Research papers
← vista completaPublished on September 1, 2011 | http://doi.org/10.5867/medwave.2011.09.5133
Small study groups: a search for evidence
Grupos de estudio pequeños: una búsqueda de evidencias
Abstract
Introduction: With the aim of seeking for empiric evidence as to support the hypothesis that small-size studies analyzed with no explicit operational hypothesis may lead to erroneous conclusions, we carried out an analysis of subgroups resulting from the breakdown of a previous published large study and using on them the same analytical procedures.
Methods: The 584 patients from a sample as calculated upon an operational hypothesis whose test results are already published, were grouped in six subgroups 100 individuals each, on whom the same analytical procedures were performed using the same statistical methods.
Results: The outcomes showed that half the groups had association statistics in one direction (above one) whereas the other half had it on the opposite (below one). Only one subgroup reached statistical significance (p <0.05).
Comments: These results illustrate with empirical evidence that small- size study groups analyzed with no explicit hypothesis yield inconsistent results, this way supporting the hypothesis that these small-size groups yields do not have the meaning of complete analytical ones and hence, their eventual importance lies only on descriptive grounds.
