2003-09-25_PWETC_AgendaPacket - Laserfiche WebLink

Home

II " II . I~ '. i " 16 The night-day accident ratios were as follows: Location 132nd to 146th 146th to 167th Before 3.0:1 2.9:1 After 1.3:1 2.0:1 The chi-square test was applied to the night-day ratios. The probability of the differ- ence in rates being due to chance occurrence was found to l?:e less than 1 in 100 (a chi- square value significant at 1 percent). This result suggests that the accident rate ratio is significantly lower after lighting was Installed. However, the data consist of only 12 observations, which does not fuifill the generally accepted requirements of at least 20 observations for a 2 by 2 chi-square contingency test (each cell shouid have an expected value of at least 5). Hence, the result of thi,s test must be treated with some caution. As a second statistical test, it was assumed that any trend in day accident figures would also be representative of any trend in night accidents. From the day values for the 2 study sections, a trend was established during the period covering both the before and the after conditions. This trend was applied to the before night figures to obtain the expected number of accidents during the after period, if lighting had not been in- stalled. A t-test was then'performed on the difference between this expected figure and the actual figure after installation of lighting. The t-value for both sections together was significant at the 10 percent level. For the north section only, a t-va1ue of 5 per- cent was found. These tests indicate that the installation of lighting quite possibly lowered the night accident rate, but an exhaustive statistical confirmation is lacking. LIGHTED VERSUS UNLIGHTED SECTIONS OF SAME ROUTE Interstate 85 in Atlanta is lighted, but its extension in DeKa1b County is unlighted. Direct comparison studies were performed in which 2 years of data from the lighted section and 4 years from the unlighted section were used. The lighted portion is 3.4 miles in length and extends northeast from a major inter- change point. The basic design consists of two 24-ft roadways with a 14-ft median. There are 4 interchanges along this section, in addition to the main freeway junction. The junction is of a directional type, and the interchanges are essentially of the dia- mond type. During the 1966 and 1967 study period, the route experienced 189 MVMT. Fixed lighting utilizes a 28-ft mounting height, type 3 distribution, 6-ft overhang, and 120-ft staggered spacing. Measurements taken at an interchange location on the route during 1967 showed a maintained illumination level of 0.38 HFC and a uniformity ratio of 2.7 to 1. However, subsequent measurements in a location more typical of the over- ail route average found a level of 0.33 HFC and a uniformity ratio of 16 to 1. The unlighted section is 8.8 miles in length and is generaily similar to the lighted portion except for a wider median and lower traffic volume. The section has 7 inter- changes, generally of the diamond type. One is a major interchange with a partiaily completed circumferentiai freeway loop. During the 1963 to 1967 study period, the un- lighted section had 523 MVMT. The relation of the number of accidents, roadway elem"ents, and day or night condi- tions is gtven in Table 4. The accidents have been grouped into the foilowing types: Accident Type Rear end 1 Other vehicular 2 Pedestrian and parked car 3 Fixed object and other off-road 4 The day and night accidents/MVM and the night-day ratios are aiso given in Table 4. (Ratios are omitted where only small accident sample sizes were available.) A substantial difference in direct accident rates exists between the 2 sections. It is, therefore, necessary to relate these rates on a ratio basis. The ratios for the unlighted route are higher for type 1 thanior type 2 accidents. Conversely, the ratio for type 3