RESUMO
Moderators are variables that change the relations among other variables. Moderators are variables that are substantive just as the variables whose relations are moderated. In the present article, we propose using individuals as moderators. Specifically, we propose using Configural Frequency Analysis, that is, investigating moderators from a person-oriented perspective. The question asked is whether variable relations vary across individuals. Base models are specified for Configural Frequency Analysis that allow one to identify individuals that differ in variable relations. In a data example, it is shown that not a single individual in a sample of alcoholics shows the pattern of association between subjective stress and beer consumption that was found for the entire sample. Extensions of the configural moderator model are discussed.
RESUMO
Serial dependence often prevents researchers from obtaining unbiased parameter estimates. In this article, we propose taking serial dependence into account, and exploiting the information that comes with serial dependence. This can be done in the form of shifted variables that are included in addition to the original variables, when models are specified. This way, models become more complex but relations can be considered that, otherwise, cannot be analyzed. Two fields of application are discussed. The first is log-linear modeling. This method is variable-oriented, but it has found applications in person-oriented research. The gain from including shifted variables in log-linear models is that new, specific variable relations can be analyzed. The second field is that of Configural Frequency Analysis. This method is person-oriented, and it allows researchers to detect local relations that, without consideration of shifted variables, cannot be detected. Application examples are given in the context of single-case analysis.
Assuntos
Modelos Estatísticos , Humanos , Modelos Lineares , Interpretação Estatística de DadosRESUMO
In observational data, understanding the causal link when estimating the causal effect of an independent variable (x) on a dependent variable (y) often requires researchers to identify the role of a third variable in the x â y relationship. Mediation, confounding, and colliding are three key third-variable effects that yield different theoretical and methodological implications for drawing causal conclusions. Commonly used covariance-based statistical methods, such as linear regression and structural equation modeling, cannot distinguish these effects in practice, however. In this study, we introduce a statistical approach for distinguishing mediators, confounders, colliders, and potential M-bias structures that uses higher-order moment information from the data. We propose a two-step procedure that uses the Hilbert-Schmidt independence criterion within the direction dependence analysis framework. Results from Monte Carlo simulations show that our proposed approach accurately recovers the true data-generating process of the third variable. We provide an empirical example to demonstrate the application of our proposed approach in psychological research. Finally, we discuss implications and future directions of our work. (PsycInfo Database Record (c) 2023 APA, all rights reserved).
RESUMO
Methods of causal discovery and direction of dependence to evaluate causal properties of variable relations have experienced rapid development. The majority of causal discovery methods, however, relies on the assumption of causal effect homogeneity, that is, the identified causal structure is expected to hold for the entire population. Because causal mechanisms can vary across subpopulations, we propose combining methods of model-based recursive partitioning and non-Gaussian causal discovery to identify such subpopulations. The resulting algorithm can discover subpopulations with potentially varying magnitude and causal direction of effects under mild parameter inequality assumptions. Feasibility conditions are described and results from synthetic data experiments are presented suggesting that large effects and large sample sizes are beneficial for detecting causally competing subgroups with acceptable statistical performance. In a real-world data example, the extraction of meaningful subgroups that differ in the causal mechanism underlying the development of numerical cognition is illustrated. Potential extensions and recommendations for best practice applications are discussed.
RESUMO
Understanding causal mechanisms is a central goal in the behavioral, developmental, and social sciences. When estimating and probing causal effects using observational data, covariate adjustment is a crucial element to remove dependencies between focal predictors and the error term. Covariate selection, however, constitutes a challenging task because availability alone is not an adequate criterion to decide whether a covariate should be included in the statistical model. The present study introduces a non-Gaussian method for covariate selection and provides a forward selection algorithm for linear models (i.e., non-Gaussian forward selection; nGFS) to select appropriate covariates from a set of potential control variables to avoid inconsistent and biased estimators of the causal effect of interest. Further, we demonstrate that the forward selection algorithm has properties compatible with principles of direction of dependence, i.e., probing whether the causal target model is correctly specified with respect to the causal direction of effects. Results of a Monte Carlo simulation study suggest that the selection algorithm performs well, in particular when sample sizes are large (i.e., n ≥ 250) and data strongly deviate from Gaussianity (e.g., distributions with skewness beyond 1.5). An empirical example is given for illustrative purposes.
RESUMO
In item response theory (IRT) modeling, the magnitude of the lower and upper asymptote parameters determines the degree to which the inflection point shifts above or below P = 0.50. The current study examines the one-parameter negative log-log model (NLLM), which is characterized by a downward shift in the inflection point, among other distinctive psychometric properties. After detailing the statistical foundations of the NLLM, we present a series of simulation studies to establish item and person parameter estimation accuracy and to demonstrate that this parsimonious model addresses the "slipping" effect (i.e., unexpectedly incorrect answers) via an inflection point < 0.50 rather than through computationally difficult estimation of the upper asymptote. We then provide further support for these simulation results through empirical data analysis. Finally, we discuss how the NLLM contributes to recent methodological literature on the utility of asymmetric IRT models.
RESUMO
In this article, we demonstrate that latent variable analysis can be of great use in person-oriented research. Starting with exploratory factor analysis of metric variables, we present an example of the problems that come with generalization of aggregate-level results to subpopulations. Oftentimes, results that are valid for populations do not represent subpopulations at all. This applies to confirmatory factor analysis as well. When variables are categorical, latent class analysis can be used to create latent variables that explain the covariation of observed variables. In an example, we demonstrate that latent class analysis can be applied to data from individuals, when the number of observation points is sufficiently large. In each case of latent variables analysis, the latent variables can be considered moderators of the structure of covariation among observed variables.
RESUMO
A variety of health and social problems are routinely measured in the form of categorical outcome data (such as presence/absence of a problem behavior or stages of disease progression). Therefore, proper quantitative analysis of categorical data lies at the heart of the empirical work conducted in prevention science. Categorical data analysis constitutes a broad dynamic field of methods research and data analysts in prevention science can benefit from incorporating recent advances and developments in the statistical evaluation of categorical outcomes in their methodological repertoire. The present Special Issue, Advanced Categorical Data Analysis in Prevention Science, highlights recent methods developments and illustrates their application in the context of prevention science. Contributions of the Special Issue cover a wide variety of areas ranging from statistical models for binary as well as multi-categorical data, advances in the statistical evaluation of moderation and mediation effects for categorical data, developments in model evaluation and measurement, as well as methods that integrate variable- and person-oriented categorical data analysis. The articles of this Special issue make methodological advances in these areas accessible to the audience of prevention scientists to maintain rigorous statistical practice and decision making. The current paper provides background and rationale for this Special Issue, an overview of the articles, and a brief discussion of some potential future directions for prevention research involving categorical data analysis.
Assuntos
Modelos Estatísticos , Comportamento Problema , Humanos , Problemas Sociais , Pesquisa sobre Serviços de Saúde , Análise de DadosRESUMO
Homogeneity of variance (HOV) is a well-known but often untested assumption in the context of multilevel models (MLMs). However, depending on how large the violation is, how different group sizes are, and the variance pairing, standard errors can be over or underestimated even when using MLMs, resulting in questionable inferential tests. We evaluate several tests (e.g., the H statistic, Breusch Pagan, Levene's test) that can be used with MLMs to assess violations of HOV. Although the traditional robust standard errors used with MLMs require at least 50 clusters to be effective, we assess a robust standard error adjustment (i.e., the CR2 estimator) that can be used even with a few clusters. Findings are assessed using a Monte Carlo simulation and are further illustrated using an applied example. We show that explicitly modeling the heterogenous variance structures or using the CR2 estimator are both effective at ameliorating the issues associated with the fixed effects of the regression model related to violations of HOV resulting from between-group differences.
Assuntos
Modelos Estatísticos , Simulação por Computador , Análise Multinível , Método de Monte CarloRESUMO
Traditional item response theory (IRT) models assume a symmetric error distribution and rely on symmetric (logit or probit) link functions to model the response probabilities. As an alternative, we investigated the one-parameter complementary log-log model (CLLM), which is founded on an asymmetric error distribution and results in an asymmetric item response function with important psychometric properties. In a series of simulation studies, we demonstrate that the CLLM (a) is estimable in small sample sizes, (b) facilitates item-weighted scoring, and (c) accounts for the effect of guessing, despite the presence of a single parameter. We then provide further evidence for these claims by applying the CLLM to empirical data. Finally, we discuss how this work contributes to the growing psychometric literature on model complexity.
Assuntos
Psicometria , Humanos , Psicometria/métodos , Simulação por Computador , Probabilidade , Tamanho da AmostraRESUMO
Comparative measures such as paired comparisons and rankings are frequently used to evaluate health states and quality of life. The present article introduces log-linear Bradley-Terry (LLBT) models to evaluate intervention effectiveness when outcomes are measured as paired comparisons or rankings and presents a combination of the LLBT model and model-based recursive partitioning (MOB) to detect treatment effect heterogeneity. The MOB LLBT approach enables researchers to identify subgroups that differ in the preference order and in the effect an intervention has on choice behavior. Applicability of MOB LLBT models is demonstrated using an artificial data example with known data-generating mechanism and a real-world data example focusing on drug-harm perception among music festival visitors. In the artificial data example, the MOB LLBT model is able to adequately recover the "true" (population) model. In the real-world data example, the standard LLBT model confirms the existence of a situational willingness among festival visitors to trivialize drug harm when peer consumption behavior is made cognitively accessible. In addition, MOB LLBT results suggest that this trivialization effect is highly context-dependent and most pronounced for participants with low-to-moderate alcohol intoxication who also proactively contacted a substance counselor at the festival venue. Both data examples suggest that MOB LLBT models allow for more nuanced statements about the effectiveness of interventions. We provide R code examples to implement MOB LLBT models for paired comparisons, rankings, and rating (Likert-type) data.
Assuntos
Julgamento , Música , Humanos , Qualidade de VidaRESUMO
In standard statistical data analysis, the effects of intervention or prevention efforts are evaluated in terms of variable relations. Results from application of regression-type methods suggest whether, overall, intervention is successful. In this article, we propose using configural frequency analysis (CFA) either in tandem with regression-type methods or by itself. CFA allows one to adopt a person-oriented perspective in which individuals are targeted that can be characterized by particular profiles. The questions asked in CFA concern these individuals instead of variables. In prevention research, one can ask whether, for particular profiles, the preventive measures are successful. In three real-world data examples, CFA is applied and compared to standard log-linear modeling. Examples consider non-randomized (observational) and randomized intervention settings. The results of these analyses suggest that person-oriented CFA and standard variable-oriented methods of analysis respond to different questions. We show that integrating person- and variable-oriented perspectives can help researchers obtain a fuller picture of intervention effectiveness. Extensions of the CFA approach are discussed.
Assuntos
Análise de Regressão , Humanos , Interpretação Estatística de DadosRESUMO
BACKGROUND: The demographic determinants of hesitancy in Coronavirus Disease-2019 (COVID-19) vaccination include rurality, particularly in low- and middle-income countries. In the second year of the pandemic, in South Tyrol, Italy, 15.6 percent of a representative adult sample reported hesitancy. Individual factors responsible for greater vaccination hesitancy in rural areas of central Europe are poorly understood. METHODS: A cross-sectional survey on a probability-based sample of South Tyrol residents in March 2021 was analyzed. The questionnaire collected information on sociodemographic characteristics, comorbidities, COVID-19-related experiences, conspiracy thinking, and the likelihood of accepting the national vaccination plan. A logistic regression analysis was performed. RESULTS: Among 1426 survey participants, 17.6% of the rural sample (n = 145/824) reported hesitancy with COVID-19 vaccination versus 12.8% (n = 77/602) in urban residents (p = 0.013). Rural residents were less likely to have post-secondary education, lived more frequently in households with children under six years of age, and their economic situation was worse than before the pandemic. Chronic diseases and deaths due to COVID-19 among close relatives were less frequently reported, and trust in pandemic management by national public health institutions was lower, as was trust in local authorities, civil protection, and local health services. Logistic regression models confirmed the most well-known predictors of hesitancy in both urban and rural populations; overall, residency was not an independent predictor. CONCLUSION: Several predictors of COVID-19 vaccine hesitancy were more prevalent in rural areas than in urban areas, which may explain the lower vaccine uptake in rural areas. Rurality is not a determinant of vaccine hesitancy in the economically well-developed North of Italy.
RESUMO
BACKGROUND: Belief in complementary and alternative medicine practices is related to reduced preparedness for vaccination. This study aimed to assess home remedy awareness and use in South Tyrol, where vaccination rates in the coronavirus pandemic were lowest in Italy and differed between German- and Italian-speaking inhabitants. METHODS: A population-based survey was conducted in 2014 and analyzed using descriptive statistics, multiple logistic regression, and latent class analysis. RESULTS: Of the representative sample of 504 survey respondents, 357 (70.8%) participants (43.0% male; primary language German, 76.5%) reported to use home remedies. Most commonly reported home remedies were teas (48.2%), plants (21.0%), and compresses (19.5%). Participants from rural regions were less likely (odds ratio 0.35, 95% confidence interval 0.19-0.67), while female (2.62, 1.69-4.10) and German-speaking participants (5.52, 2.91-9.88) were more likely to use home remedies. Latent classes of home remedies were "alcoholic home remedies" (21.4%) and "non-alcohol-containing home remedies" (78.6%). Compared to the "non-alcohol-containing home remedies" class, members of the "alcoholic home remedies" class were more likely to live in an urban region, to be male and German speakers. CONCLUSION: In addition to residence and sex, language group membership associates with awareness and use of home remedies. Home remedies likely contribute to socio-cultural differences between the language groups in the Italian Alps. If the observed associations explain the lower vaccination rates in South Tyrol among German speakers requires further study.
Assuntos
Terapias Complementares , Medicina Tradicional , Humanos , Estudos Transversais , Inquéritos e Questionários , TelefoneRESUMO
BACKGROUND: German is a minority language in Italy and is spoken by the majority of the inhabitants of the Autonomous Province of Bolzano, South Tyrol. Linguistic group membership in South Tyrol is an established determinant of health information-seeking behavior. Because the COVID-19 incidence and vaccination coverage in the second year of the pandemic in Italy was the worst in South Tyrol, we investigated whether linguistic group membership is related to COVID-19 vaccine hesitancy. METHODS: A cross-sectional survey was conducted on a probability-based sample of 1425 citizens from South Tyrol in March 2021. The questionnaire collected information on socio-demographics, including linguistic group membership, comorbidities, COVID-19-related experiences, conspiracy thinking, well-being, altruism, and likelihood of accepting the national vaccination plan. Multiple logistic regression analyses were performed to identify the significant predictors of vaccine hesitancy. RESULTS: Overall, 15.6 percent of the sample reported vaccine hesitancy, which was significantly higher among German speakers than among other linguistic groups. Increased hesitancy was mostly observed in young age, the absence of chronic disease, rural residence, a worsened economic situation, mistrust in institutions, and conspiracy thinking. In the multiple logistic regression analyses, linguistic group membership was not an independent predictor of vaccine hesitancy. CONCLUSION: Although German is a minority language in Italy and COVID-19 vaccine hesitancy was higher in the German native language group than in the Italian, linguistic group membership was not an independent predictor of hesitancy in the autonomous province. Known predictors of vaccine hesitancy are distributed unevenly across language groups. Whether language group-specific intervention strategies to promote vaccine hesitancy are useful requires further study.
RESUMO
The usefulness of mean aggregates in the analysis of intervention effectiveness is a matter of considerable debate in the psychological, educational, and social sciences. In addition to studying "average treatment effects," the evaluation of "distributional treatment effects," (i.e., effects that go beyond means), has been suggested to obtain a broader picture of how an intervention affects the study outcome. We continue this discussion by considering distributional causal effects. We present formal definitions of causal effects that go beyond means and utilize a distributional regression framework known as generalized additive models for location, scale, and shape (GAMLSS). GAMLSS allows one to characterize an intervention effect in its totality through simultaneously modeling means, variances, skewnesses, kurtoses, as well as ceiling and floor effects of outcome distributions. Based on data from a large-scale randomized controlled trial, we use GAMLSS to evaluate the impact of a teacher classroom management program on student academic performance. Results suggest the teacher classroom management training increased mean academic competence as well as the chance to obtain the maximum score on the academic competence scale. These effects would have been completely overlooked in a traditional evaluation of mean aggregates. (PsycInfo Database Record (c) 2022 APA, all rights reserved).
RESUMO
Differences in the ability of students to judge images can be assessed by analyzing the individual preference order (ranking) of images. To gain insights into potential heterogeneity in judgement of visual abstraction among students, we combine Bradley-Terry preference modeling and model-based recursive partitioning. In an experiment a sample of 1,020 high-school students ranked five sets of images, three of which with respect to their level of visual abstraction. Additionally, 24 art experts and 25 novices were given the same task, while their eye movements were recorded. Results show that time spent on the task, the students' age, and self-reported interest in visual puzzles had significant influence on rankings. Fixation time of experts and novices revealed that both groups paid more attention to ambiguous images. The presented approach makes the underlying latent scale of visual judgments quantifiable.
RESUMO
Background/Objective: The aim of the present study was to compare competing psychometric models and analyze measurement invariance of the Hospital Anxiety and Depression Scale (HADS) in cancer outpatients. Method: The sample included 3,260 cancer outpatients. Latent structure of the HADS was analyzed using confirmatory factor analysis (CFA) with robust maximum likelihood estimation (MLR). Measurement invariance was tested for age, time of response, gender, and cancer type by comparing nested multigroup CFA models with parameter restrictions. Results: Except for the one-factor solutions, all models showed acceptable model fit and measurement invariance. The model with the best fit was the originally proposed two-factor model with exclusion of two items. The one-factor solutions showed inacceptable model fit and were not invariant for age and gender. Conclusions: The HADS has a robust two-factor structure in cancer outpatients. We recommend excluding item 7 and 10 when screening for anxiety and depression.
RESUMO
Unless very large samples are available, the number of variables and variable categories that can be simultaneously used in categorical data analysis is small when models are estimated. In this article, an approach is proposed that can help remedy this problem. Specifically, it is proposed to perform, in a first step, principal component analysis or factor analysis. These methods help reduce the dimensionality of the data space without loss of important information. In a second step, sectors are created in the component or factor space. These sectors can, in a third step, be subjected to Configural Frequency analysis (CFA). CFA identifies those sectors that contradict a priori-specified hypotheses. It is also proposed to take into account the ordinal nature of the sectors. In addition, distributional assumptions can be considered. This is illustrated in data examples. Possible extensions of the proposed approach are discussed.
RESUMO
BACKGROUND: Evidence suggests an increasing demand for culturally and linguistically responsive disease prevention programs and health interventions. It is important to understand how individuals seek health information to address the potential needs of the health care system. METHODS: Latent classes of health information-seeking behaviors in a linguistically mixed region of Italy were explored through a population-based telephone survey of ten health information sources. Data were collected in August and September 2014 from 504 adults in South Tyrol, Italy (primary language German, 68%; Italian, 28%), and analyzed using latent class analysis and latent class multinomial logistic regression models. RESULTS: Three classes of health information-seeking behaviors emerged: "multidimensional" (23.3%), "interpersonal" (38.6%) and "technical/online" (38.1%). Compared to the "technical/online" class, "interpersonal" class members were older, had lower education than high school, and were less likely to be of Italian ethnicity. "Multidimensional" class members were more likely to be female, older, and of German ethnicity than those in the "technical/online" class. CONCLUSIONS: Linguistic ethnicity explains membership in classes on health information-seeking behavior. Policymakers and healthcare providers should consider the health information-seeking behaviors of population subgroups to promote health and medical care in linguistic minority groups.