Medindo Valores Com Parcelas De Itens: Contribuições à Teoria Funcionalista dos Valores

• Ano de defesa: 2012
• Autor(a) principal: Souza, Luana Elayne Cunha de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraí­ba; BR; Psicologia Social; Programa de Pós-Graduação em Psicologia Social; UFPB
• Palavras-chave em Português: Medida; Valores humanos; Parcela de itens; Validade; Measure; Human values; Item parcels; Validity; CIENCIAS HUMANAS::PSICOLOGIA
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/tede/6904

ABSTRACT - The current thesis aimed to create a new instrument for measurement of human values, based on the functionalist theory of human values, for diagnosis and screening. To this end, it took out three studies. Study 1 aimed to build up a database of items that would meet an essential criterion: to represent the subfunctions recommended by the functionalist theory. Initially, this scale was comprised by 180 items, 30 items representing each of the six subfunctions proposed. The study included 10 judges, who with a minimum of 70% agreement between them, judged the relevance of 127 items, 53 items were eliminated. Study 2 aimed to build an instrument, resulting from the items in the previous study and gather evidence of its psychometric properties. Participants were 210 students with ages ranging from 14 to 28 years (m = 18.28; sd = 2.94), mostly female (61,4%). They answered the set of 127 items on a 7-point scale ranging from 1 (totally unimportant) to 7 (extremely important), and demographic questions. Initially, there was a semantic validation of the instrument with the participation of 12 students the first year of high school. Found that there were no questions, remained the proposed version. Confirmatory factor analysis were performed considering the covariance matrix and adopting the maximum likelihood estimator for the set of items in each subfunction separately. Through the factor loadings, we selected the top three items in each human value, resulting in a structure of nine items per subfunction of values. The rates of Cronbach s alpha ranged from 0.44 (maturity) to a 0.94 (religion), the homogeneity s indices ranged from 0.21 (maturity) to 0.60 (sexuality), and composite reliability was 0.61 (pleasure) to 0.92 (religiosity). These indices gather evidence of the reliability of the instrument. The average variance extracted, which was 0.46 (maturity) to 0.91 (religiosity) and comparison of the square root of the average variance extracted with the correlations among the factors, in this case the subfunctions values, demonstrating evidence of validity convergent and discriminant, respectively, of the measure. Study 3 aimed to confirm the factor structure using item parcels technique and, again, present their psychometric parameters. Participants were 231 students with ages ranging from 16 to 34 years (m = 19.98; sd = 2.77), mostly female (61.9%). They answered the set of items resulting from the second study in the same response scale as well as demographic questions. By comparing the rates of adjusting the model item parcels from the model built with the original items, was observed that the first best fits the data. The alpha values for the factor structure proposed ranged from 0.61 (normative) to 0.80 (interactive), regarding to homogeneity s index, composite reliability and average variance extracted, all the subfunctions above the recommended values presented in the literature. Finally, all subfunctions showed discriminant validity, except the relationship between existence and achievement and between existence and suprapersonal. In conclusion, trust that the goals proposed in this thesis have been achieved, this time presenting a contribution to a field of the measurement of human values.