Bayesian Multilevel Analysis Research Articles

Global perspectives on physical and nonphysical discipline: A Bayesian multilevel analysis

Background and Objective: Sixty countries worldwide have banned the use of physical punishment, yet little is known about the association of physical and nonphysical forms of child discipline with child development in a global context. The objective of this study is to examine whether physical punishment and nonphysical discipline are associated with child socioemotional functioning in a global sample of families from 62 countries and whether country-level normativeness of physical punishment and nonphysical discipline moderated those associations. Methods: Data for this study are from 215,885 families in the fourth and fifth rounds of the United Nations Children’s Fund Multiple Indicator Cluster Surveys. Bayesian multilevel logistic models were used to analyze the associations of physical punishment and nonphysical discipline (i.e., taking away privileges and verbal reasoning) with three different outcomes representing children’s socioemotional functioning: getting along well with other children, aggression, and becoming distracted. Results: The use of physical punishment was not associated with getting along with other children, was associated with increased aggression, and was associated with increases in distraction. Taking away privileges was associated with lower levels of getting along with other children, higher levels of aggression, and higher levels of becoming distracted. Verbal reasoning (i.e., explaining why a behavior was wrong) was associated with higher levels of getting along with other children, higher levels of aggression, and higher levels of becoming distracted. Country-level normativeness moderated some of these associations but in general the direction of effects was consistent. Conclusions: Results suggest that eliminating physical punishment would benefit children across the globe and align with the United Nations Convention on the Rights of the Child, which calls for all children to be free from physical violence. More attention needs to be focused on the associations of nonphysical forms of discipline with child functioning across the globe.

Controllability over stressor decreases responses in key threat-related brain areas

Controllability over stressors has major impacts on brain and behavior. In humans, however, the effect of controllability on responses to stressors is poorly understood. Using functional magnetic resonance imaging (fMRI), we investigated how controllability altered responses to a shock-plus-sound stressor with a between-group yoked design, where participants in controllable and uncontrollable groups experienced matched stressor exposure. Employing Bayesian multilevel analysis at the level of regions of interest and voxels in the insula, and standard voxelwise analysis, we found that controllability decreased stressor-related responses across threat-related regions, notably in the bed nucleus of the stria terminalis and anterior insula. Posterior cingulate cortex, posterior insula, and possibly medial frontal gyrus showed increased responses during control over stressor. Our findings support the idea that the aversiveness of stressors is reduced when controllable, leading to decreased responses across key regions involved in anxiety-related processing, even at the level of the extended amygdala.

Bayesian Multilevel Analysis of Postnatal Care Utilization in Ethiopia

The postnatal period is defined as the first six weeks after birth; it is critical to the health and survival. The study was conducted to determine the key factors of maternal postnatal care using Bayesian multilevel analysis using 7,192 mothers aged between 15-49 of EDHS 2016 data. Descriptive statistics revealed that 675 (9.4%) women received postnatal care. The Bayesian multilevel model used to analyze postnatal care utilization of woman. The predictor variables like mother's age, region, place of residence, media exposure, religion, wealth index, parents' education level, and mothers' occupation were found to be significant determinants of variation in women postnatal care utilization in Ethiopia. The Bayesian multilevel logistic regression analysis also estimates the variance of random effect at the regional level. Thus, the value of var (u0j) = δ2= 0.127 is significant at the 5% level of significance. In order to convey policies, it is important not only to comprehend the effect of reforms on the issues, but also on the how it diverges with demographic and economic characteristics.

Real-World Language Use With Familiar Versus Unfamiliar Interlocutors in Young and Older Adults

Real-world contexts may compensate for age-related changes in language production. We compared age effects on vocabulary richness (i.e., entropy) and grammatical complexity (i.e., clause length) in conversations with familiar interlocutors (i.e., significant other, friends, family members) versus with strangers. We collected thousands of 30-seconds speech samples from 61 young and 48 healthy older adults across four days using a portable audio recording device — the Electronically Activated Recorder (EAR). Bayesian multilevel analyses showed that participants used richer vocabulary and more complex grammar with familiar interlocutors than strangers. Young adults used richer vocabulary than older adults. Furthermore, older adults produced equally complex grammar with the significant other as young adults did, but simpler grammar with friends and family members. We found no age group differences in grammatical complexity with strangers (lacking statistical power). In sum, familiarity with the significant other may benefit older adults in producing complex grammar in real-world conversations.

Investigation of influential factors of tire/pavement noise: A multilevel Bayesian analysis of full-scale track testing data

Tire/pavement noise is a primary source of traffic noise that continuously interferes with human life. In this study, we employed a Bayesian multilevel model to reveal the causal relationship between the resulting noise and the considered variables and quantify their uncertainties. The experimental design comprised four primary factors: pavement type, tire type, driving speed, and trailer weight. We also included the impacts of pavement surface characteristics and environmental factors. The results show the proposed Bayesian model is more predictive than its fixed effects counterpart. Among the considered factors, the influence of speed dominates all other factors on the noise levels. The tire/pavement noise increases with macrotexture depth while decreases with the porosity of the surface material. The porous asphalt pavement attenuates noise more significantly at a higher speed. The noise level increases statistically when the tire load increases, while it decreases as the air temperature increases.

A Bayesian multilevel analysis of the longitudinal associations between relationship quality and suicidal ideation and attempts among youth with bipolar disorder.

Youth with bipolar disorder (BD) are at high risk for suicidal thoughts and behaviors and frequently experience interpersonal impairment, which is a risk factor for suicide. Yet, no study to date has examined the longitudinal associations between relationship quality in family/peer domains and suicidal thoughts and behaviors among youth with BD. Thus, we investigated how between-person differences - reflecting the average relationship quality across time - and within-person changes, reflecting recent fluctuations in relationship quality, act as distal and/or proximal risk factors for suicidal ideation (SI) and suicide attempts. We used longitudinal data from the Course and Outcome of Bipolar Youth Study (N=413). Relationship quality variables were decomposed into stable (i.e., average) and varying (i.e., recent) components and entered, along with major clinical covariates, into separate Bayesian multilevel models predicting SI and suicide attempt. We also examined how the relationship quality effects interacted with age and sex. Poorer average relationship quality with parents (β=-.33, 95% Bayesian highest density interval (HDI) [-0.54, -0.11]) or friends (β=-.33, 95% HDI [-0.55, -0.11]) was longitudinally associated with increased risk of SI but not suicide attempt. Worsening recent relationship quality with parents (β=-.10, 95% HDI [-0.19, -0.03]) and, to a lesser extent, friends (β=-.06, 95% HDI [-0.15, 0.03]) was longitudinally associated with increased risk of SI, but only worsening recent relationship quality with parents was also associated with increased risk of suicide attempt (β=-.15, 95% HDI [-0.31, 0.01]). The effects of certain relationship quality variables were moderated by gender but not age. Among youth with BD, having poorer average relationship quality with peers and/or parents represents a distal risk factor for SI but not suicide attempts. Additionally, worsening recent relationship quality with parents may be a time-sensitive indicator of increased risk for SI or suicide attempt.

Tenets of self-determination theory as a mechanism behind challenge demands: a within-person study

PurposeThis study investigated a mechanism by which challenge stressors may affect employee well-being outcomes. This study tested a within-person longitudinal model in which the effects of challenge demands relate to basic psychological need satisfaction/thwarting and worker well-being outcomes. In particular, basic psychological need satisfaction and thwarting were hypothesized to mediate challenge demands and outcomes at the intraindividual level.Design/methodology/approachData from 84 employees from a weekly survey across four weeks (308 observations) were used in Bayesian multilevel path analyses to test hypotheses.FindingsAlthough significant indirect effects showed that basic psychological needs mediate between demands and worker outcomes, only a few specific indirect effects (e.g. the path from time pressure via thwarting the need for autonomy to emotional exhaustion) operated as hypothesized. Interestingly, in this study, time pressure was only mediated via thwarting the need for autonomy when considering undesirable worker outcomes (i.e. increased emotional exhaustion, decreased job satisfaction). Job complexity, however, led to decreased emotional exhaustion via the need for competence satisfaction. Implications for need satisfaction and thwarting as mechanisms in the challenge–hindrance framework are discussed.Originality/valueThis study (1) extends the challenge–hindrance framework to include basic psychological needs as a mechanism, (2) expands basic psychological needs to include need thwarting and (3) may enhance our understanding of stressor categories.

Testing the Homogeneity Hypothesis of Personality: Replication and Extension across European Countries, Industry Sectors and Organizations

This paper builds on previous studies testing the homogeneity hypothesis of personality as presented in seminal works in the field. Using a sample of 2,745 managers from 165 organizations operating in 51 sectors, spanning 30 European countries, we explore the Attraction‐Selection‐Attrition (ASA) framework in European organizations, and examine managerial personality variability within and between: (1) organizations; (2) industry sectors; and (3) European countries. To explore personality, the Myers‐Briggs Type Indicator (MBTI) was used and the hypotheses tested with Bayesian Multilevel analysis, along with the within‐group interrater reliability statistic (rWG) and average deviation from the mean (ADM) agreement indices. Results revealed significant variance in personality between organizations and countries – but not within sectors. Evaluation of within group agreement revealed that agreement in personality profiles within organizations is relatively low (10–39%); the agreement is higher within countries (56–68%) and within industry sectors (30–63%). The results reveal that perhaps organizations are not as homogeneous as previously suggested. Earlier cautions about the consequences of Attraction‐Selection‐Attrition yielding relative homogeneity are discussed as unnecessary, and complementary fit is suggested. This is the first study to empirically test within group agreement across levels.

Determinants of malaria among under-five children in Ethiopia: Bayesian multilevel analysis

BackgroundIn Ethiopia, malaria is one of the public health problems, and it is still among the ten top leading causes of morbidity and mortality among under-five children.However, the studies conducted in the country have been inconclusive and inconsistent. Thus, this study aimed to assess factors associated with malaria among under-five children in Ethiopia.MethodsWe retrieved secondary data from the malaria indicator survey data collected from September 30 to December 10, 2015, in Ethiopia. A total of 8301 under-five-year-old children who had microscopy test results were included in the study. Bayesian multilevel logistic regression models were fitted and Markov chain Monte Carlo simulation was used to estimate the model parameters using Gibbs sampling. Adjusted Odd Ratio with 95% credible interval in the multivariable model was used to select variables that have a significant association with malaria.ResultsIn this study, sleeping under the insecticide-treated bed nets during bed time (ITN) [AOR 0.58,95% CI, 0.31–0.97)], having 2 and more ITN for the household [AOR 0.43, (95% CI, 0.17–0.88)], have radio [AOR 0.41, (95% CI, 0.19–0.78)], have television [AOR 0.19, (95% CI, 0.01–0.89)] and altitude [AOR 0.05, (95% CI, 0.01–0.13)] were the predictors of malaria among under-five children.ConclusionsThe study revealed that sleeping under ITN, having two and more ITN for the household, altitude, availability of radio, and television were the predictors of malaria among under-five children in Ethiopia. Thus, the government should strengthen the availability and utilization of ITN to halt under-five mortality due to malaria.

With a little help from familiar interlocutors: real-world language use in young and older adults

Objectives Functional psychologists are concerned with the performance of cognitive activities in the real world in relation to cognitive changes in older age. Conversational contexts may mitigate the influence of cognitive aging on the cognitive activity of language production. This study examined effects of familiarity with interlocutors, as a context, on language production in the real world. Method We collected speech samples using iPhones, where an audio recording app (i.e. Electronically Activated Recorder [EAR]) was installed. Over 31,300 brief audio files (30-second long) were randomly collected across four days from 61 young and 48 healthy older adults in Switzerland. We transcribed the audio files that included participants’ speech and manually coded for familiar interlocutors (i.e. significant other, friends, family members) and strangers. We computed scores of vocabulary richness and grammatical complexity from the transcripts using computational linguistics techniques. Results Bayesian multilevel analyses showed that participants used richer vocabulary and more complex grammar when talking with familiar interlocutors than with strangers. Young adults used more diverse vocabulary than older adults and the age effects remained stable across contexts. Furthermore, older adults produced equally complex grammar as young adults did with the significant other, but simpler grammar than young adults with friends and family members. Conclusion Familiarity with interlocutors is a promising contextual factor for research on aging and language complexity in the real world. Results were discussed in the context of cognitive aging.

Selective Influence of Smoking on Periodontal Treatment Outcomes after 3 Years of Follow-up.

The impact of smoking habits on periodontal treatment has not been clearly elucidated. This study aimed to specify the effects of cigarette consumption and nicotine addiction on periodontal therapy. In this retrospective case-control study, 20 moderate smokers and 20 non-smokers with severe periodontitis were examined after initial diagnosis, and non-surgical active and supportive therapies for 1-6 years (mean follow-up = 3.37 years). Fagerström's test of nicotine dependence (FTND) was evaluated at re-examination. Treatment efficacy was assessed by periodontal pocket probing depth (PPD) changes and number of teeth lost per year (TL). Bayesian multilevel and regression analyses were performed at site, tooth, and patient levels. During the mean follow-up period of > 3 years including active and supportive periodontal therapies, mean PPD, PPD > 3 mm and PPD > 7 mm percentage reductions were 1.03, 1.48 and 2.57 times statistically significantly less pronounced, respectively, in smokers than in non-smokers. Multilevel analysis showed that the variability of PPD > 7 mm reduction was mainly associated with patient-level factors. Smokers presented a higher risk for periodontitis progression. In smokers, periodontal parameter improvement was less pronounced in the maxilla and molars. The mean TL was related to the FTND score, not to cigarette consumption. Regression analysis did not demonstrate other influences of demographic and periodontal treatment characteristics on treatment outcomes, except patient age. Smoking negatively impacted periodontal treatment outcomes at specific tooth sites (deep pockets, maxillary molars) and periodontitis progression, independent of other risk factors.

Effect of socioeconomic inequalities and contextual factors on induced abortion in Ghana: A Bayesian multilevel analysis.

There is a dearth of information on induced abortion in Ghana, possibly owing to the sensitive nature of the subject. In this study, we examine the effect of socioeconomic and contextual factors on induced abortion in Ghana. This study draws on data from the 2017 Ghana Maternal Health Survey. The study used a Bayesian multilevel logistic regression analysis to estimate both individual- and contextual-level factors affecting induced abortion levels in Ghana. The results show a total induced abortion prevalence of 19.6% coupled with considerable district-level disparities. Induced abortion is significantly associated with socioeconomic factors such as educational attainment, wealth status, and marital status at the individual-level. The risk of induced abortion is considerably higher among the educated, wealthy, and cohabiting women. The current age of women, age at first sex, religious affiliation, parity, and type of residence are the demographic factors having an association with induced abortion levels. At the contextual-level, district health insurance coverage and poverty rate have a significant association with induced abortion. Induced abortion appears to be prevalent in Ghana and is underpinned by both individual-level socioeconomic and aggregate-level factors. Addressing induced abortion levels in Ghana may require policies that take a multilevel approach by focusing on the socioeconomic status of women and district-level contextual factors.

Bayesian Multilevel Analysis of Utilization of Antenatal Care Services in Ethiopia.

In sub-Saharan Africa, 72% of pregnant women received an antenatal care visit at least once in their pregnancy period. Ethiopia has one of the highest rates of maternal mortality in sub-Saharan African countries. So, this high maternal mortality levels remain a major public health problem. According to EDHS, 2016, the antenatal care (ANC), delivery care (DC), and postnatal care (PNC) were 62%, 73%, and 13%, respectively, indicating that ANC is in a low level. The main objective of this study was to examine the factors that affect the utilization of antenatal care services in Ethiopia using Bayesian multilevel logistic regression models. The data used for this study comes from the 2016 Ethiopian Demographic and Health Survey which was conducted by the Central Statistical Agency (CSA). The statistical method of data analysis used for this study is the Bayesian multilevel binary logistic regression model in general and the Bayesian multilevel logistic regression for the random coefficient model in particular. The convergences of parameters are estimated by using Markov chain Monte-Carlo (MCMC) using SPSS and MLwiN software. The descriptive result revealed that out of the 7171 women who are supposed to use ANC services, 2479 (34.6%) women were not receiving ANC services, while 4692 (65.4%) women were receiving ANC services. Moreover, women in the Somali and Afar regions are the least users of ANC. Using the Bayesian multilevel binary logistic regression of random coefficient model factors, place of residence, religion, educational attainment of women, husband educational level, employment status of husband, beat, household wealth index, and birth order were found to be the significant factors for usage of ANC. Regional variation in the usage of ANC was significant.

Deliberate Practice, Functional Performance and Psychological Characteristics in Young Basketball Players: A Bayesian Multilevel Analysis.

Background: Early sport specialization has increased its popularity mostly based on the deliberate practice theory premises. In this study, we examined the influence of the age of onset of deliberate basketball practice on body size, functional performance (countermovement jump, line drill and yo-yo intermittent recovery level 1), motivation for achievement and competitiveness, motivation for deliberate practice and sources of enjoyment among young Brazilian basketball players. In addition, we adjusted for the influence of gender, age group, maturity status and state basketball federation on the outcomes. Methods: The sample included 120 female and 201 male adolescent basketball players aged 14.0 (1.7) years, on average. We grouped players by the age of onset of deliberate basketball practice as related to biologic maturation milestones (pre-puberty deliberate practice onset, mid-puberty deliberate practice onset and late-puberty deliberate practice onset). Results: There was no substantial variation among contrasting players by the onset of deliberate practice in all of the outcomes. Adjusting for gender, male players with late-puberty deliberate practice onset had better functional performance than players with pre- and mid-puberty onset of practice. Females players with late-puberty deliberate practice onset had slightly worst functional performance than players with pre- and mid-puberty onset of practice. Conclusions: Early deliberate basketball practice does not appear to provide an advantage for the development of physiological functions. Likewise, enjoyment, motivation for deliberate practice and motivation for achievement and competition do not appear to be negatively influenced by early deliberate basketball practice. The debate about the relationship between time spent in deliberate practice and performance development in young athletes will need to emphasize the coaching pedagogical quality and the training environment and account for informal practice and deliberate play.

Determinants of multidrug-resistant tuberculosis in São Paulo-Brazil: a multilevel Bayesian analysis of factors associated with individual, community and access to health services.

ObjectiveMultidrug‐resistant tuberculosis (MDR‐TB) remains a serious public health problem worldwide. Accordingly, this study sought to identify individual, community and access to health services risk factors for MDR‐TB.MethodsRetrospective cohort of all TB cases diagnosed between 2006 and 2016 in the state of São Paulo. A Bayesian spatial hierarchical analysis with a multilevel design was carried out.ResultsIt was identified that the history of previous TB treatment (Odds Ratios [OR]:13.86, 95% credibility interval [95% CI]:12.06–15.93), positive sputum culture test (OR: 5.26, 95% CI: 4.44–6.23), diabetes mellitus (OR: 2.34, 95% CI: 1.87–2.91), residing at a standard address (OR: 2.62, 95% CI: 1.91–3.60), positive sputum smear microscopy (OR: 1.74, 95% CI: 1.44–2.12), cavitary pulmonary TB (OR: 1.35, 95% CI: 1.14–1.60) and diagnosis performed due to spontaneous request (OR: 1.26; 95% CI: 1.10–1.46) were associated with MDR‐TB. Furthermore, municipalities that performed HIV tests in less than 42.65% of patients with TB (OR: 1.50, 95% CI: 1.25–1.79), that diagnosed TB cases only after death (OR: 1.50, 95% CI: 1.17–1.93) and that had more than 20.16% of their population with income between ¼ and ½ of one minimum wage (OR: 1.56, 95% CI: 1.30–1.87) were also related to the MDR‐TB.ConclusionsKnowledge of these predictive factors may help to develop more comprehensive disease prevention strategies for MDR‐TB, avoiding the risks expressed regarding drug resistance expansion.

The Influence of Fitness-App Usage on Psychological Well-Being and Body Awareness-A Daily Diary Randomized Trial.

Self-tracking via fitness apps is popular and has been described as a means to enhance body awareness and well-being. However, the effects of fitness-app use and specific app functions on well-being and body awareness have yet to be targeted in controlled experimental studies. In two randomized groups, a fitness tracker was used for 6 weeks, and in one group a daily step target was implemented. In a third control group, participants documented their physical activity. A daily diary method was used to measure well-being and body trusting. In Bayesian multilevel analyses, no time, group, or interaction effects were found. These results were robust when controlling for diverse variables. It can be concluded that exercise-related self-tracking and specific step goals do not substantially influence psychological well-being and body trusting. Considering the large variability in effects, potential effects can be assumed under conditions that are to be identified in further studies.

Sea ice predicts long-term trends in Adélie penguin population growth, but not annual fluctuations: Results from a range-wide multiscale analysis.

Understanding the scales at which environmental variability affects populations is critical for projecting population dynamics and species distributions in rapidly changing environments. Here we used a multilevel Bayesian analysis of range-wide survey data for Adélie penguins to characterize multidecadal and annual effects of sea ice on population growth. We found that mean sea ice concentration at breeding colonies (i.e., "prevailing" environmental conditions) had robust nonlinear effects on multidecadal population trends and explained over 85% of the variance in mean population growth rates among sites. In contrast, despite considerable year-to-year fluctuations in abundance at most breeding colonies, annual sea ice fluctuations often explained less than 10% of the temporal variance in population growth rates. Our study provides an understanding of the spatially and temporally dynamic environmental factors that define the range limits of Adélie penguins, further establishing this iconic marine predator as a true sea ice obligate and providing a firm basis for projection under scenarios of future climate change. Yet, given the weak effects of annual sea ice relative to the large unexplained variance in year-to-year growth rates, the ability to generate useful short-term forecasts of Adélie penguin breeding abundance will be extremely limited. Our approach provides a powerful framework for linking short- and longer term population processes to environmental conditions that can be applied to any species, facilitating a richer understanding of ecological predictability and sensitivity to global change.

Evaluating the influence of data collector training for predictive risk of death models: an observational study

BackgroundSeverity-of-illness scoring systems are widely used for quality assurance and research. Although validated by trained data collectors, there is little data on the accuracy of real-world data collection practices.ObjectiveTo evaluate...

Spatial variations and socioeconomic determinants of modern contraceptive use in Ghana: A Bayesian multilevel analysis.

The literature on spatial patterns and contextual factors affecting modern contraceptive use is not well-documented in Ghana. This study describes the spatial variations and estimates the socioeconomic determinants of modern contraception among women in reproductive age in Ghana. Using data from the 2014 Ghana Demographic and Health Survey, both global and local Moran’s I test were performed to show spatial autocorrelation and Bayesian multilevel models estimated to determine socioeconomic factors affecting modern contraceptive use. The results show evidence of low prevalence and spatial clustering of modern contraception use across the country. There were also notable regional disparities in contraceptive use that favour mostly the southern regions. Modern contraceptive use is significantly associated with socioeconomic factors such as educational attainment, work status, and marital status, as well as age, religious affiliation, and parity. Contextual factors such as the convenient location of health facility and family planning messages exposure also have a considerable positive effect on modern contraceptive use. Uneducated, unemployed and never-married women are considerably disadvantaged in the utilisation of modern contraception in Ghana. Socioeconomic and contextual factors play a key role in modern contraceptive use in Ghana.

Bayesian analysis for social data: A step-by-step protocol and interpretation

The paper proposes Bayesian analysis as an alternative approach for the conventional frequentist approach in analyzing social data. A step-by-step protocol of how to implement Bayesian multilevel model analysis with social data and how to interpret the result is presented. The article used a dataset regarding religious teachings and behaviors of lying and violence as an example. An analysis is performed using R statistical software and a bayesvl R package, which offers a network-structured model construction and visualization power to diagnose and estimate results.•The paper provides guidance for conducting a Bayesian multilevel analysis in social sciences through constructing directed acyclic graphs (DAGs, or "relationship trees") for different models, basic and more complex ones.•The method also illustrates how to visualize Bayesian diagnoses and simulated posterior.•The interpretations of visualized diagnoses and simulated posteriors of Bayesian inference are also discussed.