Outliers in Price’s Law

• Autor(a) principal: Silva, Dávilla Vieira Odízio da
• Data de Publicação: 2014
• Outros Autores: Maroldi, Alexandre Masson, Lima, Luís Fernando Maia
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Em Questão (Online)
• Texto Completo: https://seer.ufrgs.br/index.php/EmQuestao/article/view/49088
• topic: Outlier. Exploratory Data Analysis. Price’s Law. Bibliometrics. Statistics.; Outlier; Análise Exploratória de Dados; Lei do Elitismo; Bibliometria; Estatística; Outlier

Looking at the most productive subset of authors as outliers, this study proposes the use of Exploratory Data Analysis (EDA) to detect outliers, and thus the productive elite. Price’s Law or the square root criteria proposed by Price has two major problems: first, the square root of the total number of authors is not always an integer value, creating the need to employ truncate or rounded values. Second, the cutoff point is not easy to select, since the theoretical value rarely fits the observed data. Besides, the detection of outliers from EDA finds a unique way to show the elite. In the end of the study we conclude that outliers may give non integer values for author production, but offers only one value to identify the elite set in a given area of knowledge.