Nested Structure Of Data Research Articles

Consequences of ignoring nested data structure on item parameters in Rasch/1P-IRT model

The present study investigated the consequences of ignoring a nested data structure on the Rasch/one parameter item response theory model. Although most large-scale educational assessment data do exhibit a nested data structure, current practice often ignores such data structure and applies the standard Rasch/IRT models to conduct measurement analyses. We hypothesized that this practice would produce negative consequences on the item parameter estimates. Using simulation, we investigated this hypothesis by comparing the results from an incorrectly specified two level model which ignored the nested data structure to those from a correctly specified three-level hierarchical generalized linear model. Use of the incorrect two-level model did, in fact, result in negative consequences in estimating the standard errors, although the point estimates were unbiased and identical to the ones from the three-level analysis. A real data set from the IEA Civic education study in 1999 was used to illustrate the simulation results.

Take It, or Leave It? Analyzing How Unsubsidized Federal Loans Affect Six-Year Degree Attainment Across Income Groups

This study examined the effects of unsubsidized federal Stafford loans on six-year degree attainment at 4-year colleges and universities in the U.S., and how these differentially impact students across income groups. For this, nationally representative data from the Beginning Postsecondary Students (BPS:04/09) and the Integrated Postsecondary Education Data System (IPEDS) was merged to analyze N=6,561 students attending n=651 four-year institutions. To account for possible selection bias in loan borrowing and the nested data structure, this study employs a propensity score matching, multilevel modeling approach. In addition to financial aid measures, the analytic model draws from the heterogeneous research approach and incorporates students’ college experiences and institutional context to estimate effects. Results reveal a comparatively large, negative association with the outcome measure for low-income students. For every $1,000 borrowed additionally in unsubsidized loans, low-income students are found to be 5.45% (p

Assessing treatment integrity in personalized CBT: the inventory of therapeutic interventions and skills

ABSTRACTThe third wave of cognitive behavioral therapy (CBT) has increased the heterogeneity of today’s CBT practice, while developments in patient-focused research are paving the road to the empirical personalization of CBT. This paper presents the development and psychometric properties of a therapy video rating instrument, which was designed to adequately assess the treatment integrity (adherence and competence) of modern, personalized CBT. The Inventory of Therapeutic Interventions and Skills (ITIS) was developed based on two existing CBT adherence and competence scales and augmented with third wave content and overarching therapeutic strategies. The instrument was then applied by graduate students and post-graduate clinicians to rate N = 185 therapy videos from N = 70 patients treated at a university outpatient clinic. Descriptive results, inter-rater reliability, item structure, and associations with session outcome and alliance were examined. Average inter-rater reliability was excellent for Interventions items and good for Skills items. Intercorrelations were low between Interventions items, but higher and significant between Skills items, which loaded on a single factor. Several ITIS items were shown to be predictive of session outcome and alliance, even after controlling for the nested data structure. Implications of these results for future research and clinical training are discussed.

Associations between mindfulness and general change mechanisms in individual therapy: Secondary results of a randomised controlled trial

AbstractObjectiveThe aim of this study was to investigate the possible effects of the interactions between mindfulness‐based session introductions and general change mechanisms on outcome in cognitive behavioural therapy (CBT).MethodWe investigated the effects of different session introductions on change mechanisms and their prediction of clinical outcome over the course of 25 sessions of training therapy. We compared session introductions with mindfulness (M) elements with progressive muscle relaxation (PMR) and treatment as usual (TAU) in routine CBT sessions conducted by therapists in training. Participants included 162 patients who were diagnosed primarily with either depression or anxiety disorder and 48 therapists. To address the nested data structure, we used multilevel models.ResultsMultilevel modelling (MLM) indicated that the interaction between an increase in several general change mechanisms and treatment condition (TAU + M vs. TAU + PMR) on therapeutic outcome was significant. From the patient perspective, MLM indicated that only the interaction between an increase in one general change mechanism (clarification of meaning) and treatment condition (TAU + M vs. TAU + PMR) turned out significantly. From both perspectives, correlation analyses revealed that general change mechanism effects on outcome were stronger in TAU + M than in TAU + PMR.ConclusionMindfulness has an impact on general change mechanisms, especially for therapists. These associations lead to a significantly higher effect on outcome compared to introductions with PMR.

Incidental Second-Level Dependence in Educational Survey Data with a Nested Data Structure

Many national and international educational data collection programs offer researchers opportunities to investigate contextual effects related to student performance. In those programs, schools are often used in the first-stage sampling process and students are randomly drawn from selected schools. However, the incidental dependence of students within classrooms, which are not part of the sampling design, may violate assumptions of statistical models, but this nesting also offers the opportunity for educational researchers to evaluate contextual effects. In this manuscript, we utilize the Early Childhood Longitudinal Study-Kindergarten dataset to demonstrate impacts of incidental dependence using a two-level model and a three-level model. We then illustrate, through a simulation, that both models can yield unbiased parameter estimates. However, two-level models tend to provide underestimated standard errors for fixed effects at the incidental level, and variance components of the random effect at the incidental level are divided into the flanking levels when it is ignored. In addition, another method of modeling nested data, using generalized estimating equations, was also compared with the model-based methods.

Investigating the relationship between district-level socioeconomic status and individual obesity in Taiwanese adolescents: A large-scale cross-sectional analysis

The current study aimed to assess the prevalence of obesity and to explore the relationship between socioeconomic status and obesity among adolescents in Taiwan, a transitioning country. Data from the Taiwan School Physical Fitness Database on 1,875,627 Taiwanese adolescents aged 10–18 years were analyzed. The average family income per household in each district was collected from the national statistical institutional database. Descriptive statistics, Chi-square tests, Pearson correlation analysis, and mixed model analyses were used. The overall prevalence of combined overweight and obesity was 28.1%. The prevalence of overweight/obesity significantly differed according to gender and age. Furthermore, the average family income per household was negatively associated with the district-level prevalence of obesity. Additionally, when controlling for physical fitness, the average family income per household remained negatively associated with adolescent obesity. In addition, multilevel analysis was also applied to explore the relationship between district-level socioeconomic status and individual-level obesity to prevent the nested data structure from affecting the results. The results revealed that the average family income per household negatively correlated with individual obesity. These findings provide insight for public health officials into preventing and managing adolescent obesity.

Online estimation of individual-level effects using streaming shrinkage factors

It has become increasingly easy to collect data from individuals over long periods of time. Examples include smart-phone applications used to track movements with GPS, web-log data tracking individuals’ browsing behavior, and longitudinal (cohort) studies where many individuals are monitored over an extensive period of time. All these datasets cover a large number of individuals and collect data on the same individuals repeatedly, causing a nested structure in the data. Moreover, the data collection is never ‘finished’ as new data keep streaming in. It is well known that predictions that use the data of the individual whose individual-level effect is predicted in combination with the data of all the other individuals, are better in terms of squared error than those that just use the individual mean. However, when data are both nested and streaming, and the outcome variable is binary, computing these individual-level predictions can be computationally challenging. Five computationally-efficient estimation methods which do not revise “old” data but do account for the nested data structure are developed and evaluated. The methods are based on existing shrinkage factors. A shrinkage factor is used to predict an individual-level effect (i.e., the probability to score a 1), by weighing the individual mean and the mean over all data points. The performance of the existing and newly developed shrinkage factors are compared in a simulation study. While the existing methods differ in their prediction accuracy, the differences in accuracy between the novel shrinkage factors and the existing methods are extremely small. The novel methods are however computationally much more appealing.

Robot Acceptance at Work: A Multilevel Analysis Based on 27 EU Countries

Robots are increasingly being used to assist with various tasks ranging from industrial manufacturing to welfare services. This study analysed how robot acceptance at work (RAW) varies between individual and national attributes in EU 27. Eurobarometer surveys collected in 2012 (n = 26,751) and 2014 (n = 27,801) were used as data. Background factors also included country-specific data drawn from the World Bank DataBank. The study is guided by the technology acceptance model and change readiness perspective explaining robot acceptance in terms of individual and cultural attributes. Multilevel studies analysing cultural differences in technological change are exceptionally rare. The multilevel analysis of RAW performed herein accounted for individual and national factors using fixed and random intercepts in a nested data structure. Individual-level factors explained RAW better than national-level factors. Particularly, personal experiences with robots at work or elsewhere were associated with higher acceptance. At a national level, the technology orientation of the country explained RAW better than the relative risk of jobs being automated. Despite the countries’ differences, personal characteristics and experiences with robots are decisive for RAW. Experiences, however, are better enabled in countries open to innovations. The findings are discussed in terms of possible mechanisms through which the technological orientation and social acceptance of robots may be related.

Competence Differences in Literacy, Numeracy, and Problem Solving According to Sex

This study aimed to determine individual- and country-level latent classes in literacy, numeracy and problem-solving competencies of individuals participating in the Programme for the International Assessment of Adult Competencies 2015. Specifically, it sought to distinguish these classes in relation to individuals’ sex and to identify the state of prediction of the determined latent classes by each person’s level of education. The study population consisted of 116,301 adults aged 16 to 65 years in 20 countries. Multilevel latent class analysis was conducted to consider the nested data structure and determine the number of latent classes. According to the results of the multilevel latent class analysis, Turkey and Chile were in the low achievement group in all skills, while Japan was in the most successful group. Moreover, the results revealed that sex and education level had a considerable influence on certain competence levels.

Columnar data processing for HEP analysis

In the last stages of data analysis, physicists are often forced to choose between simplicity and execution speed. In High Energy Physics (HEP), high-level languages like Python are known for ease of use but also very slow execution. However, Python is used in speed-critical data analysis in other fields of science and industry. In those fields, most operations are performed on Numpy arrays in an array programming style; this style can be adopted for HEP by introducing variable-sized, nested data structures. We describe how array programming may be extended for HEP use-cases and an implementation known as awkward-array. We also present integration with ROOT, Apache Arrow, and Parquet, as well as preliminary performance results.

Examining Determinants of Sport Event Attendance: A Multilevel Analysis of a Major League Baseball Season

Attendance data in sports leagues are organized at more than one level and thus are nested data with factors varying by individual games (e.g. single game attendance figure), which are nested within the factors varying by teams (e.g. local population and income level, stadium). Such a nested data structure requires a multilevel modeling approach for an accurate examination of attendance. Also, although the previous application of multilevel modeling to sport attendance research has been limited to the data aggregated and averaged by seasons or by teams, no multilevel modeling of attendance data at the individual game level is available. The uniqueness of the current study is its extension of the multilevel modeling approach by analyzing the audience data at the individual game level. Thus, this study examined the relationship between attendance determinants and single game attendance in a Major League Baseball (MLB) season by using multilevel modeling with 14 game level variables and 12 team level variables. Twelve predictors (e.g. home/visiting team playoff appearances, day of game, season progress, ticket price, stadium capacity, game uncertainty) significantly (p < .05) influenced attendance.

Residual Normality Assumption and the Estimation of Multiple Membership Random Effects Models

While conventional hierarchical linear modeling is applicable to purely hierarchical data, a multiple membership random effects model (MMrem) is appropriate for nonpurely nested data wherein some lower-level units manifest mobility across higher-level units. Although a few recent studies have investigated the influence of cluster-level residual nonnormality on hierarchical linear modeling estimation for purely hierarchical data, no research has examined the statistical performance of an MMrem given residual non-normality. The purpose of the present study was to extend prior research on the influence of residual non-normality from purely nested data structures to multiple membership data structures. Employing a Monte Carlo simulation study, this research inquiry examined two-level MMrem parameter estimate biases and inferential errors. Simulation factors included the level-two residual distribution, sample sizes, intracluster correlation coefficient, and mobility rate. Results showed that estimates of fixed effect parameters and the level-one variance component were robust to level-two residual non-normality. The level-two variance component, however, was sensitive to level-two residual non-normality and sample size. Coverage rates of the 95% credible intervals deviated from the nominal value assumed when level-two residuals were non-normal. These findings can be useful in the application of an MMrem to account for the contextual effects of multiple higher-level units.

Problem-Specific Coaching Interventions Influence Goal Attainment via Double-Loop Learning

Abstract. How do coaching interventions work? This study contributes to process-outcome research in coaching by examining the impact of coaches’ problem-specific interventions on coaching results. Paying particular attention to the mediation processes in coaching seems to be a promising research area, which has received little attention to date. Using several points of measurement over time and taking nested data structure into account, 33 coaching dyads were analyzed using longitudinal multilevel analyses. Results showed a general coaching effect as well as the mediating role of learning processes. As expected, clients’ perceived goal attainment increased over the course of the coaching sessions. Furthermore, problem-specific interventions by the coach, especially clarification of meaning and mastery/coping, led to an increase in goal attainment mediated by clients’ perceived learning experience (double-loop learning). Our results underscore the importance of stimulating learning processes by specific interventions to improve coaching effectiveness.

Adolescents' academic engagement mediates the association between Internet addiction and academic achievement: The moderating effect of classroom achievement norm

Although the association between Internet addiction and academic achievement has been widely observed in adolescents in recent years, little is known about the mechanisms of Internet addiction and its adverse effects on academic achievement in terms of individual factors and classroom contexts. The present study examined the relationship between Internet addiction and academic outcomes, the mediating role of academic engagement, and the cross-level moderating effect of classroom achievement norm. By using a short-term longitudinal design, two-wave data were obtained from 2407 students from 47 classes in China over six-months. A multilevel structural equation model was used to explain the nested data structure. Results indicated that the negative effect of Internet addiction on adolescents' later academic achievement was mediated by academic engagement and disaffection at the individual level. Furthermore, classroom achievement norm moderated the association between academic disaffection and achievement, with a significant indirect effect in the high classroom achievement norm classes and a non-significant effect in the low classroom achievement norm classes. These results held after controlling for an individual's gender, age and academic achievement in the first term at the individual level and the class size and school at the classroom level. In conclusion, this study reveals the internal mechanism of how students' Internet addiction leads to the decline of academic achievement. It also provids a practical and effective outlook on the prevention and intervention of the negative effects of Internet addiction in the future teaching practice from the perspective of personal motivation and classroom environment.

Client predictors of the therapeutic alliance in individual resiliency training for first episode psychosis

Individuals experiencing their first episode of psychosis (FEP) are often reluctant to seek treatment, and are difficult to engage and retain in mental health services. The therapeutic alliance (TA), or the affective and collaborative bond between therapist and client, is predictive of better treatment outcomes for clients with FEP; thus, it is important to understand the predictors of the TA in order to determine how best to foster a positive alliance with these individuals. The primary aim of the present study was to examine whether baseline client characteristics, including severity of symptoms, social functioning, and insight, were associated with the TA. The exploratory aim was to examine associations between demographic variables (age, race, and gender) and the TA. The present study included a subsample of participants (n = 134) who received Individual Resiliency Training (IRT) as part of the NAVIGATE treatment in the Recovery After An Initial Schizophrenia Episode Early Treatment Program study. Four trained research assistants rated the TA from early audiotaped sessions of IRT. Multilevel modeling was utilized given the nested data structure. Results indicated that more severe positive and less severe negative symptoms were significantly and uniquely associated with a better therapeutic alliance, as was female gender. The findings suggest that client symptom profiles should be considered when developing a TA with FEP clients.

Communication Among Team Members Within the Patient-centered Medical Home and Patient Satisfaction With Providers: The Mediating Role of Patient-Provider Communication.

The Patient-centered Medical Home (PCMH) uses team-based care to improve patient outcomes, including satisfaction. The quality of patients' communication with their primary care providers (PCPs) is a key determinant of patient satisfaction. A shift to team-based care could disrupt the therapeutic relationship between patients and their PCPs and reduce patient satisfaction if communication and coordination among primary care team members is poor. Little is known about the relationship between intrateam communication within a PCMH and patient satisfaction with PCPs, and whether patient-provider communication might mediate this relationship. To examine the relationship between intrateam communication in a PCMH and patients' satisfaction with assigned PCPs, and whether patient-provider communication mediates this relationship. Cross-sectional surveys of Veterans Health Administration PCPs (2011-2012, n=149) matched with their assigned patients' surveys (n=3329). Mediation analyses using a nested data structure, controlling for patient and provider characteristics. Patient satisfaction with PCPs, patient-reported patient-provider communication, and PCP-reported intrateam communication within the PCMH. Intrateam communication and patient-provider communication were independently associated with patients' satisfaction with their PCPs. Patient-provider communication mediated 56% of the association between intrateam communication and patient satisfaction. Better intrateam communication combined with better patient-provider communication predicted high satisfaction (81%), compared with poor intrateam communication and poor patient-provider communication (22%). PCMH environments with better communication among team members are likely to experience better patient-provider communication and high patient satisfaction. PCMH practices with low ratings of patient satisfaction may need to look beyond individual PCPs to communication within and across teams.

Suicidal Ideation and Skill Use During In-patient Dialectical Behavior Therapy for Borderline Personality Disorder. A Diary Card Study.

Associations between suicidal ideation and skill use were investigated during in-patient dialectical behavior therapy (DBT) for borderline personality disorder (BPD). Participants were N = 44 patients with BPD undergoing a 5-week in-patient DBT program in a psychiatric clinic. They filled in a diary card each treatment day resulting in 1,334 skill use ratings and 1,364 suicidal ideation ratings. Treatment days were categorized as days with successful skill use (using skills and perceiving them as effective), days with no skill use, days with unsuccessful skill use (using skills but perceiving them as ineffective). Multilevel models were performed to account for the nested data structure. The results showed that suicidal ideation improved more for patients who applied skills successfully more often during treatment (p < 0.05). Moreover, suicidal ideation was lower on treatment days with successful skill use compared to treatment days with no skill use and compared to treatment days with unsuccessful skill use (p < 0.05). When treatment days with no skill use were compared to treatment days with unsuccessful skill use, suicidal ideation was higher on treatment days with unsuccessful skill use (p < 0.05). To conclude, using skills successfully on as many treatment days as possible is associated with lower suicidal ideation.

Model selection in Bayesian framework to identify the best WorldView-2 based vegetation index in predicting green biomass of salt marshes in the northern Gulf of Mexico

Coastal wetlands are among the most productive ecosystems globally but have experienced dramatic degradation and loss within the past several decades. Vegetation biomass of coastal wetlands is not only the key component of blue carbon storage but also plays an important role in vertical accretion, important for maintaining these habitats under relative sea-level rise. Remote sensing offers a cost-effective approach to study vegetation biomass at a broad spatial scale. We developed statistical models to predict peak aboveground green biomass of Spartina alterniflora and Juncus roemerianus, two dominant species of salt marshes using WorldView-2 satellite imagery at the Grand Bay National Estuarine Research Reserve (NERR) on the Mississippi coast in the northern Gulf of Mexico. The model accounted for nested data structures in the sampled biomass, assimilated uncertainties from data, parameters and model structures, and helped determine the best vegetation index among a variety of commonly-used indices to predict aboveground green biomass. We developed a series of mixed-effects models, which included different combinations of fixed effect(s), random intercept, and random slope(s). The fixed effects were species and one of the 60 vegetation indices derived from a WorldView-2 image obtained on 6 October 2012. The random effect used was site. We implemented the models in a Bayesian framework and selected the best model structure and vegetation index based on minimum posterior predictive loss and deviance information criterion. The results showed that the best vegetation index to predict peak green biomass was the green chlorophyll index derived from the reflectance values of band 8 (near-infrared) and band 3 (green), and its effect on biomass prediction varied among sites. The inclusion of species as a fixed effect improved the model prediction. The study demonstrated the need to account for spatial dependence of data in developing a robust model, and the importance of the second WorldView-2 near-infrared band (860–1040 nm) in predicting aboveground green biomass for the Grand Bay NERR. The analysis using mixed-effects modeling in Bayesian inference which coherently combined field and WorldView-2 data with uncertainties accounted for provides a robust and nondestructive tool for resource managers to monitor the status of coastal wetlands at a high spatial resolution in a timely manner. Through this study, we hope to emphasize the importance of appropriately accounting for nested data structures using mixed-effects models and promote wider application of Bayesian inference to facilitate assimilation of uncertainties in remote sensing applications.

Belief in a Just World as a Resource of Victimized Students

In a cross-sectional questionnaire study with N = 2593 German students, aged between 12 and 17 years (M = 14.1, SD = 0.5), we investigated the relation between students’ personal belief in a just world (BJW) and their victimization in bullying situations. According to the just-world theory and research, we expected that the more strongly the students endorsed the personal BJW, the less likely they were to report being victimized by other students. We aimed to extend previous findings that failed to confirm this negative relation by considering students’ personal experience of teacher justice as a possible mediator in this relation, while statistically controlling for sex and school type. We further considered the nested data structure with regard to school classes in our analyses. The results of latent mediation analyses at the individual and group levels showed that the more the students endorsed personal BJW, the more they evaluated their teachers’ behavior toward them personally as being just, and the less likely they were to report that they were bullied. However, the students’ personal experience of teacher justice did not mediate the relation between personal BJW and victimization at the individual or group level when controlled for sex and school type. We discussed the adaptive functions of BJW and implications for future school research and practice.

The Utility of Between-Nation Subjective Wellbeing Comparisons Amongst Nations Within the European Social Survey

Between-nation differences in wellbeing are frequently reported. Such differences are attributed to between-nation differences in social, economic and political factors. However, there is a likelihood that between-nation differences are over-estimated as they fail to account for the extent to which wellbeing varies within-nation owing to within-nation factors. Participant data for 43,000 participants from 23 countries was obtained from wave 3 of the European Social Survey in 2006. Analyses were undertaken in a multi-level framework with citizens nested within-nation in order to derive maximum likelihood estimates and standard error which adjust for the nested data hierarchy. Participant data was adjusted for (1) a design weight which adjusted for a sampling probability reflecting their likelihood of being recruited for the study, and (2) a population weight which adjusts for the extent individuals reflected a nation’s population. Across wellbeing indicators, most variance was accounted for at the within-nation level (> 95%). Within-nation factors were the strongest drivers of wellbeing. Best linear unbiased predictions indicated that raw national aggregated well-being means over-estimate between-nation wellbeing differences. Many prior cross-national wellbeing comparisons likely overestimate between-nation differences as they fail to account for the nested data structure in which individual citizens are nested within countries. Between-nation factors were not substantive drivers of wellbeing outcomes in comparison with within-nation effects and interpretation of any between-nation effects need to be carefully considered since so little wellbeing variance is accounted for at the between-nation level.