Statistical tools in published articles of a public health journal in 2013 and 2014: bibliometric cross-sectional study.

Autor:	Arcila Quiceno V; Facultad de Medicina, Universidad CES, Medellin, Colombia. Address: Tr 5 # 41-15, Bogotá, Colombia. Email: montenegro.gino@uces.edu.co., García Restrepo E; Universidad CES, Facultad de Medicina, Medellin, Colombia., Gómez Rúa N; Universidad CES, Facultad de Medicina, Medellin, Colombia., Montenegro Martínez G; Facultad de Medicina, Universidad CES, Medellin, Colombia., Silva Ayçaguer LC; Universidad CES, Facultad de Medicina, Medellin, Colombia; Escuela Nacional de Salud Pública, La Habana, Cuba.
Jazyk:	English; Spanish; Castilian
Zdroj:	Medwave [Medwave] 2015 Aug 31; Vol. 15 (7), pp. e6238. Date of Electronic Publication: 2015 Aug 31.
DOI:	10.5867/medwave.2015.07.6238
Abstrakt:	Introduction: Research projects use statistical resources to express in numerical or graphic terms different magnitudes like frequencies, differences or associations. Objectives: The purpose of this paper is to describe the statistics tools utilization, with special emphasis in the use of conventional statistical tests and confidence intervals, to communicate results in a renowned public health peer reviewed journal in Colombia. Methods: We included the 84 articles published in the journal between 2013 and 2014. Results: The most used resource is frequency analysis (89.3%), followed by p values (65.5%) and confidence intervals (53.6%); 48.9% of the papers used confidence intervals together with p values; 29.8% use neither of them; 16.7% of the articles only used p values and 4.8% only confidence intervals. Conclusions: Descriptive statistics is a tool widely used in research results presentation; the critics and caveats suggesting to avoid the exclusive use of the statistical signification test in the results presentation are not followed in the analysis and presentation of the research results.
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu