ммммммммммммм ммм мммм мллллллллплпппппллм ммппппмммм млм мллллллллмммммлпп плл пллллллллллллллм плл млллмллммм пллллм плллллллмллмммллн лп пппллллллллллм он лллллллллллпплмоллл лллллм пплллоп Mo.iMP мллм плллллллнл оллллллллл олллл олллллн плп мллллллл ллллллллн оллллллллн ллл лллллл млллллллн олллллллл олллллллл п оллллллм мл млллллллн лллллллл ллллллллн оолллллллллп млплллллл мм ллллллллн ллолллллн оллллллпп млпллллллмллллмолллллллл ол пллллл м лнллллл м мл оллллллллллп ллллллллл плм плллмм ммлллп ол олллн ммлл лл ллллллллп ллллллллллм плм пплллллллллп мммп ллллмммммммлллллп плм млллп пллллллллллм ппмм ппмллпп пллм ппплплллллллпп ппппп ппллп ппппп ппппппппппппп ARRoGANT CoURiERS WiTH ESSaYS Grade Level: Type of Work Subject/Topic is on: [ ]6-8 [ ]Class Notes [Lab: Response Results ] [x]9-10 [ ]Cliff Notes [from Stimulus ] [ ]11-12 [x]Essay/Report [ ] [ ]College [ ]Misc [ ] Dizzed: o4/95 # of Words:951 School: ? State: ? ФФФФФФФФФ>ФФФФФФФФФ>ФФФФФФФФФ>Chop Here>ФФФФФФФФФ>ФФФФФФФФФ>ФФФФФФФФФ>ФФФФФФФФФ Part 1 In this signal detection experiment, there were 3 parts, each with 50 trials. On half of the trials, a signal 'x' was presented along with visual noise 'z'. The other half of the trials showed only visual noise. The task was to locate the signal stimulus 'x' among all other noise signals 'z' present. All this was done on a computer generated program. Another 2 students' results of the experiment were obtained for further analysis of signal detection. During the trials there were many 'z' flashing on the screen. The objective was to determine if the symbol 'x' was also flashed on the screen. The responses to be given were : If you are sure you saw 'x' press 5 If you are fairly sure you saw 'x' press 4 If you are unsure you saw 'x' press 3 If you are fairly sure you didn't see 'x' press 2 If you are sure you didn't see 'x' press 1 The dependent variable in this experiment was the response in pressing 1,2,3,4 or 5. The independent variable was the signal stimulus 'x' which flashed across the screen. The design of this experiment is the within-subject design since the conditions were randomized from trial to trial and students were tested under mainly the same experimental conditions. The Questions which can be addressed with this design are how would the results have changed if the design had been a between- subject design? Also what would happen if the trials were not randomized? How would this factor distort the results in this within-subject design. Part 2 Results Group 1 Condition 1 Condition 2 Condition 3 0.01 sec. 0.05 sec. 0.1 sec. 1+2+3+4+5 5 1+2+3+4+5 5 1+2+3+4+5 5 Signal 20 13 29 29 25 24 Noise 30 0 21 2 25 1 P(Resp +/S) 0.6875 1 1 P(Resp +/N) 0 0.952 3.846 D' -4.7652 -1.374 -1.797 Group 2 Condition 1 Condition 2 Condition 3 0.01 sec. 0.05 sec. 0.1 sec. 1+2+3+4+5 5 1+2+3+4+5 5 1+2+3+4+5 5 Signal 25 7 25 24 25 24 Noise 25 0 25 6 25 1 P(Resp +/S) 0.68 0.96 0.96 P(Resp +/N) 0 0.24 0.04 D' -4.794 -0.857 -1.752 Group 3 Condition 1 Condition 2 Condition 3 0.01 sec. 0.05 sec. 0.1 sec. 1+2+3+4+5 5 1+2+3+4+5 5 1+2+3+4+5 5 Signal 18 13 5 20 4 25 Noise 19 0 25 0 20 0 P(Resp +/S) 0.419 0.84 0.862 P(Resp +/N) 0 0 0 D' -4.491 -4.933 -1.602 In group 1, the probability of the response signal stimulus 'x' was the greatest in condition 2 and condition 3 where it was 1. The probability of a noise signal was lowest in condition 1 where it was zero. The distance between the means of the two distributions was closest in condition 2. The value for D' in condition 2 was -1.374. So overall in group 1, the best response for the signal stimulus is shown in condition 2, where the stimulus flash duration was 0.05 seconds. In group 2, the probability of the response signal stimulus was greatest in condition 2, with 0.96. The lowest probability of a noise response was in condition 1, with the result of 0. The closest distribution of the 2 means was in condition 2 with -0.857. The best response to the signal stimulus 'x' was again in condition 2 with a stimulus flash duration of 0.05 seconds. In group 3, the highest probability of a response signal being a 'hit' was in condition 3, which was 0.862. The lowest probability for a false alarm was in all 3 conditions which equalled 0. The distance between the means of the two distributions was the lowest in group 3. Therefore in group 3 the best condition for responding to a hit was in condition 3, which had a stimulus flash duration of 0.1 seconds. Signal detection approaches provide ways of examining sensory factors and response bias. A hit rate in signal detection gives us an estimate of the subjects sensitivity. The higher the hit rate is, the more sensitive the subject. Also the higher the false alarm rate, the less the subject is going to respond with a hit rate. In responding to the stimuli the subject will either sense or not sense the stimuli, and in addition will also decide what to report. So because the subject decides what to report this is why all three groups had quite different results. Part 3 Two other variables that could influence the sensitivity dimension of the subjects response are: 1) The number of choices to be made if the signal stimulus was present. Five choices are too many (1,2,3,4,5). It would have been better if there were only three choices : sure you saw 'x', unsure, and sure you didn't see 'x'. 2) The time when the signal stimulus appeared on the screen. If it was the first or last signal shown, 'x' could be identified very easily. Another variable that may influence the response-criterion dimension of the subjects response was how dispersed the signals were on the screen. Sometimes they were bunched together and so it was easier to locate 'x'. But if the signals were all spread out across the screen then it was harder to locate the signal stimulus. A confounding variable in this signal detection experiment could be how you were positioned in your distance from the computer screen. If you were not positioned a constant distant away from the screen, this would affect how you saw the signals. If the signals were spread out and you were very close to the screen then you would not be able to clearly recognize the signal stimulus 'x' among all the other noise signals 'z'.