Different statistical analyses may lead to categorically distinct conclusions

Detalhes bibliográficos
Autor(a) principal: Lima Jr, Ronaldo Mangueira
Data de Publicação: 2021
Outros Autores: Garcia, Guilherme Duarte
Tipo de documento: Artigo
Idioma: por
Título da fonte: Revista da ABRALIN (Online)
Texto Completo: https://revista.abralin.org/index.php/abralin/article/view/1790
Resumo: In this study, we illustrate the potential variability of statistical significance by comparing four different methods, namely, t-test, ANOVA (followed by Tukey HSD), simple linear regression, and mixed effects linear regression. In our demonstration, we model reaction times as a function of different affixes in Danish, and show how our conclusions regarding the effect of certain affixes can change categorically depending on which of the aforementioned methods we choose to use. Finally, we echo recent studies (e.g., BARR et al., 2013), and suggest that mixed effects models be the norm whenever grouped data is analyzed. With our comparison, we hope to raise researchers’ awareness to the need for well-informed and ethical analytical decisions in linguistic studies.
id UFPR-12_40e195d34c8f6f18f4b7b88998c179f1
oai_identifier_str oai:ojs.revista.ojs.abralin.org:article/1790
network_acronym_str UFPR-12
network_name_str Revista da ABRALIN (Online)
repository_id_str
spelling Different statistical analyses may lead to categorically distinct conclusionsDiferentes análises estatísticas podem levar a conclusões categoricamente distintasAnálise quantitativa de dadosModelos de regressãoTestes estatísticosSignificância estatísticaQuantitative data analysisRegression modelsStatistical testsStatistical significanceIn this study, we illustrate the potential variability of statistical significance by comparing four different methods, namely, t-test, ANOVA (followed by Tukey HSD), simple linear regression, and mixed effects linear regression. In our demonstration, we model reaction times as a function of different affixes in Danish, and show how our conclusions regarding the effect of certain affixes can change categorically depending on which of the aforementioned methods we choose to use. Finally, we echo recent studies (e.g., BARR et al., 2013), and suggest that mixed effects models be the norm whenever grouped data is analyzed. With our comparison, we hope to raise researchers’ awareness to the need for well-informed and ethical analytical decisions in linguistic studies.Neste estudo, demonstramos como significância estatística pode variar a partir da comparação de quatro métodos distintos: teste t, ANOVA (seguida de Tukey HSD), modelo linear simples, e modelo linear de efeitos mistos. Em nossa demonstração, modelamos tempos de reação em função de diferentes afixos em dinamarquês, e mostramos como nossas conclusões a respeito do efeito de certos afixos podem mudar categoricamente dependendo de qual dos métodos mencionados acima decidimos utilizar. Por fim, reiteramos o que dizem estudos recentes (e.g., BARR et al., 2013), e sugerimos que modelos de efeitos mistos devam ser a norma sempre que dados agrupados forem analisados. Esperamos, com este estudo, alertar pesquisadores da área para a importância de decisões analíticas bem informadas e éticas em estudos linguísticos.Associação Brasileira de Linguística2021-08-05info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionTutorialTutorialapplication/pdftext/xmlhttps://revista.abralin.org/index.php/abralin/article/view/179010.25189/rabralin.v20i1.1790Revista da ABRALIN; V. 20, N. 1 (2021); 1-19Revista da ABRALIN; V. 20, N. 1 (2021); 1-190102-715810.25189/rabralin.v20i1reponame:Revista da ABRALIN (Online)instname:Universidade Federal do Paraná (UFPR)instacron:UFPRporhttps://revista.abralin.org/index.php/abralin/article/view/1790/2316https://revista.abralin.org/index.php/abralin/article/view/1790/2383Copyright (c) 2021 Ronaldo Mangueira Lima Jr, Guilherme Duarte Garciainfo:eu-repo/semantics/openAccessLima Jr, Ronaldo MangueiraGarcia, Guilherme Duarte 2021-08-05T18:00:53Zoai:ojs.revista.ojs.abralin.org:article/1790Revistahttps://revista.abralin.org/index.php/abralinPUBhttps://revista.abralin.org/index.php/abralin/oairkofreitag@uol.com.br || ra@abralin.org2178-76031678-1805opendoar:2021-08-05T18:00:53Revista da ABRALIN (Online) - Universidade Federal do Paraná (UFPR)false
dc.title.none.fl_str_mv Different statistical analyses may lead to categorically distinct conclusions
Diferentes análises estatísticas podem levar a conclusões categoricamente distintas
title Different statistical analyses may lead to categorically distinct conclusions
spellingShingle Different statistical analyses may lead to categorically distinct conclusions
Lima Jr, Ronaldo Mangueira
Análise quantitativa de dados
Modelos de regressão
Testes estatísticos
Significância estatística
Quantitative data analysis
Regression models
Statistical tests
Statistical significance
title_short Different statistical analyses may lead to categorically distinct conclusions
title_full Different statistical analyses may lead to categorically distinct conclusions
title_fullStr Different statistical analyses may lead to categorically distinct conclusions
title_full_unstemmed Different statistical analyses may lead to categorically distinct conclusions
title_sort Different statistical analyses may lead to categorically distinct conclusions
author Lima Jr, Ronaldo Mangueira
author_facet Lima Jr, Ronaldo Mangueira
Garcia, Guilherme Duarte
author_role author
author2 Garcia, Guilherme Duarte
author2_role author
dc.contributor.author.fl_str_mv Lima Jr, Ronaldo Mangueira
Garcia, Guilherme Duarte
dc.subject.por.fl_str_mv Análise quantitativa de dados
Modelos de regressão
Testes estatísticos
Significância estatística
Quantitative data analysis
Regression models
Statistical tests
Statistical significance
topic Análise quantitativa de dados
Modelos de regressão
Testes estatísticos
Significância estatística
Quantitative data analysis
Regression models
Statistical tests
Statistical significance
description In this study, we illustrate the potential variability of statistical significance by comparing four different methods, namely, t-test, ANOVA (followed by Tukey HSD), simple linear regression, and mixed effects linear regression. In our demonstration, we model reaction times as a function of different affixes in Danish, and show how our conclusions regarding the effect of certain affixes can change categorically depending on which of the aforementioned methods we choose to use. Finally, we echo recent studies (e.g., BARR et al., 2013), and suggest that mixed effects models be the norm whenever grouped data is analyzed. With our comparison, we hope to raise researchers’ awareness to the need for well-informed and ethical analytical decisions in linguistic studies.
publishDate 2021
dc.date.none.fl_str_mv 2021-08-05
dc.type.driver.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
Tutorial
Tutorial
format article
status_str publishedVersion
dc.identifier.uri.fl_str_mv https://revista.abralin.org/index.php/abralin/article/view/1790
10.25189/rabralin.v20i1.1790
url https://revista.abralin.org/index.php/abralin/article/view/1790
identifier_str_mv 10.25189/rabralin.v20i1.1790
dc.language.iso.fl_str_mv por
language por
dc.relation.none.fl_str_mv https://revista.abralin.org/index.php/abralin/article/view/1790/2316
https://revista.abralin.org/index.php/abralin/article/view/1790/2383
dc.rights.driver.fl_str_mv Copyright (c) 2021 Ronaldo Mangueira Lima Jr, Guilherme Duarte Garcia
info:eu-repo/semantics/openAccess
rights_invalid_str_mv Copyright (c) 2021 Ronaldo Mangueira Lima Jr, Guilherme Duarte Garcia
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
text/xml
dc.publisher.none.fl_str_mv Associação Brasileira de Linguística
publisher.none.fl_str_mv Associação Brasileira de Linguística
dc.source.none.fl_str_mv Revista da ABRALIN; V. 20, N. 1 (2021); 1-19
Revista da ABRALIN; V. 20, N. 1 (2021); 1-19
0102-7158
10.25189/rabralin.v20i1
reponame:Revista da ABRALIN (Online)
instname:Universidade Federal do Paraná (UFPR)
instacron:UFPR
instname_str Universidade Federal do Paraná (UFPR)
instacron_str UFPR
institution UFPR
reponame_str Revista da ABRALIN (Online)
collection Revista da ABRALIN (Online)
repository.name.fl_str_mv Revista da ABRALIN (Online) - Universidade Federal do Paraná (UFPR)
repository.mail.fl_str_mv rkofreitag@uol.com.br || ra@abralin.org
_version_ 1798329771399053312

Registros relacionados