We reassess the mathematical expression for statistical distributions of latitudinal extents of electron acceleration events; the distributions have been obtained by Newell et al. (1996, hereinafter referred to as N96) using precipitation data from all Defense Meteorological Satellite Program satellites. Paper N96 proposed a basic hypothesis that within a (given) large-scale region of upward field-aligned currents (FACs), any 1-s spectrum has a constant probability, P, of experiencing electron acceleration independently of neighboring spectra. Directly from this hypothesis, the probability of observing n consecutive accelerated spectra, denoted by Qn, is newly derived as the correction of "Pn-1(1 - P)" that was assumed in N96. Necessarily, Qn is a function of the probability P and the latitudinal width W of a large-scale FAC region; hence it is written as Qn(W, P). The occurrence frequency distribution (histogram) of latitudinal extents of observed acceleration events can be identified with a statistical superposition of a number of Qn(W, P) in relatively wide ranges of W and P. For fitting to the histograms obtained in N96, 0.1¿ $lesssim$ W and 0.7 $lesssim$ P $lesssim$ 0.98 are appropriate. This prediction is quite in contrast with N96's supposition that P is nearly constant and that normally the width of a large-scale FAC region is sufficiently large to accommodate many accelerated spectra. (The appearance of region 1 FACs of narrow widths has already been identified both observationally and theoretically.)