THE 7 DEADLY SINS OF STATISTICAL-ANALYSIS

Citation
Wm. Kuzon et al., THE 7 DEADLY SINS OF STATISTICAL-ANALYSIS, Annals of plastic surgery, 37(3), 1996, pp. 265-272
Citations number
9
Categorie Soggetti
Surgery
Journal title
ISSN journal
01487043
Volume
37
Issue
3
Year of publication
1996
Pages
265 - 272
Database
ISI
SICI code
0148-7043(1996)37:3<265:T7DSOS>2.0.ZU;2-E
Abstract
In a pedantic but playful way, we discuss some common errors in the us e of statistical analysis that are regularly observed in our professio nal plastic surgical literature. The seven errors we discuss are (1) t he use of parametric analysis of ordinal data; (2) the inappropriate u se of parametric analysis in general; (3) the failure to consider the possibility of committing type II statistical error; (4) the use of un modified t-tests for multiple comparisons; (5) the failure to employ a nalysis of covariance, multivariate regression, nonlinear regression, and logistical regression when indicated; (6) the habit of reporting s tandard error instead of standard deviation; and (7) the underuse or o veruse of statistical consultation. Confidence and common sense are ad vocated as a means to balance statistical significance with clinical i mportance.