economic factors, and school and neighbourhood disadvantage were more strongly associated with teacher and parent rated mental health than with children's own ratings. Gender differences in externalising problems were most pronounced for teacher ratings, and least so for child ratings; the opposite held true for emotional problems. Effect sizes for combined latent scores fell near the upper end of the range of effect sizes estimated for the three individual informants. Results showed good replication across the two samples. CONCLUSIONS: The study highlights that there is substantial variation across informants in the links between associated factors and child psychopathology.

The health and cognitive growth of Latino toddlers: at risk or immigrant paradox?

Epidemiologists have shown how birth outcomes are generally robust for immigrant Latina mothers, despite often situated in poor households, advanced by their strong prenatal and nutritional practices. But little is known about (1) how these protective factors may differ among Latino subgroups, (2) the extent to which birth outcomes, ongoing maternal practices, and family supports advance Latino toddlers' health and physical growth, and (3) whether the same processes advance toddlers' early cognitive growth. We drew on a national probability sample of 8,114 infants born in 2001, including 1,450 of diverse Latino origins. Data come from birth records, maternal interviews when the child was 9 and 24 months of age, and direct assessments of health status, physical growth, and cognitive proficiencies. Descriptive analyses compared Mexican-heritage and other Latino mothers and toddlers relative to middle-class whites. Multivariate regression techniques identified predictors of child health, weight, and BMI, as well as cognitive proficiencies at 24 months. Infants of Mexican-heritage or less acculturated Latina mothers displayed robust birth outcomes, compared with other ethnic groups. The low incidence of premature births and low birthweight among these mothers continued to advance their cognitive growth through 24 months of age. Yet Latino children overall displayed smaller gains in cognitive proficiencies between 9 and 24 months, compared with middle-class populations, attributable to Latinas' lower levels of maternal education, weaker preliteracy practices, and a higher ratio of children per resident adult. Health practitioners should recognize that many Latina mothers display healthy prenatal practices and give birth to robust infants. But these early protective factors do not necessarily advance early cognitive growth. Screening practices, early interventions, and federal policy should become more sensitive to these countervailing dynamics.

Bayesian Comparison of Latent Variable Models: Conditional Versus Marginal Likelihoods.

Typical Bayesian methods for models with latent variables (or random effects) involve directly sampling the latent variables along with the model parameters. In high-level software code for model definitions (using, e.g., BUGS, JAGS, Stan), the likelihood is therefore specified as conditional on the latent variables. This can lead researchers to perform model comparisons via conditional likelihoods, where the latent variables are considered model parameters. In other settings, however, typical model comparisons involve marginal likelihoods where the latent variables are integrated out. This distinction is often overlooked despite the fact that it can have a large impact on the comparisons of interest. In this paper, we clarify and illustrate these issues, focusing on the comparison of conditional and marginal Deviance Information Criteria (DICs) and Watanabe-Akaike Information Criteria (WAICs) in psychometric modeling. The conditional/marginal distinction corresponds to whether the model should be predictive for the clusters that are in the data or for new clusters (where "clusters" typically correspond to higher-level units like people or schools). Correspondingly, we show that marginal WAIC corresponds to leave-one-cluster out cross-validation, whereas conditional WAIC corresponds to leave-one-unit out. These results lead to recommendations on the general application of the criteria to models with latent variables.

Classical latent variable models for medical research.

Latent variable models are commonly used in medical statistics, although often not referred to under this name. In this paper we describe classical latent variable models such as factor analysis, item response theory, latent class models and structural equation models. Their usefulness in medical research is demonstrated using real data. Examples include measurement of forced expiratory flow, measurement of physical disability, diagnosis of myocardial infarction and modelling the determinants of clients' satisfaction with counsellors' interviews.

CFA Models with a General Factor and Multiple Sets of Secondary Factors.

We propose a class of confirmatory factor analysis models that include multiple sets of secondary or specific factors and a general factor. The general factor accounts for the common variance among manifest variables, whereas multiple sets of secondary factors account for the remaining source-specific dependency among subsets of manifest variables. A special case of the model is further proposed which constrains the specific factor loadings to be proportional to the general factor loadings. This proportional model substantially reduces the number of model parameters while preserving the essential structure of the general model. Furthermore, the proportional model allows for the interpretation of latent variables as the expected values of the observed manifest variables, decomposition of the variances, and the inclusion of interactions, similar to generalizability theory. We provide two applications to illustrate the utility of the proposed class of models.

Modifying the Impact of Eveningness Chronotype ("Night-Owls") in Youth: A Randomized Controlled Trial.

OBJECTIVE: To determine whether an intervention to reduce eveningness chronotype improves sleep, circadian, and health (emotional, cognitive, behavioral, social, physical) outcomes. METHOD: Youth aged 10 to 18 years with an evening chronotype and who were "at risk" in 1 of 5 health domains were randomized to: (a) Transdiagnostic Sleep and Circadian Intervention for Youth (TranS-C; n = 89) or (b) Psychoeducation (PE; n = 87) at a university-based clinic. Treatments were 6 individual, weekly 50-minute sessions during the school year. TranS-C addresses sleep and circadian problems experienced by youth by integrating evidence-based treatments derived from basic research. PE provides education on the interrelationship between sleep, stress, diet, and health. RESULTS: Relative to PE, TranS-C was not associated with greater pre-post change for total sleep time (TST) or bed time (BT) on weeknights but was associated with greater reduction in evening circadian preference (pre-post increase of 3.89 points, 95% CI = 2.94-4.85, for TranS-C, and 2.01 points, 95% CI = 1.05-2.97 for PE, p = 0.006), earlier endogenous circadian phase, less weeknight-weekend discrepancy in TST and wakeup time, less daytime sleepiness, and better self-reported sleep via youth and parent report. In terms of functioning in the five health domains, relative to PE, TranS-C was not associated with greater pre-post change on the primary outcome. However, there were significant interactions favoring TranS-C on the Parent-Reported Composite Risk Scores for cognitive health. CONCLUSION: For at-risk youth, the evidence supports the use of TranS-C over PE for improving sleep and circadian functioning, and improving health on selected outcomes. CLINICAL TRIAL REGISTRATION INFORMATION: Triple Vulnerability? Circadian Tendency, Sleep Deprivation and Adolescence. https://clinicaltrials.gov; NCT01828320.

A randomized, controlled trial with 6-month follow-up of repetitive transcranial magnetic stimulation and electroconvulsive therapy for severe depression.

OBJECTIVE: Repetitive transcranial magnetic stimulation (rTMS) has been reported to be as effective as electroconvulsive therapy (ECT) for major depression. The authors conducted a multicenter randomized, controlled trial to test the equivalence of rTMS with ECT. METHOD: Forty-six patients with major depression referred for ECT were randomly assigned to either a 15-day course of rTMS of the left dorsolateral prefrontal cortex (N=24) or a standard course of ECT (N=22). The primary outcome measures were the score on the 17-item Hamilton Depression Rating Scale (HAM-D) and the proportion of patients with remissions (Hamilton score,