A. Nguessan et C. Langrand, ON THE LIMITING DISTRIBUTION OF SOME STAT ISTICS USED IN ROAD SAFETY STUDIES, Comptes rendus de l'Academie des sciences. Serie 1, Mathematique, 317(4), 1993, pp. 401-404
To evaluate the effectiveness of a road safety measure, the number of
accidents before and after the measure, classified in r categories, ar
e observed on s sites, with theta(k) the effect of the safety measure
for the site k (k = 1, 2, . . ., s). We propose a statistic to test th
e hypothesis H0:theta1 = theta2 = . . . = theta(s) = theta. Its asympt
otic distribution is given under certain conditions.