Resumen de Ensino de Análise Funcional Baseada em Tentativas: Vídeo-modelo versus vídeo-feedback

Gleice de Oliveira Cordeiro, Isadora Peresi Ferrari, Christian Vichi, Bruna Vilas Boas Rios Costa

• English

  Functional Behavior Assessment (FBA) seeks to analyze how environmental factors interfere in the occurrence and maintenance of behavioral patterns of an individual. Given the complexity of conducting this assessment, some adaptations were made so that professionals with no prior knowledge in Behavior Analysis could investigate behavioral functions in natural contexts, by experimentally arranging antecedent and consequent contexts. It is in this scenario that the Trial-based Functional Analysis (TBFA) is designed. The literature has been investigating in which conditions teachers could be taught how to conduct TBFA in a variety of scenarios, including classroom settings. This study aimed to evaluate the use of Video-feedback (VF) and Video-modeling (VM) to train teachers to implement the TBFA with at least 80% procedural fidelity. Participants were six Brazilian teachers working at the Specialized Educational Attendance (AEE) from a public school in Pernambuco State, Brazil. Each of them conducted TBFA with two students with disabilities. They were referred for problem behavior such as self-harm, aggression, destruction of proper materials, and escape from classroom activities. The procedures took place in seven phases: Initial Evaluation, Baseline (BL), Training with VF or VM, Posttraining Probe, Retraining, Post-Retraining Probe and Generalization Probe. The experimental conditions evaluated were Attention, Tangible, Demand and Ignore. Interobserver agreement was 85.75%. The verification of training effectiveness was evaluated by calculating the percentage of procedural fidelity during implementation, and then, by comparing the BL with the Posttraining and Retraining Probes. It was pointed out that although five teachers did not reach the post-training learning criterion defined in this project, there was an improvement in the TBFA repertoire, especially when comparing the general averages in BL (15%) and (17%) with Post-training Probes (70%) and (61%), respectively for those trained with VF and VM. It has been found an average fidelity of 60% in Retraining Probes, regardless of training configuration. At the Generalization Probes, the average performance of the application of TBFA was maintained (75%), this time with a second student, weeks after the training. Comparing to Video-modeling (VM) results, VF had the best cost-benefit ratio by ensuring faster learning, with fewer training exposures to ensure performance maintenance throughout the Probes. TBFA training was described, procedural limitations and indications for future research were also discussed. This research has expanded the field of TBFA by teaching Brazilian Special Education teachers on how to conduct such procedures with high procedural fidelity. It has described the effects of the Video-feedback and Video-modeling component trainings, commonly used in training packages, over teachers’ abilities to conduct TBFA. It has also presented the limitations of the VF and VM by analyzing teachers’ errors in each training condition.

• português

  Foram avaliados os efeitos de Vídeo-feedback (VF) e Vídeo-modelo (VM), de forma isolada e combinada, para o ensino de Análise Funcional Baseada em Tentativas (TBFA) para seis professores de Atendimento Educacional Especializado (AEE), divididas em dois grupos. Após treinamento, cada uma delas avaliou dois estudantes público-alvo da educação especializada com problemas de comportamento. Os procedimentos de treinamento incluíram Linha de Base (LB), Treinos iniciando com VF ou VM, Sondas Pós-treino (com aluno A) e Sondas de Generalização (com aluno B). As condições experimentais treinadas foram Atenção, Tangível, Demanda e Ignorar. Os dados de LB foram comparados com as médias de desempenho nos Pós-treinos e observou-se que, embora ambos os grupos das professoras não tenham atingido o critério de aprendizagem (>80%), apresentaram melhora significativa na condução do TBFA, sendo LB: 15% e 17%, Pós-treino: 70% e 61%, para VF e VM, respectivamente. Observou-se também a manutenção do desempenho de aplicação do TBFA (VF =73% e VM=77%) nas sondas de generalização, aplicadas dois meses após o treino. Também foram discutidos tipos de erros e frequências em cada condição para cada professora e por grupo segundo estratégia de treinamento, limitações dos treinos e do procedimento do TBFA, e indicações de pesquisas futuras.