Equívocos estatísticos: perguntas que você sempre quis fazer, mas nunca teve coragem / Statistical misconceptions: questions you've always wanted to ask, but never dared

Este artigo visa elucidar algumas dúvidas enfrentadas ou equívocos estatísticos cometidos por pesquisadores de diversas áreas. São explanados os temas: "tradução não é validação", "análise fatorial exploratória ou confirmatória", "nem todo estudo com dois grupos tem delineamento caso-controle", "teste ou ajuste de Bonferroni", "tamanho de amostra para teste de hipóteses e/ou para intervalo de confiança", e "testes ou dados paramétricos". A abordagem é realizada em uma linguagem acessível ao público leigo, utilizando exemplos e sugerindo referências para aprofundar o conhecimento.(AU)

This article aims to answer some questions and elucidate statistical misconceptions of researchers from different fields. The following topics are addressed: "translation is not validation", "exploratory or confirmatory factor analysis", "not every study with two groups is a case-control study", "Bonferroni test or adjustment", "sample size for testing hypotheses and/or for confidence intervals", and "parametric data or tests". The topics are explained in lay terms, using examples and suggesting references to advance knowledge.(AU)

What is the proper way to apply the multiple comparison test? / 대한마취과학회지

Multiple comparisons tests (MCTs) are performed several times on the mean of experimental conditions. When the null hypothesis is rejected in a validation, MCTs are performed when certain experimental conditions have a statistically significant mean difference or there is a specific aspect between the group means. A problem occurs if the error rate increases while multiple hypothesis tests are performed simultaneously. Consequently, in an MCT, it is necessary to control the error rate to an appropriate level. In this paper, we discuss how to test multiple hypotheses simultaneously while limiting type I error rate, which is caused by α inflation. To choose the appropriate test, we must maintain the balance between statistical power and type I error rate. If the test is too conservative, a type I error is not likely to occur. However, concurrently, the test may have insufficient power resulted in increased probability of type II error occurrence. Most researchers may hope to find the best way of adjusting the type I error rate to discriminate the real differences between observed data without wasting too much statistical power. It is expected that this paper will help researchers understand the differences between MCTs and apply them appropriately.

A Simple Scalable Association Hypothesis Test Combining Gene-wide Evidence from Multiple Polymorphisms.

Aims: In single-nucleotide polymorphism (SNP) scans, SNP-phenotype association hypotheses are tested, however there is biological interpretation only for genes that span multiple SNPs. We demonstrate and validate a method of combining gene-wide evidence using data for high-density lipoprotein cholesterol (HDLC). Methodology: In a family based study (N=1782 from 482 families), we used 1000 phenotype-permuted datasets to determine the correlation of z-test statistics for 592 SNPHDLC association tests comprising 14 genes previously reported to be associated with HDLC. We generated gene-wide p-values using the distribution of the sum of correlated zstatistics. Results: Of the 14 genes, CETP was significant (p=4.0×10-5 <0.05/14), while PLTP was significant at the borderline (p=6.7×10-3 <0.1/14). These p-values were confirmed using empirical distributions of the sum of χ2 association statistics as a gold standard (2.9×10-6 and 1.8×10-3, respectively). Genewide p-values were more significant than Bonferronicorrected p-value for the most significant SNP in 11 of 14 genes (p=0.023). Genewide p- values calculated from SNP correlations derived for 20 simulated normally distributed phenotypes reproduced those derived from the 1000 phenotype-permuted datasets were correlated with the empirical distributions (Spearman correlation = 0.92 for both). Conclusion: We have validated a simple scalable method to combine polymorphism-level evidence into gene-wide statistical evidence. High-throughput gene-wide hypothesis tests may be used in biologically interpretable genomewide association scans. Genewide association tests may be used to meaningfully replicate findings in populations with different linkage disequilibrium structure, when SNP-level replication is not expected.