IIMYC   23581
INSTITUTO DE INVESTIGACIONES MARINAS Y COSTERAS
Unidad Ejecutora - UE
artículos
Título:
El sesgo del operador en la replicabilidad de los estudios tafonómicos comparativos
Autor/es:
RITTER, MATIAS DO NASCIMENTO; LUZ, NATHÁLIA CARVALHO; PAIM, PROTÁSIO A.V.; KUHN, LIDIA A.; FRANCISCHINI, HEITOR; MORAIS, ANDERSON L.M.; MICHELS, FERNANDO H.; DE FRANCESCO, CLAUDIO GERMÁN; XAVIER, PEDRO L.A.
Revista:
REVISTA BRASILEIRA DE PALEONTOLOGIA
Editorial:
SOC BRASILEIRA PALEONTOLOGIA
Referencias:
Lugar: Porto Alegre; Año: 2016 vol. 19 p. 449 - 464
ISSN:
1519-7530
Resumen:
The operator effect is a well-known analytical bias already quantifiedin some taphonomic studies. However, the influence of operator bias in thereplicability on taphonomic studies has still not been considered. Here, wequantified for the first time this bias using different multivariatestatistical techniques, testing if the operator effect is related to thereplicability. We analyzed the results reported by 15 operators working on thesame dataset. Each operator analyzed 30 bioclasts (bivalve shells) by site,from a total of five sites, considering the following taphonomic attributes:shell fragmentation, edge rounding, corrasion, bioerosion, and coloralteration. The operator effect followed the same pattern reported in previousstudies, characterized by a worse correspondence for those attributes havingmore than two levels of damage categories. However, the effect did not appearto have relation to replicability, because nearly all operators founddifferences among sites. Despite the binary attribute, bioerosion exhibited 83%of correspondence among operators it was the taphonomic attribute that showedthe highest dispersion among operators (28%). Therefore, we concluded thatbinary attributes despite indicating a reduction of the operator effect diminishreplicability, resulting in different interpretations of concordant data. Wefound that a variance value of nearly 8% among operators, was enough togenerate a different taphonomic interpretation, in a Q-mode cluster analysis.The results reported here showed that the statistical method employedinfluences the level of replicability and comparability of a study and that theavailability of results may be a valid alternative to reduce bias.