Interval estimation as a function of constant and interrupted stimulation

backward m asking fl\J1ction in vision. Psychonornic Seienee, 1965,2,75-76. WEISSTEIN, N., JURKFNS, T., & ONDERISIN, T. Effeet of force<1 choice vs m agni tude·estimation mcasurC.., on the waveforrns of rnetacontrast funcuons. Journal 01' the OpticaJ Soeicty 01' Ameriea, 1970, 60, 97S-98 1. Interval estimation as a function of constant and interrupted stimulation* GERALD LONG and SUCHOON MO The University ofDetroit, Detroit, Mich. 48221 The present study investigated the relationship between a temporal interval's perceived length and the amount of stimulus and stimulus change within that interval. In three separate tests, completely mied intervals were found to be significantly overestimated by the Ss in comparison with empty intervals. In contrast to this, interrupted or divided intervals were underestimated. There appears to be some confusion over the effects of stimulation upon short temporal estimations. It is generally agreed that, in comparison with an "empty" interval, a completely "filled" temporal interval (Le., some stimulus duration) is usually overestimated by S (Meumann, 1896; Fraisse, 1963). However, the effects of "interrupting" or "dividing" temporal intervals are still somewhat in question. Following Hall & Jastrow (1886), Fraisse (1961, 1963, 1965) has to draw an analogy attempted oetween uppers illusion 01' visual extent, in which a divided line appears longer than an unbroken line, and the interrupting of a 'The authors wish to thank Mr. Jerorne Heckenmue11cr and Mr. Robert Nove11y for their aid in analyzing the data and Miss fathy McIvor for her help in its co11cction. Psychon. Sei., 1970, Vol. 21 (2) temporal interval with various stimuli. Such an interrupted or broken interval shou1d then appear longer or be overestimated by S when compared to an empty interval of objective1y the same 1ength. Other researchers, however, have found opposite results in their studies, results perhaps more in line with the common-sense notion of change making time appear to pass more quickly (Benussi, 1917; Roe10fs & Zeeman, 1949, 1951; Hawkins & Meyer, 1965). The present study attempts to clarify this area in regard to the effects of both constant and interrupted stimulation upon S's short temporal estimations. METHOD The Ss were 13 male and 12 female u n dergraduates at the University of Detroit. The apparatus employed were a Lafayette reaction-time machine, an Industrial Timer Corporation stopclock, a buzzer, and a tele graph key. By depressing a button, E could both start the timer and present a light of moderate intensity to S. S had before hirn a button which, by depressing, would terminate the light stimulus and stop the timer. E also controlIed a buzzer by means of a tele graph key, thereby enabling hirn to present the buzzing sound at any point during any of the trials. Finally, all trials were run in a semidark and sound-deadened room. Test 1 Each of 15 Ss was initially run through 10 practice trials in which E presented the light and, after 3.25 sec, instructed S to terminate it by pressing his button. This was done to acquaint S with the approximate length of a 3.25-sec interval. Forty test trials now followed in which S was instructed to "pay attention" to the light before him. When it went on, he was to estimate 3.25 sec and then turn it off by pressing the button. There was approximately a 5-sec intertrial interval. Randomly mixed within the 40 trials were four different auditory conditions (10 trials of each), which differed as to the amount of buzzer presented by E during the interval. On any particular trial, E presented the buzzer once (B 1) or twice (B2) or continuously (BC) or not at all (NB). S's temporal productions of his estimate of a 3.25-sec light duration were recorded by E from the stopclock after each trial. S was instructed to sit quite still and not to use any form of rhythmical tapping in estimating the duration. Test 2 The instructions given Sand the procedure followed in Test 2 were alm ost identical with those of Test 1, except that the length of the light duration to be 1:1 ~ zo 3. \m . ;:: u ::> . c ~ ;( '"0 3. ~ST'" 3. ~ 2.8 .::: J TEST I NB BI B2 Be AMOUNT OF AUDITORY STIMULI Fig. l. The effects of varying amounts of auditory stimuli upon S's temporal estimations. 87 / INTERRUPTED INTfRVALS oL-~ ______ ~~~ ______ ~ __ ';r THE NUMIER OF STIMULI FIlLING THE INTElVAL THE NUMIEI Of TlMES THE INTEIVAL WAS INTEI.UPTED Fig. 2. The effects of the type and degree of stimulation within the intervaI on S's productions in Test 1. estimated by S was changed. After 10 practice trials with a 7.00-sec light duration, S was instructed to terminate the light after 7.00 sec rather than 3.25 sec as in Test 1. Trus test was run approximately 1 week after Test 1, using 10 Ss who had participated in the earlier test. Test 3 A second group of 10 Ss was run through the experiment producing estimates of 3.25-sec intervals, but without any light stimulus. Each S was instructed to listen for the "click" made by the start of the re action-time machine. After hearing the click, S was to wait 3.25 sec and then press his button. No light stimulus was presented in any of these trials. In regard to the four auditory conditions, Test 3 was identical with Test 1. RESULTS Presented graphicaIly in Fig. 1 are the overall results of the three tests. It can be seen that, in general, the effects of the varying amounts of auditory stimuli within each test were very similar. Approximately the same data trend is revealed in a11 three curv~s. _ Test 1 The up-down character of the graph in Fig. 1 may at first appear confusing, unless the data are analyzed in terms of a possibly fundamental difference existing within the four auditory conditions. This possibility exists in that the first and last auditory condition (NB and Be) were basically different from the two middle conditions in terms of stimulus quality. Both the first and last condition were actually completely "filled" intervals of time [ftlled with light only (NB) or with light and buzzer (BC)], while the other two conditions were auditorily "interrupted" or divided intervals in the presence of the constant light stimulus. In Fig. 2 it can be seen that, graphed with this distinction kept in mind, two different and opposite effects appear to result from ftlling vs interrupting the intervals. The effect of 88 filiing an intervai was a slgni6cant underproduction (p< .01) by S, in comparison with interrupting the interval. An unexpected, but interesting, resuIt was that filling the interval even more led to further significan t underproduction (p< .01), whereas interrupting more to INTERRUPTED more overproduction (p < .0 I). INTf.RVAlS Test 2 As evidenced in Fig. 1, the same apparent relationships are also clearly fwice present in Test 2, which simply employed Ion ger (7 .00 sec) temporal productions tHE NUMBER OF THl: NUMBER OF TlMES THE INTERVAL STIMULI FllliNG than Test 1. (...truncated)