' " :,.. "-.' j Drivers' Brake Reaction Times GUNNAR JOIIANSSON and KARE RUMAR, Department of Psychology, Ulliversityof Uppsala, Uppsala, Sweden rhe ob u( 0/ (his inl'es(igalion was (0 determine /he dis/ribu/ion of brake reaction limes .... hich can be expec/ed from dril'ers who have 10 brake ruddenly and comp/euly unexpectedly in tra/lic situations. The experimellts were carried oul as follows: I. Broke reaclion time was measured on a large group of drivers flU) in an anticipa/ed si/ualion on Ihe road (Blake rl?ac/io/l lime I). 2. A small group 0/ drivers (5) was repea/edly tes/I?d ill Ihe same way (Brake reaction lime 2). l. The same small group was repealedly lesled in a surprise silualion (Brake reaclion time J). 4. The ra/io of brake reaction lime 3 10 brake reaclion lime 2 was usrd as a correction factor and applied 10 brake reac/ion time I. The cofTecled median of Ihe rerulting distributi01l was 0.9 sec. 25% of Ihe group was eSlinUlled to "ave a brake (cac/ion (ime 10llger Ihan 1.2 sec. INTRODUCTION The usual driver reaction to a potential accident situation is sudden braking. As a consequence, one of the main factors deter�� mining whether or not the accident will be avoided is the driver's reaction time. This report describes a me thodological investigation of brake reaction time. An impor. tant methodological problem in studying brake reaction time has been the difficulty of gelling measurements which would be applicable to ordinary traffic situations. Furthermore, from "the point of view of traffic safety it is desirable to know the distribution of brake reaction time over the population of drivers. The specifIC objectives of the study were as follows: I. To obtain a correction factor for esti�� mating brake reaction time when braking is unexpected from brake reaction time measures obtained when braking is expected. 2. To determine the brake reaction time distribution in a representative sample of drivers when braking is unexpected. Reaction time has long been the object of study in the fields of physiology and psychol�� ogy. Two excellent surveys of reaction time investigations rekvant "to driving situations are those by Teichner (1954) and Forbes and Katz (1957). Th~" pi~ture" that has emerged from the 23 studies reviewed is that reaction time varies with the type of lask involved, with level of attention over different sensory areas, and from individual to individual. Thus, for an accurate assessment of braking reaction time, the measurements must be" made under conditions which are comparable to those in which the results are to be applied. Brake reaction time measures obtained in the laboratory and applied directly to actual traffic situations are likely to result in incorrect conclusions. Lister (1950) summarized investigations which" were directly concerned with brake reaction time. In the majority of these, how- ever, the measurements have been made "in "stationary" situations, usually under laboca�� tory conditions, and always with the subject havillg a fairly high degree of.. expectancy. The procedure typically called for the subject to press a pe4al as soon as a lamp was lit. The mean brake reaction time obtained under these conditions was found to vary between .45 and .60 sec." Lister, in his investigations, split the total brake reaction lime inlo perception time (time from the presentation of the stimulus until the foot slarts to move) and movement time (time from the start of the movement until the foot reaches the brake pedal). He also compared brake reaction time measured in a stationary vehicle with that measured in a vehicle moving

. 24-February, 1971 t ~'. :. ,~: " HUMAN FACTORS at different speeds. When tbe expectancy of the The accuracy of tbe measurements were drivers was relatively high, the perceplion time obviously dependent on the reliabilily of the proved to be about��.25 sec. (with very little assistant's reaction tir!1e. Thus the standard variation) and the movement time about .IS error of his reaction time should be negligible sec. He ascribed increased brake reaction time compared to that of the brake reaction lime. variation in condition oC.lower expectancy to 4" (The criterion used was that the spread of the increased variation in perception time. d.- ,' :' assistant's reaction lime should be less than ,j'" .' 2C1Xo of tbe brake reaction time spread. Thus, interval beN times more I It was f familiarizing .............. -��the ftrSt si enough had he realized \ for. nus COL' in brake reae the fust Out foc allowing to the buzzt correspondin measurement results. In a curve bad rea signa!. MEASUREMENT METHODS Two methods of measurement wert em�� ployed in this study, The brake reaction~ime Of each of a large group of drivers was measured once under norma) highway driving conditions, but with some degree of brak.ing expectation. The second method involved repeated measure�� ment of a small group of drivers, including' measures obtained . under no-expectancy conditions. Method I: Single Measures 0/ a Large Sample Every southbound car along a road was stopped by the poUce and lhedriver was asked if he was willing to take part in the experiment (in fact, every driver participated). After he had received his braking instructions, he was aHowed to proceed with his journey. Sometime during the next 10 Ion., the driver WQuld hear a loud klaxon hom sound al Ihe side of his car. This was the signal for immediate braking. The car was not to be broughl 10 a standstill, however the driver was to release the brake pedal and. continue his i9umey as soon as the klaxon was silent. About 5 km. from the police trap was a concealed measuring station. When a car passed 8 photoceU this triggered a klaxon hidden in a letlerbox at the side of the road. Simultane�� ously an electronic timer (accurate to .001 sec.) was started. A speciaUy trained assistant, also well hidden but with a good view of the road, was on the alert to press a contact key as soon 'as he saw the cat~s braking lights glow. This contact also fUllctiooed as a circuit breaker so -'that' the klaxon became silent and the time between its onset and the breaking of the circuit was recorded on lhe timer. .. ,,- , ~~ .' ,'.. .' with 02 tol . :: O'BRT t a 2 J1.T' reactioo time contnlmles only 4% of the total vuiance when the ratio ORT/VBRT :: 1:5). The accuracy of tbe assistant as a mea~.lIing instrument was controlled as conowS: 1. The assistant was chosen on the grounds of his having a relativelyconstanl simple reaction time, as measured in pretrials. He was then trained so that he would respond with as constant (not as short) a reaction time as possible, 2. During the actual measuring of the drivers' . brake reaction time, the assistant was regulady i'calibrated" with respect to his reac�� tion time. During the course of the investi�� gation proper, 322 such calibralion meuure�� ments were made and were found to h.m: a near-normal distnoution with M : .244 tee. and 0::: .016 and a range of .20-.29 sec. Thismcans that the spread of the assistant's reaction time about the mean .24 sec. can be estimated as not shifting any single brake reaction time n1ue more than .OS sec. in either direction. Method /I: Repealed Measures 0/ II" Smail Sample The apparatus used for repeated measures was connected to both the ignition and braling systems of the car, and functioned as foUows: When the car's ignition system was twitched on, an electric clockwork system started, which l . ," The main { bution of bral from an appro drivers who 2 -~::::... tion" within 1 Procedure All cars an exception of vehicles (amb drivers who p who had earli, lion. braked i signal was nol those drivers w before reachlnl cases where e~ could lead to vehicles travell closed the circuit to the buzzer and the electric . however, occur: timer at different set intervals. When the brut ! reason to belic\' ~~ ~~li~~~~~n~ir':tcl:~ ~:pk~~.~n ~ . \ . :~.-- the tesults. '~7' way the time interval between the onset of the .~ " ������. Resul~ bu.zzer and the braking reaction was noted, with an accuraCy of .1 I��C, The intem..l t ~:i' Brake reacth between the signals was often Jet so that it "'IS '~r(hiveri the d~s greater than an hour, and since the car 'IflI:S I .~ (1l1.e reaction til mainly used for traveling to and from wor". UtbcJe-....._-.....................��-----��--,��'. .-.~ ... - ... ~ .. . -4.~