Title
Empirical Statistics: V. Evidence On The Extent Of The Steepness Bias In Visual Estimation Of Trends
Abstract
The existence of a pervasive and pronounced steepness bias in visually fitting a line to data in a scattergraph was investigated. 262 undergraduate business students were asked to fit lines visually to scattergraphs, to correspond to a least squares regression fit. These visually estimated lines strongly overestimated the steepness of the actual trend. As visual inspection of data is an important step for the detection of linear trends and outliers prior to regression analysis, this bias in subjective perception of the line may result in inappropriate deletion or retention of outliers, transferring the subjective bias into the results of least squares regression. In addition, textbooks on management accounting discuss the use of visually fitted lines to estimate cost behavior.
Publication Date
1-1-1996
Publication Title
Perceptual and Motor Skills
Volume
84
Issue
3
Number of Pages
731-734
Document Type
Article
Personal Identifier
scopus
DOI Link
https://doi.org/10.2466/pms.1996.82.3.731
Copyright Status
Unknown
Socpus ID
1842639898 (Scopus)
Source API URL
https://api.elsevier.com/content/abstract/scopus_id/1842639898
STARS Citation
Bailey, Charles D., "Empirical Statistics: V. Evidence On The Extent Of The Steepness Bias In Visual Estimation Of Trends" (1996). Scopus Export 1990s. 2227.
https://stars.library.ucf.edu/scopus1990/2227