Search

There are major challenges in olfactory measurements in clinical practice; therefore, a handheld digital scent device (DSD; Noar MultiScent 20) was developed as a tablet with an integrated storage system for odors. The DSD is a self-administered, handheld device that controls the duration of odor release to the nasal cavity through a touchscreen digital interface with automatic database generation. In this study we aimed to determine the feasibility of this DSD as an olfactory assessment test.We recruited 180 participants (age [mean ± standard deviation], 34.58 ± 9.71 years; 114 women and 66 men) to participate in smell tests using both the DSD and the 40-item Smell Identification Test (SIT-40), which contained the same type and order of odors and the same multiple-choice answers. The scores were compared and evaluated for correlation between the tests, and test-retest reliability was calculated.The DSD test scores were higher than the SIT-40 scores (median [interquartile range], 32 [5.0] vs 31 [7.0]; p = 0.005). The completion time was less for the DSD test than for the SIT-40 (12.5 [5.0] vs 16 [6.0] minutes; p 0.001). The tests were strongly correlated (Spearman rho = 0.74; p 0.001) and exhibited a high level of agreement (Bland-Altman regression coefficient = 0.672; p = 0.003). The DSD test-retest reliability coefficient was 0.820.The DSD is feasible as an olfactory assessment test. The digitalization of olfactory assessment combined with data science may enable new research perspectives in the field of olfaction.

Peers are the preferred source of help for Brazilian adolescents who experience dating violence. However, they are not always the best informants for effective responses for dealing with situations of violence in romantic relationships among peers. This experimental study aimed to evaluate the short-term efficacy of three aspects of a peer- and bystander approach-based intervention: the intent to offer help, empathy, and bystander attitudes in response to dating violence in a Brazilian sample of adolescents. The study’s participants were 33 adolescents randomized in two groups: experimental group (EG,n = 14) and control group (CG,n = 19). The EG underwent three weekly intervention sessions of 90 min each on the healthy versus violent romantic relationships, the quality of friendship in the peer network, and the role of the bystander, while the CG received no intervention. Evaluations were performed 1 week before and two and half months after the intervention. Statistically significant differences between EG and CG at post-test were not found in intention to help, empathy, and bystander attitudes. Future studies should include evaluations of larger samples and mid- and long-term follow-ups to identify patterns of change over the long term as well as examine barriers to utilization of bystander behaviors by adolescents in Brazilian culture.

A scoping review, based on the RE-AIM framework, was conducted to analyze evidence of reach, effectiveness, adoption, implementation, and maintenance of the Strengthening Families Program (10-14), a preventive family-based substance abuse program for adolescents. Sixty-five articles were included. The results disclosed that effectiveness, implementation, and maintenance at the individual-level were the most evaluated aspects, while reach, maintenance at the setting-level, and adoption were the least investigated aspects. Positive effects on drug abuse prevention and protective parenting factors were found in the U.S. studies. Likewise, Latin American studies have shown the improvement of parenting practices. However, European studies have produced mixed results, with predominantly null effects on substance abuse. The implementation quality was high. There is no available evidence of adoption and maintenance at the setting-level by the organizations that implemented it. New studies must examine the reach, adoption, and sustainability of the program to lay foundations for its future use as an instrument of public policies. Supplementary Information The online version contains supplementary material available at 10.1186/s41155-021-00182-z.

Adolescents living in low- and middle-income countries (LMICs) are not only exposed to underprivileged conditions - they are also at greater risk of being raised with negative parenting practices that may have detrimental effects on their psychosocial development and health. Thus, the implementation and evaluation of family-based interventions that foster positive parenting practices among vulnerable populations are necessary. The aim of this study is to examine if the Brazilian Strengthening Families Program (SFP) has a differential impact on parenting styles in the short, medium, and long terms. This was achieved through a single-group longitudinal design with four data collection moments. Parenting styles were assessed using the Parental Practices Scale at baseline and then after delivering the program, with follow-ups after 6/8 and 10/12 months. The sample comprised 361 adolescents aged 10–14 years from low-income families in northeastern Brazil. A combination of hierarchical and non-hierarchical cluster-analysis methods offered a four-cluster solution representing the four parenting styles: authoritative, authoritarian, indulgent, and negligent. Mixed ANOVAs revealed a main effect of time on both responsiveness and demandingness, indicating that overall levels were higher post-test than pre-test. There were improvements in the responsiveness levels among authoritarian parents, as well as improvements in the demandingness levels among indulgent parents. The Brazilian SFP (10–14) appears promising for improving parenting styles, which it does by strengthening the weakest parental aspect to enhance positive parenting in families living in underprivileged conditions.

In 2013, Brazil’s Ministry of Health adopted the Strengthening Families Program (SFP 10–14), developed internationally for preventing drug abuse by enhancing family bonds. The social validity of the objectives, procedures, and perceived impacts of the program were investigated for participants and facilitators in northeastern Brazil. Focus groups with parents/guardians (N = 199), adolescents (N = 111), and facilitators (N = 100) were implemented. Content analysis revealed that the program’s objectives were considered socially relevant and that there was a positive short-term perceived impact on family cohesion, authoritative parenting style, adolescent life skills, and the facilitators’ professional capacity. The parents/guardians and adolescents presented a positive perception of the appropriateness of the program’s methodology, while facilitators indicated the need to adapt it to vulnerable families and improve its implementation conditions. Future studies may benefit from these findings when developing similarly viable and scalable interventions in low-resource settings. Brazilian Trial Register RBR-7q9xh5. Registered 5 August 2017, http://www.ensaiosclinicos.gov.br/rg/RBR-7q9xh5/

Os dados ausentes são comuns nas avaliações educacionais. Por isso, o uso de métodos adequados torna-se fundamental para reduzir o impacto da perda de informação. O objetivo deste estudo é comparar o desempenho de quatro métodos de tratamentos de dados ausentes (imputação pela média, listwise deletion, máxima verossimilhança e imputação múltipla), tendo como base o uso de modelos de regressão aplicados aos dados da avaliação educacional realizada no estado do Ceará. Foram utilizadas informações de 7.000 estudantes, simulando-se diversos cenários de acordo com o percentual e o tipo de ausência. A imputação pela média apresentou o pior desempenho em todos os cenários simulados e os demais métodos mostraram resultados semelhantes entre si. Verificou-se ainda que o uso de variáveis auxiliares na estimação por máxima verossimilhança e imputação múltipla reduziu o viés das estimativas de parâmetros importantes do modelo quando a ausência simulada não é ao acaso.Palavras-chave: Tratamento de Dados Ausentes; Avaliação da Educação; Desempenho Acadêmico; Simulação. Datos ausentes en evaluaciones educacionales: comparación de métodos de tratamientoLos datos ausentes son comunes en las evaluaciones educacionales. Por ello el uso de métodos adecuados se hace fundamental para reducir el impacto de la pérdida de información. El objetivo de este estudio es comparar el desempeño de cuatro métodos de tratamientos de datos ausentes (imputación por el promedio, listwise deletion, máxima verosimilitud e imputación múltiple), en base al uso de modelos de regresión aplicados a los datos de la evaluación educacional realizada en el estado de Ceará. Se utilizaron informaciones de 7.000 estudiantes, simulando diversos escenarios de acuerdo al porcentual y al tipo de ausencia. La imputación por el promedio presentó el peor desempeño en todos los escenarios simulados y los demás métodos mostraron resultados similares entre ellos. También se verificó que el uso de variables auxiliares en la estimación por máxima verosimilitud e imputación múltiple redujo el sesgo de las estimaciones de parámetros importantes del modelo cuando la ausencia simulada no se debe a la casualidad.Palabras clave: Tratamiento de Datos Ausentes; Evaluación de la Educación; Desempeño Académico; Simulación. Missing data in educational assessment: a comparison of data treatment methodsMissing data are common in educational assessments. Using the appropriate methods has, therefore, become essential to reduce the impact of the loss of information. The present study aims to compare the performance of four methods for dealing with missing data (mean imputation, listwise deletion, maximum likelihood and multiple imputation), all based on regression models applied to the educational assessment of data collected in the State of Ceará. Information about 7,000 students was used, simulating various scenarios according to the percentage and the type of the missing data. The mean imputation method showed the worst performance in all simulated scenarios and the other methods showed similar results among themselves. Moreover, the use of auxiliary variables in the estimation by maximum likelihood and multiple imputation proved to reduce the bias of estimates of some important parameters of the model, when the simulated missing data is not random.Keywords: Treatment of Missing Data; Education Assessment; Academic Performance; Simulation.

Substance abuse and violence are among the primary health concerns regarding Brazilian adolescents. This study sought to explore the short-term effects of the Strengthening Families Program (SFP 10-14), a preventive program for families with adolescents, adapted to Brazil. A pre-experimental design was used, with a pretest and 10-12-month follow-up evaluation. A qualitative study was carried out using in-depth interviews held one to three months after the intervention to examine the use of skills learned. The sample included 126 adolescents (pre-test and follow-up comparison) and 23 adolescents (interviews) between 10 and 14 years of age from low-income families residing in northeastern Brazil. The comparison between pretest and follow-up showed an increase in learning self-efficacy and school absence without parental permission. Null effects were found on the consumption of alcohol in the last month; episodes of binge drinking in the last month; antisocial behavior; parenting practices regarding emotional support factors, intrusiveness, and behavior supervision; future time perspective; doing homework; grade repetition; school grades; school dropout; and satisfaction with one's relationship with school. The majority of the interviewed adolescents reported applying the learned skills during family interaction and with friends. Future studies should examine the contexts and mechanisms linked to such mixed results.

Tese (doutorado)—Universidade de Brasília, Instituto de Psicologia, Departamento de Psicologia Social e do Trabalho, Programa de Pós-Graduação em Psicologia Social, do Trabalho e das Organizações, 2016. O objetivo desta tese foi contribuir para o desenvolvimento da avaliação educacional por meio de: (a) estudo de simulação visando a avaliar o desempenho de alguns métodos de tratamento e o efeito dos dados ausentes nos resultados, (b) comparação de técnicas estatísticas avançadas usadas para tratamento de dados ausentes nos estudos longitudinais, (c) apresentação de um novo método para tratamento de dados ausentes e (d) aplicação do novo método na identificação de fatores associados ao desempenho escolar, tendo como base os dados de uma avaliação longitudinal. Para isso, foram utilizados os dados da avaliação educacional realizada no estado do Ceará. A tese está dividida em três manuscritos. O primeiro apresenta uma introdução à teoria relacionada aos dados ausentes, as metodologias geralmente utilizadas pelos pesquisadores e os possíveis impactos desses dados nos resultados das pesquisas. Por meio de um estudo de simulação, quatro métodos de tratamentos de dados ausentes (imputação pela média, listwise deletion, máxima verossimilhança e imputação múltipla) foram comparados. A imputação pela média apresentou o pior desempenho em todos os cenários e os demais métodos apresentaram resultados semelhantes. Um outro resultado do estudo de simulação foi que o uso de variáveis auxiliares na estimação por máxima verossimilhança e na imputação múltipla reduziu o viés das estimativas quando a ausência simulada não é ao acaso. O segundo manuscrito discute a classificação proposta por Rubin com ênfase na ausência de dados em estudos longitudinais. Esse manuscrito apresenta uma nova metodologia para o tratamento de dados ausentes não ao acaso (MNAR) no contexto de avaliações educacionais. Um estudo de simulação comparou os procedimentos listwise deletion, imputação múltipla e a metodologia proposta. Tendo como base o modelo de crescimento linear, verificou-se que o procedimento listwise deletion superestimou a taxa média de aprendizado. A imputação múltipla e a metodologia proposta geraram maiores estimativas para os coeficientes das variáveis independentes, e ainda identificaram efeitos de interação. Os resultados evidenciaram a importância da escolha da abordagem a ser utilizada no tratamento de dados faltantes. No terceiro manuscrito, a metodologia proposta para tratamento de dados ausentes foi utilizada no estudo de fatores associados ao desempenho escolar. Em uma amostra composta por 8.681 estudantes do ensino médio, 25,7% estava ausente em pelo menos um momento da avaliação. Verificou-se que a ausência estava relacionada às características dos estudantes e ao desempenho escolar avaliado. A taxa média de aprendizado estimada foi de 8,96 pontos, mas essa taxa varia significativamente entre os estudantes. Com a utilização de dados longitudinais e técnicas de tratamento de dados ausentes, os resultados corroboram estudos transversais de fatores associados ao desempenho escolar. Além disso, demonstra que variáveis relacionadas à idade, número de reprovações e período noturno têm efeitos negativos tanto na proficiência inicial, quanto na taxa de aprendizado. The main objective of this thesis was to contribute to the development of educational assessment through: (a) simulation study to assess the performance of some treatment methods and the impact of missing data on results, (b) comparison of advanced statistical techniques for missing data treatment in longitudinal studies, (c) introducing a new method for missing data treatment and (d) application of the new method to identify factors associated with academic performance, based on a longitudinal assessment data. For this purpose, the data of an educational assessment carried out in Ceará State was used. The thesis is divided into three manuscripts. The first manuscript presents an introduction of missing data theory, methodologies generally used by researchers and the potential impacts of such data in research results. A simulation study was used to compare four missing data treatments (mean imputation, listwise deletion, maximum likelihood and multiple imputation). Mean imputation had the worst performance in all scenarios while the other three methods showed similar results. Additionally, the use of auxiliary variables with maximum likelihood estimation and multiple imputation reduced estimation bias when the simulated missingness is not at random. The second manuscript presents the classification proposed by Rubin emphasizing missing data in longitudinal studies. This manuscript proposes a new methodology for the treatment of missing not at random data in the educational assessment context. Listwise deletion, multiple imputation and the proposed methodology were compared in a simulation study. The linear growth model was used for data analysis and comparisons. The average learning rate in Mathematics was overestimated when listwise deletion was used. Multiple imputation and the proposed methodology estimated higher impacts of the independent variables and identified interaction effects. The results highlighted the importance of the method that is chosen to deal with missing data, which is directly related to the assumptions about missing data generating mechanism. In the third manuscript, the proposed methodology for missing data treatment was used to identify factors associated with academic performance. The sample was composed by 8,681 high school students, 25.7% of them were absent at least one moment of follow-up. It was found that the missingness was related to students’ characteristics and school performance. The linear growth model showed that the annual learning rate was 8.96 points on average, however it varies significantly among students. Using longitudinal data and missing data treatment techniques, the results corroborate those from cross-sectional studies of factors associated with school performance. Moreover, it shows that variables related to age, school repetition and evening classes have negative effects on both initial proficiency and learning rate.

Several studies have indicated that student performance is strongly influenced by school context. Within this frame of reference, the present study sought to identify contextual variables that affect proficiency in mathematics taking the multilevel structure of schools into consideration. The information of 4,338,885 students who participated in the Brazilian national assessment of educational progress of 2009 was used. Multilevel analysis showed an intra-class correlation of .27 and .19 for the 5th grade of elementary school and 9th grade of middle school, the last year of middle school. In both models, more than 50% of school level variance and more than 11% of the variance on student level was explained. The final model is composed of control variables, 13 student level variables, 9 school level variables, and random and interaction effects. The results contribute to a better comprehension of factors that affect school performance, which permits the orientation of public policies. Resumo Vários estudos têm apontado que o aprendizado do aluno é fortemente influenciado pelo contexto escolar. Dessa forma, este estudo teve como objetivo identificar variáveis contextuais que interferem no desempenho em matemática, considerando a estrutura multinível da conjuntura escolar. Utilizaram-se informações de 4.338.885 alunos avaliados na Prova Brasil de 2009. A análise multinível mostrou uma correlação intraclasse de 0,27 e 0,19, para o 5° e 9° ano do ensino fundamental, respectivamente. No modelo final do 5° e 9° ano, mais de 50% da variância no nível da escola e mais de 11% da variância no nível do aluno foram explicadas. Esse modelo foi composto por variáveis de controle, 13 variáveis do nível do aluno, 9 do nível da escola, efeitos randômicos e efeitos de interação. Os resultados corroboram estudos anteriores e contribuem com o entendimento dos fatores que influenciam o desempenho, o que permite direcionar políticas públicas. Resumen Diversos estudios señalan que el aprendizaje del alumno está fuertemente influenciado por el contexto escolar. De esta forma, la presente investigación tuvo como objetivo identificar variables contextuales que interfieren en el desempeño de matemáticas, teniendo en cuenta la estructura de varios niveles del entorno escolar. Se utilizó la información de 4.338.885 alumnos evaluados en la Prueba Brasil de 2009. El análisis de varios niveles mostró una correlación intraclase de 0,27 y 0,19 para el 5º y 9º año de enseñanza primaria, respectivamente. En el modelo final de 5º y 9º año, más del 50% de la variancia a nivel de escuela y más de 11% de variancia a nivel de alumno fue explicada. El modelo fue compuesto por variables de control, 13 variables del nivel del alumno, 9 del nivel de la escuela, efectos aleatorios y efectos de interacción. Los resultados corroboran estudios anteriores y contribuyen a la compresión de los factores que influyen en el desempeño, lo que permite direccionar las políticas públicas.

La disponibilidad de información relacionada con evaluaciones educacionales es grande en América Latina, principalmente después de las reformas educativas de las últimas décadas. El objetivo de este estudio es trazar un mapa de cómo los datos de las evaluaciones educativas realizadas en Brasil, Chile y Argentina son utilizados por los investigadores. La metodología empleada para la identificación y selección de los estudios fue la búsqueda de publicaciones, entre 1994 y 2014, por medio de la base de datos de la agencia CAPES. Utilizando las palabras clave: “rendimiento escolar” (en español, portugués – rendimento escolar – e inglés – school performance), SAEB, SIMCE y “Operativo Nacional de Evaluación”; se seleccionaron 27 estudios empíricos que utilizan datos de las evaluaciones de los tres países en cuestión. La revisión mostró que la técnica más utilizada fue el análisis de regresión multinivel. Se verificó que casi todos los trabajos seleccionados analizaban el impacto de factores asociados al desempeño escolar. Sin embargo, también se utilizaron los datos de las evaluaciones educativas para evaluar el papel de la escuela en la promoción de la equidad y estimar el impacto de programas educativos. Los resultados obtenidos en este estudio permiten concluir que, en general, los artículos revisados proporcionan información que puede apoyar políticas que contribuyan a la mejora de la educación, A great amount of information related to the assessment of educational progress is available in Latin America, especially after the implementation of educational reforms during the last decades. This article aims to map how researchers use data of educational assessments conducted in Brazil, Chile and Argentina. The methodology used for identifying and selecting studies was a search for publications through the CAPES database, between 1994 and 2014. Using the keywords: "school performance", “rendimiento escolar” (in Spanish) and “rendimento escolar” (in Portuguese), SAEB, SIMCE and "National Operational Assessment"; 27 empirical studies using assessment data related to the three countries concerned were selected. The most frequently used technique in these articles was multilevel regression analysis. The review indicated that practically all of the selected articles analyzed the impact of factors associated to school performance. However, educational assessment data were also used to evaluate the role of schools in promoting equity and to estimate the impact of educational programs. The results of the present study allow the general conclusion that the revised articles supply information that can support policies that contribute the improvement of education