Issues of the statistical analysis in the Russian ophthalmic journals in 2021 compared to 2015
Year: 2024
Type: article
Abstract: Purpose. To evaluate changes in statistical analysis in Russianophthalmic journals over 6 years — in 2021 compared to 2015. Material and methods. An analysis was carried out of the quality of statistical data processing and the formation of research material in 48 original articlespublished in the first issue of 2021 in 7 leading ophthalmological journals. Literature reviews, descriptions of clinical cases, articles in the sections«To help the practitioner» etc. were not evaluated. A comparison wasmade with the results of a previous analysis of articles published in 2015. Results. By 2021 there has been a positive trend in the use of statisticalmethods in ophthalmology journals. A correctly formatted “StatisticalAnalysis” section was present in articles approximately 1.5 times moreoften. The proportion of publications that did not contain significant errorsin the statistical processing of materials and the formation of study groupsincreased significantly from 21 to 48% (p=0.004). In 2021 somewhat more often than in 2015, there was such a serious error as including both eyesof subjects in the analysis without taking into account the correlation ofthe parameters of the fellow eyes. At the same time, there was a trendtowards a decrease in the frequency of using incorrect methods in multiplecomparisons and when comparing proportions. Lack of indication of thedata format M±σ or M±m, as well as lack of indication of the used versionof the Student’s test or the use of its incorrect version were encounteredin 2021 almost 2 times more often. In 2021 as before, in all cases, therequired size of the group(s) of subjects was not calculated, in more than80% of the articles the sampling order was not indicated; in more thanhalf of the cases, exclusion criteria were not formulated; in some articlesthere was an unreasonably small group size (2–8 subjects). An exampleof a gross error in the organization of research due to the incorrectchoice of normative data in the absence of its own control group is given. Conclusion. Thus, certain positive changes in statistical analysis werenoted in Russian ophthalmic journals over 6 years — in 2021 compared to2015. At the same time, serious errors and violations of both the statisticaldata processing itself and the organization of research and the formationof their material (groups of subjects) are still common. Recommendationsfor reviewing articles are given to eliminate these errors and violations. Key words: statistical analysis, errors, fellow eyes, multiplecomparisons, formation of study groups
Authors: S. Fyodorov Eye Microsurgery Federal State Institution, Moscow, Russian Federation, A.A. Shpak
Language: English
Subfield: History and Philosophy of Science
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W4407795135