Pregled bibliografske jedinice broj: 901768
Effect size and its statistical significance in published papers as metadata in horticultural sciences for application of meta-analysis
Effect size and its statistical significance in published papers as metadata in horticultural sciences for application of meta-analysis // BIOSTAT Book of Abstracts / Jazbec A, Pecina M, Sonicki Z, Šimic D, Vedriš M, (ur.).
Zagreb: Hrvatsko BioMetrijsko društvo, 2015. str. 32-32 (predavanje, međunarodna recenzija, sažetak, znanstveni)
CROSBI ID: 901768 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Effect size and its statistical significance in published papers as metadata in horticultural sciences for application of meta-analysis
Autori
Bosancic Borut, Pecina Marija, Micic Nikola
Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni
Izvornik
BIOSTAT Book of Abstracts
/ Jazbec A, Pecina M, Sonicki Z, Šimic D, Vedriš M, - Zagreb : Hrvatsko BioMetrijsko društvo, 2015, 32-32
Skup
22nd International Scientific Symposium on Biometrics
Mjesto i datum
Dubrovnik, Hrvatska, 28.06.2015. - 01.07.2015
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
biometrics, meta-analysis, horticulture
Sažetak
In order to conduct meta-analysis abundance of quality metadata are required. Meta-analysis in horticultural sciences is novel approach and the standard methodology for finding metadata is to conduct the data search from related research papers through available databases. In the course of planned meta-analysis available databases were searched for metadata on hail protection of apple orchards specifically regarding impact of hail nets on fruit characteristics.It is observed that in horticultural literature sometimes the case that effect sizes are reported without clearly specifying estimates of variations, sample sizes and even without reporting p-values of applied statistical tests. This is especially often in the cases when several groups are compared with post-hoc test hence indicating the grouping pattern with letters rather than actual biometrical test results. Such practice is understandable from the perspective of simplification of data presentation and comprehension of research results. However from the perspective of possible summarizing and reviewing the results from the multiple researches especially for application in meta-analysis a full set of average effect sizes, with related variability estimates and tested values with its exact significance are recommended to be included in publications.
Izvorni jezik
Engleski
Znanstvena područja
Matematika, Biologija, Poljoprivreda (agronomija)
