Abstract

Those unfamiliar with statistical methods tend to over-estimate their usefulness. To this group let it be pointed out that (1) statistics <i>alone</i> cannot take the place of rational physical theory in developing a good analytical method; (2) statistics <i>alone</i> cannot take the place of careful procedure and manipulations in carrying through a method; and (3) statistics <i>alone</i> cannot improve the results obtained by a particular method regardless of how poor the results may be. Thus, statistics is by no means a magic wand which permits the chosen few who possess it to use any sample preparation, any excitation, any spectrograph, any development procedure, any densitometer, any spectrum lines, in fact, any old method at all, yet always get the right answer by squaring a few numbers and dividing by something or other.

Application of statistical methods to the determination of slope in lidar data

David N. Whiteman
Appl. Opt. 38(15) 3360-3369 (1999)

Application of a wavelet-Galerkin method to the forward problem resolution in fluorescence Diffuse Optical Tomography

Anne Landragin-Frassati, Stéphane Bonnet, Anabela Da Silva, Jean-Marc Dinten, and Didier Georges
Opt. Express 17(21) 18433-18448 (2009)

Statistics of the slope-method estimator

Francesc Rocadenbosch, Adolfo Comerón, and Lorena Albiol
Appl. Opt. 39(33) 6049-6057 (2000)

Application of the Monte Carlo method to problems in visibility using a local estimate: an investigation

John M. Davis, Thomas B. McKee, and Stephen K. Cox
Appl. Opt. 24(19) 3193-3205 (1985)

Curvilinear coordinate method as an initial value problem: application to gratings

Cihui Pan, Richard Dusséaux, Mandiaye Fall, and Nahid Emad
J. Opt. Soc. Am. A 32(1) 143-149 (2015)