On the merits of longitudinal multiple group modelling: an alternative to multilevel modelling for intervention evaluations

2021 ◽  
pp. 1-13
Author(s):  
Todd D. Little ◽  
Daniel Bontempo ◽  
Charlie Rioux ◽  
Allison Tracy
Keyword(s):  
Multilevel Modelling ◽  
Multiple Group ◽  
Intervention Evaluations

scholarly journals Multilevel modelling and multiple group analysis of disparities in continuity of care and viral suppression among adolescents and youths living with HIV in Nigeria

2020 ◽  
Vol 5 (11) ◽  
pp. e003269
Author(s):  
Okikiolu Badejo ◽  
Christiana Noestlinger ◽  
Toyin Jolayemi ◽  
Juliette Adeola ◽  
Prosper Okonkwo ◽  
...  
Keyword(s):  
Young Adults ◽  
Viral Suppression ◽  
Group Analysis ◽  
Age Groups ◽  
Multilevel Modelling ◽  
Retention In Care ◽  
Healthcare Facilities ◽  
Individual Level ◽  
Multiple Group ◽  
Facility Level

IntroductionSubstantial disparities in care outcomes exist between different subgroups of adolescents and youths living with HIV (AYLHIV). Understanding variation in individual and health facility characteristics could be key to identifying targets for interventions to reduce these disparities. We modelled variation in AYLHIV retention in care and viral suppression, and quantified the extent to which individual and facility characteristics account for observed variations.MethodsWe included 1170 young adolescents (10–14 years), 3206 older adolescents (15–19 years) and 9151 young adults (20–24 years) who were initiated on antiretroviral therapy (ART) between January 2015 and December 2017 across 124 healthcare facilities in Nigeria. For each age group, we used multilevel modelling to partition observed variation of main outcomes (retention in care and viral suppression at 12 months after ART initiation) by individual (level one) and health facility (level two) characteristics. We used multiple group analysis to compare the effects of individual and facility characteristics across age groups.ResultsFacility characteristics explained most of the observed variance in retention in care in all the age groups, with smaller contributions from individual-level characteristics (14%–22.22% vs 0%–3.84%). For viral suppression, facility characteristics accounted for a higher proportion of variance in young adolescents (15.79%), but not in older adolescents (0%) and young adults (3.45%). Males were more likely to not be retained in care (adjusted OR (aOR)=1.28; p<0.001 young adults) and less likely to achieve viral suppression (aOR=0.69; p<0.05 older adolescent). Increasing facility-level viral load testing reduced the likelihood of non-retention in care, while baseline regimen TDF/3TC/EFV or NVP increased the likelihood of viral suppression.ConclusionsDifferences in characteristics of healthcare facilities accounted for observed disparities in retention in care and, to a lesser extent, disparities in viral suppression. An optimal combination of individual and health services approaches is, therefore, necessary to reduce disparities in the health and well-being of AYLHIV.

An Adapted Measure of Teacher Self-Efficacy for Preservice Teachers: Exploring its Validity Across two Countries

2014 ◽  
Vol 28 (3) ◽  
pp. 83-92 ◽  
Author(s):  
Franziska Pfitzner-Eden ◽  
Felicitas Thiel ◽  
Jenny Horsley
Keyword(s):  
Preservice Teachers ◽  
Self Efficacy ◽  
Confirmatory Factor Analyses ◽  
3 Dimensional ◽  
Multiple Group ◽  
Three Samples ◽  
Confirmatory Factor ◽  
Teacher Self Efficacy ◽  
Dimensional Assessment ◽  
Metric Measurement

Teacher self-efficacy (TSE) is an important construct in the prediction of positive student and teacher outcomes. However, problems with its measurement have persisted, often through confounding TSE with other constructs. This research introduces an adapted TSE instrument for preservice teachers, which is closely aligned with self-efficacy experts' recommendations for measuring self-efficacy, and based on a widely used measure of TSE. We provide first evidence of construct validity for this instrument. Participants were 851 preservice teachers in three samples from Germany and New Zealand. Results of the multiple-group confirmatory factor analyses showed a uniform 3-factor solution for all samples, metric measurement invariance, and a consistent and moderate correlation between TSE and a measure of general self-efficacy across all samples. Despite limitations to this study, there is some first evidence that this measure allows for a valid 3-dimensional assessment of TSE in preservice teachers.

Adolescent Language: Models, Assessment, and Links to Reading

2019 ◽  
Author(s):  
Amanda Goodwin ◽  
Yaacov Petscher ◽  
Jamie Tock
Keyword(s):  
Reading Comprehension ◽  
Bifactor Model ◽  
Language Models ◽  
Multiple Group ◽  
Global Factor ◽  
Eighth Grade Students ◽  
Key Aspects ◽  
Future Work ◽  
The Relationship ◽  
Best Fit

Various models have highlighted the complexity of language. Building on foundational ideas regarding three key aspects of language, our study contributes to the literature by 1) exploring broader conceptions of morphology, vocabulary, and syntax, 2) operationalizing this theoretical model into a gamified, standardized, computer-adaptive assessment of language for fifth to eighth grade students entitled Monster, PI, and 3) uncovering further evidence regarding the relationship between language and standardized reading comprehension via this assessment. Multiple-group item response theory (IRT) across grades show that morphology was best fit by a bifactor model of task specific factors along with a global factor related to each skill. Vocabulary was best fit by a bifactor model that identifies performance overall and on specific words. Syntax, though, was best fit by a unidimensional model. Next, Monster, PI produced reliable scores suggesting language can be assessed efficiently and precisely for students via this model. Lastly, performance on Monster, PI explained more than 50% of variance in standardized reading, suggesting operationalizing language via Monster, PI can provide meaningful understandings of the relationship between language and reading comprehension. Specifically, considering just a subset of a construct, like identification of units of meaning, explained significantly less variance in reading comprehension. This highlights the importance of considering these broader constructs. Implications indicate that future work should consider a model of language where component areas are considered broadly and contributions to reading comprehension are explored via general performance on components as well as skill level performance.

Longitudinal and Gender Measurement Invariance of the Gratitude Questionnaire in Chinese Adolescents

2021 ◽  
pp. 003329412110360
Author(s):  
Qingsong Tan ◽  
Jilin Zou ◽  
Feng Kong
Keyword(s):  
Measurement Invariance ◽  
Time Interval ◽  
Longitudinal Measurement Invariance ◽  
Multiple Group ◽  
Longitudinal Measurement ◽  
Confirmatory Factor ◽  
Adolescent Sample ◽  
Good Internal Consistency ◽  
And Gender ◽  
Mean Differences

The 5-item Gratitude Questionnaire (GQ-5) is one of the most commonly used instruments to measure dispositional gratitude in adolescents. The purpose of this study was to verify the longitudinal measurement invariance (LMI) and gender measurement invariance (GMI) of the GQ-5 that was administered to an adolescent sample twice over the course of 18 months ( N = 669). Single-group confirmatory factor analysis (CFA) was adopted to examine the LMI and multiple-group CFA was conducted to assess the GMI. The results showed that the GQ-5 had strong invariance (i.e., equality of factor patterns, loadings, and intercepts) across time and gender. Validation of latent factor mean differences showed that females had higher gratitude scores than males. In addition, the GQ-5 exhibited good internal consistency indices across time and a moderate stability coefficient was also found across an 18-month time interval in adolescents. In summary, our study showed that LMI and GMI of the GQ-5 are satisfactory and the GQ-5 is a reliable instrument for measuring gratitude in adolescents.

Development and Validation of the Economic Coercion Scale 36 (ECS-36) in Rural Bangladesh

2021 ◽  
pp. 088626052098781
Author(s):  
Kathryn M. Yount ◽  
Yuk Fai Cheong ◽  
Stephanie Miedema ◽  
Ruchira T. Naved
Keyword(s):  
Factor Analysis ◽  
Partner Violence ◽  
Short Form ◽  
Latent Trait ◽  
Married Women ◽  
Primary Data ◽  
Group Factor ◽  
Rural Bangladesh ◽  
Multiple Group ◽  
Economic Coercion

Assessing progress toward Sustainable Development Goal (SDG) 5, to achieve gender equality and to empower women, requires monitoring trends in intimate partner violence (IPV). Current measures of IPV may miss women’s experiences of economic coercion, or interference with the acquisition, use, and maintenance of financial resources. This sequential, mixed-methods study developed and validated a scale for economic coercion in married women in rural Bangladesh, where women’s expanding economic opportunities may elevate the risks of economic coercion and other IPV. Forty items capturing lifetime and prior-year economic coercion were adapted from formative qualitative research and prior scales and administered to a probability sample of 930 married women 16–49 years. An economic coercion scale (ECS) was validated using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) with primary data from random-split samples ( N1 = 310; N2 = 620). Item response theory (IRT) methods gauged the measurement precision of items and scales over the range of the economic-coercion latent trait. Multiple-group factor analysis assessed measurement invariance of the economic-coercion construct. Two-thirds (62.26%) of women reported any lifetime economic coercion. EFA suggested a 36-item, two-factor model capturing barriers to acquire and to use or maintain economic resources. CFA, multiple group factor analysis, and multidimensional IRT methods confirmed that this model provided a reasonable fit to the data. IRT analysis showed that each dimension provided most precision over the higher range of the economic coercion trait. The Economic Coercion Scale 36 (ECS-36) should be validated elsewhere and over time. It may be added to violence-specific surveys and evaluations of violence-prevention and economic-empowerment programs that have a primary interest measuring economic coercion. Short-form versions of the ECS may be developed for multipurpose surveys and program monitoring.

scholarly journals The impact of non-pharmaceutical interventions on COVID-19 epidemic growth in the 37 OECD member states

2021 ◽  
Author(s):  
Francisco Pozo-Martin ◽  
Heide Weishaar ◽  
Florin Cristea ◽  
Johanna Hanefeld ◽  
Thurid Bahr ◽  
...  
Keyword(s):  
Growth Rate ◽  
Early Phase ◽  
Policy Decision ◽  
Multilevel Modelling ◽  
Daily Growth ◽  
Public Places ◽  
Member States ◽  
Unit Increase ◽  
Policy Decision Making ◽  
The Impact

AbstractWe estimated the impact of a comprehensive set of non-pharmeceutical interventions on the COVID-19 epidemic growth rate across the 37 member states of the Organisation for Economic Co-operation and Development during the early phase of the COVID-19 pandemic and between October and December 2020. For this task, we conducted a data-driven, longitudinal analysis using a multilevel modelling approach with both maximum likelihood and Bayesian estimation. We found that during the early phase of the epidemic: implementing restrictions on gatherings of more than 100 people, between 11 and 100 people, and 10 people or less was associated with a respective average reduction of 2.58%, 2.78% and 2.81% in the daily growth rate in weekly confirmed cases; requiring closing for some sectors or for all but essential workplaces with an average reduction of 1.51% and 1.78%; requiring closing of some school levels or all school levels with an average reduction of 1.12% or 1.65%; recommending mask wearing with an average reduction of 0.45%, requiring mask wearing country-wide in specific public spaces or in specific geographical areas within the country with an average reduction of 0.44%, requiring mask-wearing country-wide in all public places or all public places where social distancing is not possible with an average reduction of 0.96%; and number of tests per thousand population with an average reduction of 0.02% per unit increase. Between October and December 2020 work closing requirements and testing policy were significant predictors of the epidemic growth rate. These findings provide evidence to support policy decision-making regarding which NPIs to implement to control the spread of the COVID-19 pandemic.

scholarly journals Sampling weights in multilevel modelling: an investigation using PISA sampling structures

2021 ◽  
Vol 9 (1) ◽  
Author(s):  
Julia Mang ◽  
Helmut Küchenhoff ◽  
Sabine Meinck ◽  
Manfred Prenzel
Keyword(s):  
Hierarchical Models ◽  
Large Scale ◽  
Practical Importance ◽  
Multilevel Modelling ◽  
Likelihood Method ◽  
Multilevel Regression ◽  
Sampling Weights ◽  
International Student Assessment ◽  
Software Packages ◽  
Software Programs

Abstract Background Standard methods for analysing data from large-scale assessments (LSA) cannot merely be adopted if hierarchical (or multilevel) regression modelling should be applied. Currently various approaches exist; they all follow generally a design-based model of estimation using the pseudo maximum likelihood method and adjusted weights for the corresponding hierarchies. Specifically, several different approaches to using and scaling sampling weights in hierarchical models are promoted, yet no study has compared them to provide evidence of which method performs best and therefore should be preferred. Furthermore, different software programs implement different estimation algorithms, leading to different results. Objective and method In this study, we determine based on a simulation, the estimation procedure showing the smallest distortion to the actual population features. We consider different estimation, optimization and acceleration methods, and different approaches on using sampling weights. Three scenarios have been simulated using the statistical program R. The analyses have been performed with two software packages for hierarchical modelling of LSA data, namely Mplus and SAS. Results and conclusions The simulation results revealed three weighting approaches performing best in retrieving the true population parameters. One of them implies using only level two weights (here: final school weights) and is because of its simple implementation the most favourable one. This finding should provide a clear recommendation to researchers for using weights in multilevel modelling (MLM) when analysing LSA data, or data with a similar structure. Further, we found only little differences in the performance and default settings of the software programs used, with the software package Mplus providing slightly more precise estimates. Different algorithm starting settings or different accelerating methods for optimization could cause these distinctions. However, it should be emphasized that with the recommended weighting approach, both software packages perform equally well. Finally, two scaling techniques for student weights have been investigated. They provide both nearly identical results. We use data from the Programme for International Student Assessment (PISA) 2015 to illustrate the practical importance and relevance of weighting in analysing large-scale assessment data with hierarchical models.

Sign in / Sign up

Export Citation Format

Share Document