Researchers Misunderstand Confidence Intervals and Standard Error Bars.

Abstract

Little is known about researchers' understanding of confidence intervals (CIs) and standard error (SE) bars. Authors of journal articles in psychology, behavioral neuroscience, and medicine were invited to visit a Web site where they adjusted a figure until they judged 2 means, with error bars, to be just statistically significantly different (p < .05). Results from 473 respondents suggest that many leading researchers have severe misconceptions about how error bars relate to statistical significance, do not adequately distinguish CIs and SE bars, and do not appreciate the importance of whether the 2 means are independent or come from a repeated measures design. Better guidelines for researchers and less ambiguous graphical conventions are needed before the advantages of CIs for research communication can be realized.

Exploring Small, Confirming Big: An alternative system to The New Statistics for advancing cumulative and replicable psychological research

“…Based on the outcome of these tests, researchers must then select between competing meta-analytic models (e.g., trim-and-fill v. no trim-and-fill; or the precision effect test v. the precision effect estimate with standard error; see Duval & Tweedie, 2000a, 2000b, and Stanley & Doucouliagos, 2014. (Belia, Fidler, Williams, & Cumming, 2005).…”

A Perceptual-Statistics Shading Model

“…Due to violations of sphericity according to Mauchly's test, reported degrees of freedom and p-values are Greenhouse-Geisser corrected [10,23] Figure 11 do overlap, it does not imply that the effect is insignificant at 5% level [3]. Table 1.…”

La ville compacte remise en cause ? Formes urbaines et mobilités de loisirs

“…Sur les histogrammes, ils sont représentés par un trait ombré vertical s'étalant de part et d'autre de la moyenne. Le lecteur peut ainsi généralement conclure à des différences significatives entre deux moyennes, lorsque les deux traits correspondants ne se chevauchent pas, ou lorsque le chevauchement n'excède pas 25 % de leur longueur (Belia et al, 2005 ;Payton et al, 2003). Dans aucun cas cependant, nous n'avons conclu à l'existence de différences significatives par le seul examen visuel des graphiques sans procéder à des tests de comparaisons multiples (voir ci-dessous).…”