Planejamento experimental para mudas de couve-flor e plantas de alface: definição do tamanho amostral ótimo

• Ano de defesa: 2023
• Autor(a) principal: Bittencourt, Karina Chertok
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; Agronomia; UFSM; Programa de Pós-Graduação em Agronomia - Agricultura e Ambiente; UFSM Frederico Westphalen
• Palavras-chave em Português: Brassica oleracea; Lactuca sativa; Estatísticas de precisão; Horticultura; Reamostragem bootstrap; Precision statistics; Bootstrap resampling; CNPQ::CIENCIAS AGRARIAS::AGRONOMIA
• Link de acesso: http://repositorio.ufsm.br/handle/1/32039

Defining the sample size is an important step in the planning of experiments, as collecting a sufficiently representative sample is essential to obtain reliable results. However, the optimal sample size may vary depending on the species, evaluated characters, and the subsequently estimated statistics. In this sense, research on economically important vegetable crops, such as cauliflower and lettuce, has commonly used small sample sizes, given the scarcity of studies focusing on their sample dimensioning. Therefore, the present work aimed to optimize the experimental planning of experiments with cauliflower and lettuce crops through sample sizing for different statistics and characters. Thus, a greenhouse experiment with cauliflower seedlings was conducted in the experimental area of the Federal University of Pampa, Itaqui Campus, and a field experiment with 26 lettuce genotypes was conducted in the experimental area of the Federal University of Santa Maria, Frederico Westphalen Campus. For cauliflower seedlings, the following characters were assessed: number of leaves, plant height, root length, and total length (plant height + root length), and for lettuce plants, the yield per plant (fresh weight in grams), number of leaves, plant height, stem diameter, and mean head diameter were evaluated. Precision statistics were estimated, obtaining the 95% confidence interval width. One hundred sampling scenarios were simulated for each statistic and character using bootstrap resampling with replacement, and optimal sample sizes were defined by adjusting the 95% confidence intervals to models of the power family and finding the maximum curvature point. Furthermore, four methods for obtaining the maximum curvature point were compared, and predictive equations for precision statistics based on sample size were proposed. The 95% confidence interval width of the statistics reduced as the sample size increased, until a point of stabilization. The perpendicular distance method was considered the most efficient for defining the maximum curvature point. The sample sizes varied according to statistics and characters, with this variation being greater between statistics. The F statistic stood out for obtaining larger sample sizes in all studies. The predictive equations presented excellent fitting quality, which allows for knowing the mean, maximum, and minimum values of precision statistics based on the selection of a specific sample size. Thus, the information provided by these studies contributes to optimizing the experimental planning of cauliflower and lettuce crops and may be useful for researchers in the field who wish to evaluate experimental precision through the statistics and characters described.