Perceptual and Motor Skills, 1976,42, 919-928. @ Perceptual and Motor Skills 1976
PRIMACY EFFECT I N ORIENTING RESPONSES TO AUDITORY STIMULI OF TONES AND MUSIC ELEANOR B. SIMOW California State College, Dominguez H i l h Summary.-This 26 factorial experiment investigated the primacy effect in the orienting response. The rype of stimuli (tone or "music"), stimulus intensities (loud or soft), length of adaptation period (same, 5 or 30 sec.; or different, 5 min.), interstimulus intervals ( 5 or 30 scc.), and sex were studied. College students, 32 males and 32 females were randomly assigned to each group. In the same condition, the tone (or music) was soft (or loud) for 5 sec. (or 30 sec.) in adaptation and was the0 changed alternately without interruption to loud, soft, etc. (or soft, loud, etc.) for 5 sec. (or 30 sec.). The different condition was identical except for the length of the adaptation period in which the stimuli sounded continuously for 5 min. Analyses of the GSR manifestation of the orienting responses indicated: ( a ) an over-all primacy effect with the auditory stimuli and ( b ) the primacy effect occurred in the 5-sec.-same but not in the 30-sec.-same condition as predicted.
Although much of the research concerning the orienting response both in the Soviet Union and in the U. S. has been concerned with its facilitating effects (e.g., Kintsch, 1965; Maltzman & Raskin, 1965; Simon, 1972; Sokolov, 1963) there have been indications (Maltzman, Harris, Ingram, & Wolff, 1971; Maltzman & Wolff, 1970) that it may also have inhibitory effects under certain conditions. This may occur when one orienting response follows another within a short interval (e.g., 5 sec.) producing a larger orienting response to the first stimulus change, the source of the first orienting response, than to the subsequent (second) change. This phenomenon, obtained by Maltzman, Harris, Ingram, and Wolff (1971), was termed the primacy effect. In the Maltzman, et ai. (1971) study, alternating increases and decreases (or the reverse) of light intensity presented continuously for 5-, lo-, or 30-sec. periods to separate groups, found that the mean magnitude of the first orienting response (in terms of the galvanic skin response, GSR) of each pair in the series (onset-offset, onset-offset, etc.) was reliably larger than that of the second. This result was obtained in all groups including the group given 30 sec. between presentations of stimulus-change, a relatively long interval. While a primacy effect was expected in the shorter periods, its appearance in the longer 30-sec. interval mas interesting. In this condition, the GSR, i.e., the orienting response to the first stimulus change, which is known to have a relatively long response latency ranging from approximately 1 to 3.5 sec. (Edel'This research was supported by Research Grant 1R03-MH22037-01 from the U. S. Public Health Service and an award from the Society of Sigma Xi. The author would like to thank Robert Gray for collecting the data and b o t h ~ o b e r rGray and John Douglas for their assisrance in serting up the equipment.
920
E. B. SIMON
berg, 1967; Venables & Martin, 1967), would require continuous GSR activation of an additional 27 sec. at least for it to have an inhibitory effect on the next GSR, the orienting response to the second change. According to Sokolov (1963), the GSR, once activated, returns to baseline level even while the stimulus is still in operation (see p. 116). The records presented in his book (e.g., pp. 58, 117, and 124) show that the return to normal takes place sooner than 30 sec. after the response reaches its lowest resistance level. The question arises then as to what to attribute the occurrence of the primacy effect with a 30-sec. interstimulus interval. One possible explanation may lie in the inadvertent confounding of the two variables, time and light intensity. A compounding of stimulus changes in temporal interval (a stimuluschange) and light intensity (another stimulus-change) could have produced an unexpectedly large GSR (Sokolov, 1963). A careful examination of the data shows that the change away from the initial level of illumination occurred first after a 10-min. adaptation period and again after only 30 sec. This temporal change combined with the simultaneous change in light intensity on the second trial when the direction of illumination change was away from the level experienced in adaptation may have produced a significantly larger GSR than the previous response when illumination was changed back to its initial level. Differences in the GSR for change and change-back conditions on subsequent trials in the 30-sec. group were small. In order to obviate any effect due to temporal changes, and, hence, determine whether the primacy effect in the 30sec. group was in fact due to changes in illumination, the experimental design should include a control condition wherein the duration of the adaptation period is the same as the intervals between presentations of stimulus-change. This experiment was designed to include such a control condition. This experiment also attempted to demonstrate the generality of the primacy effect by employing shifts in auditory stimulation instead of changes in illumination. According to Sokolov (1963) the modality of stimulation is irrelevant. It is any change in stimulation which produces an orienting response. Therefore, results obtained by shifts in auditory stimulation should be similar to those found for shifts in visual stimulation. In summary, the purposes of this experiment were: ( a ) to determine whether the primacy effect in the orienting response occurs when the length of the adaptation period is the same as the duration of stimulus presentations, and ( b ) to extend the generality of the primacy phenomenon from visual to auditory stimuli both simple (pure tones) and complex (music). It was hypothesized that: ( a ) there will be an over-all primacy effect; ( b ) that the primacy effect will occur in the short ( 5-sec.) same condition when the adaptation period and interstimulus interval are both 5 sec., and not i n the long (30-sec.) same condition; and ( c ) the primacy effect will occur in both the 5- and 30-sec. dif-
PRIMACY EFFECT IN ORIENTING RESPONSES
92 1
ferent conditions when the adaptation period is 5 min. and the interstimulus intervals are 5 or 30 sec. The latter finding would constitute a replication of the earlier results with illumination. METHOD Expen'nzental Design The experiment was a 2 X 2 X 2 X 2 X 2 factorial design with the following variables: ( a ) type of stimulus, tone or music; ( b ) length of adaptation period, same ( 5 or 30 sec.) or different ( 5 min.); ( c ) the interstimulus intervals, 5 or 30 sec.; ( d ) initial stimulus intensicy (in adaptation period), loud or soft; and ( e ) sex. The GSR was recorded throughout the experiment. It was measured in terms of the largest resistance change occurring between 0.5 sec. and 5 sec. following a stimulus-change. Sixty-four students from California State College, Dominguez Hills, 32 males and 32 females, served as subjects. These students were in the introductory psychology classes and volunteered to serve to fulfill a course requirement. They were assigned to each group on a predetermined random schedule.
Apprwatw by a Beltone Model 12c The tones, 1000 Hz at 45 d b and 80 db, Audiometer, were prerecorded on magnetic tape using a Wallensak T-3500 tape recorder. Similarly, the music, specifically prepared for this experiment by an electronic synthesizer, was first played and recorded simultaneously for 10 min. continuously at 45 db. Later, this tape was re-recorded such that the intensity of the music was varied at 45 db and 80 db according to the particular experimental treatment. For example, one tape played the music for 5 min. at 45 db then shifted without interruption to 80 db for 5 sec., then to 45 db for 5 sec., etc. for the required number of stimulus-changes. A separate tape was prepared for each of the experimental conditions yielding a total of 16 tapes. The tapes played on the recorder were presented to the subjects via earphones. The recorder was also hooked u p to one of the four channels on the Narco Biosysterns Model 4 Physiograph to mark the time when each stimulus was presented. This made for accurate recording of the GSR intervals during which the responses were read. Palmar GSRs were recorded with circular zinc discs 11 mm in diameter which were embedded 2.5 mm in plastic cups. The cups were filled with Beckman electrode pasre. The output from a Narco Biosystems Preamplifier using - a constant dc current of 20 pa was transmitted to one channel of a fourchannel physiograph recorder.
E. B. SIMON
The experimental room contained a comfortable lounge chair, the arms padded with foam rubber. All of the electronic equipment except for the leads which were attached to the subject were hidden behind a large wooden screen. After the subject entered the experimental room and was seated in the chair, one electrode was attached to each palm of his/her hands. Then, the following instructions were read: These pickups measure changes in the skin which occur to certain stimuli. Since the pickups are sensitive to body movements, it is important that you remain as still as possible, especially your hands. So, please try to get as comfortable as possible before the experiment begins. Are there any questions before we start?
The experimenter then placed the earphones on rhe subject, and went behind the screen to monitor the equipment. First, the polygraph was turned on to obtain the subject's baseline resistance level, then, the adaptation period and subsequent stimulus manipulations followed. The soft (45 d b ) and loud ( 8 0 d b ) 1000-Hz tones and music pre-recorded on tape were presented through the earphones. The first sound with appropriate intensity and duration constituted the adaptation period. The subsequent auditory changes (loud to soft, soft to loud, etc. and vice versa) followed immediately without: interruption. The subjects in the condition in which both lengths of adaptation period and the interstimulus interval were the same heard the sounds for 5 or 30 sec. throughout the experiment. For example, he or she listened to the soft tone for 5 sec. (adaptation period), then to the loud tone for 5 sec., then to the soft one for 5 sec., and so on. The subjects in the "different" condition listened to the loud (or soft) auditory stimulus for 5 min. then to the soft (or loud) auditory stimulus for 5 or 30 sec., and so on. Half the subjects in each group received the soft sound and half the loud sound as the initial stimulus, and half the subjects in each group were males and half were females. All subjects in the 5-sec. interstimulus interval condition received 60 trials of onset and offset, and those in the 30-sec. groups were given 10 such trials in order to equate the amount of time spent in the laboratory.
Before statistical analyses, the GSRs for the first 10 trials after the adaptation period for all groups were subjected to a log conductance change transformation (Haggard, 1949). A mixed analysis of variance with five berweengroups variables (cype of stimulus, adaptation period, stimulus intensity, interstimulus interval, and sex) and one within-groups variable (trials) was performed on these data. The results indicated that the type of stimulus used was reliable ( F 1,3? = 24.3 1, # < .001) . The over-all mean for tones (0.7982 ) was significantly larger than that for music (0.2391). Sex was statistically sig-
PRIMACY EFFECT IN ORIENTING RESPONSES
923
nificant = 8.64, p < .01), with males having a larger over-all mean GSR than females (0.6853 and 0.3520, respectively). The effect of trials was also significant ( F O ,=~ 6.47, ~ ~ p < .001), indicating that habituation over trials had occurred as expected. Fig. l a shows the trend of habituation for the group means over the 10 trials and the trends for tones and music separately over the 10 trials. The rate of habituation for tones was greater than that for music. The interaction between type of stimuli and habituation was statistically significant (p9,2s8 = 4.89, P < .01). It should also be noted that the levels of the GSR for music are far below those for tones.
Primacy Effect TO determine whether the primacy effect occurred, the analysis involved comparison of the mean GSR for Trials 1, 3, 5, 7, and 9, the first of each pair of stimuli in the series, with those for Trials 2, 4, 6, 8, and 10, the second of each pair. This constituted the within-groups variable in a mixed analysis of variance including the five between-groups variables. Again the main effects for type of stimulus = 23.8, p < .001) and sex = 8.6, p < .01) were statistically significant. The mean for tones was larger than for music ( 3.9910 vs 1.21 l l ) , and the mean GSR for males was larger than that for females (3.4420 vs 1.7601).
0.1
5
TRIALS
FIG. la Mean GSRs over 10 trials for music and tones combined, tones only, and music only
4
3
TRIALS
FIG. lb. Mean GSRs for stimuluschange and change-back for all trials. (Means for stimulus-change resulted from Trials 1, 3, 5, 7, and 7, and for change-back, Trials 2, 4, 6, 8, and 10.)
924
E. B. SIMON
The over-all means for the primacy effect were 3.0064 and 2.1956, and the analysis of difference between these means indicated the difference was reliable = 8.1, p < .01). The significance of this result lies in the fact that the GSR to the first stimulus-change in each pair of stimuli is larger than the GSR to the second change, that is, the change-back to the initial level of intensity. Fig. l b shows that mean GSRs for stimulus changes are clearly larger than the GSRs for change-back over all trials. A most interesting result of the analyses was the statistically significant triple interaction among the variables of the adaptation period (same or different), length of the interstimulus interval ( 5 or 30 sec.), and the primacy effect = 6.2, p < .05). Considering the condition in which the length of the adaptation period was the same as the interstimulus interval, the primacy effect occurred when the interval was 5 sec. in length but did not occur in the 30-sec. interstimulus interval. The mean GSRs for these conditions appear in Table 1. The analysis of the difference between the means for the 5-set.-same condition was reliable (FIq8= 13.58, p < .01). When the length of the adaptation period is different ( 5 min.) from the interstimulus interval, the means for both the 5- and 30-sec. conditions indicate the appearance of the primacy effect as expected (Maltzman, et d., 1971). However, only the result for the 30-sec. group was statistically significant (Fl,s = 7.40, p < .05). TABLE 1
Adaptation Period Different: Intentimulus Interval Same: Interstimulus Interval Primacy Effect Change Change-back
Figs. lc, d, e, and f show the primacy effect separately for the 5-sec.-same, 30-set.-same, 5-sec.-different, and 30-sec.-different conditions. The relationship is clearly evident in the 5-sec.-same condition compared with the 30-sec.-same condition. The reverse occurs in the different conditions, with the primacy effect shown in the 30-sec. but not in the 5-sec. condition. The results of the analysis of variance also indicated that the interaction between the initial stimulus intensity and the primacy effect war reliable (F1,32 = 24.2, p < .01) as was the triple interaction among type of stimulus and these variables (F1.32 = 10.2, p < .01). For tones, when the first stimuluschange after the adaptation period was from soft to loud and the change-back was from loud to soft, there is a significant primacy effect. The means for stimulus-change and change-back are 5.91 50 and 2.582 3, respectively. When
PRIMACY EFFECT I N ORIENTING RESPONSES
0.1 1
2
3
4
6
TRIALS
FIG. lc. Mean GSRs for stimuluschange and change-back for all trials i n the 5-sec-same groups
TRIALS
FIG. l d . Mean GSRs for stimuluschange and change-back for all trials in the 30-sec.-same groups
1
TRIALS
FIG. le. Mean GSRs for stirnuluschange and change-back for all trials in the 5-sec.-different groups
2
3
4
5
TRIALS
FIG. l f . Mean GSRs for stimuluschange and change-back for all trials in the 30-sec.-different groups
926
E. B. SIMON
the intensity changes were in the reverse directions, from loud to soft and soft to loud, the primacy effect did not occur. The means for stimulus-change and change-back were 3.0922 and 4.3742, respectively. For music, the means for stimulus-change are larger than the means for change-back in both situations: soft to loud, loud to soft, and vice versa. The differences, however, were too small to be significant.
DISCUSSION Habituation The habituation curves for the two types of stimuli differed in two respects: ( a ) the magnitude of the GSRs for the 1000-Hz tones was significantly larger than that for music even though the intensities and durations of these stimuli were the same for both, and ( b ) the rate of habituation for tones was more rapid than that for music. According to Sokolov (1963) and Lynn (1966), the magnitude of the GSR varies with the intensity of the stimulus; for moderate intensities, the greater the intensity the larger the magnitude of GSR. In this study, the size of the GSR should have been the same for tones and music since their intensities were the same. If there was a difference, musical stimuli being more complex than tones should have elicited larger GSRs (Berlyne, 1960; Berlyne, Craw, Salapatek, & Lewis, 1963). However, since pure tones are seldom encountered in real life experiences, whereas music (even the metallic sounding music produced on the electronic synthesizer) is heard quite frequently, it may be that the familiarity of the music resulted in reducing the magnitude of the GSRs. During the experiment, many of the subjects in the music condition sat back comfortably in the chair, closed their eyes and seemed to enjoy the music. Actually, the music was an original composition composed extemporaneously by a graduate student in music for the sole purpose of this study. It had no regular rhythm nor apparent melody. Nevertheless, the college students accepted it as real music (perhaps rock music?). Lynn (1966) has summarized the results of studies indicating that the rate of habituation differs as a function of the stimulus intensities, stimulus durations, and other variables. The results of habituation in this experiment indicate that the type of stimulus is another variable affecting the habituation rate. Some support for this finding comes from a study (Yaremko, Glanville, & Leckhart, 1972) in which one group of subjects was instructed to imagine they had heard 1000-Hz tones at equal intervals and another group actually heard them before both groups received 10 trials of real tones. The results though not statistically reliable, indicated a difference in the rate of habituation, and the real-tone group tended to have the faster rate.
Primacy Effect The finding of a significant over-all primacy effect in this study supports
PRIMACY EFFECT IN ORIENTING RESPONSES
927
that obtained by Maltzman and his colleagues (1971). Of particular importance and interest, however, are the conditions under which it occurs. When the effect of the length of the adaptation period is eliminated or at least reduced as a variable affecting the primacy effect obtained during s~~bsequent stimulus presentations, it appears that the primacy effect occurs in the brief interval of 5 sec. but not in the extended interval, i.e., 30 sec. Thus, the expectation that the primacy effect would occur in the 5-sec.-same condition and not in the 30-sec.same condition was confirmed. The results of the different conditions in this experiment should have replicated those obtained by Maltzrnan, et a!. (1971). Although there was a primacy effect in both the 5-sec. and 30-sec. groups, only the primacy effect in the 30-sec. group was statistically significant. However, these results may not be in complete agreement because of several methodological differences in the two experiments. Maltman, et al. (1971) used a 10-min. adaptation period compared to 5 min. in this study, and they used lights where this study employed tones and music. The longer adaptation period in the Maltzman, et al. ( 1971) study may have been the more critical variable since the longer temporal interval plus the change in light intensity constituted a greater stimulus-change in the 5-sec. group in that study than the 5-min. temporal interval plus the change in tone (music) in the comparable group in this study. Hence, in that study, larger initial GSRs may have occurred and produced a larger primacy effecc. It should be noted that although the over-all primacy effecc was reliable, the direction of change in stimulus intensity was not statistically significant. This finding indicates that for the over-all primacy effect, at least, the GSR evoked by the first type of stimulus-change, whether the change was from loud to soft or soft to loud, was larger than the GSR evoked by the second change, i.e., change-back from soft to loud and from loud to soft, respectively. This finding agrees with that obtained previously (Maltzman, et al., 1971). However, a careful examination of the results of the triple interaction among the variables of initial intensity, type of stimulus and primacy effect shows that the primacy effecc did not occur regardless of direction of stimulus-change. The only statistically reliable primacy effecc occurred for tones when the direction of change was from soft to loud and then reversed. For tones, there was no primacy effect in the reverse direction, from loud to soft and back again. For music, the means for stimulus-change and change-back were in the direction indicating a primacy effect, but the difference becween these means was not reliable. The interpretation based on these results is that the primacy effect occurs with auditory stimuli. REFERENCES BERLYNB,D. E Conflict, moural, and curiosity. New York: McGraw-Hill, 1960. BERLYNB,D. E.. CRAW, M. A,, SALAPATEK,P. H., & LEWIS,J. Novelry, complexity, incongruity, extrinsic motivation, and the GSR. Iousnal of Experimental Psychology, 1963, 66, 560-567.
928
E. B. SIMON
EDELBERG,R. Electrical properties of the skin. In C. C. Brown (Ed.), Methods in psychophysiology. Baltimore, Md: Williams & Wilkins, 1967. Pp. 1-53. HAGGARD,E A. On the application of analysis of variance to GSR data: the selection of an appropriate measure. JournaL of Experimental Psychology, 1949, 39, 378-392. KINTSCH, W. Habituation of the GSR component of the orienting reflex during pairedassociate learning before and after learning has taken place. JownaI of Mathematical Psychology, 1965, 2, 330-341. LYNN, R . Attention, arousal and the orientajion reaction. New York: Pergamon, 1966. MALTZMAN, I., HARRIS,L., INGRAM,E., & WOLPF, C. A primacy effect in the orientin reflex to stimulus change. Journal of Experimental Psychology, 1971, 87, 202-20t MALTZMAN,I., & RASKIN,D. C. Effects of individual differences in the orienting reflex on conditioning and complex processes. Journal of Experimental Research in Personality, 1965, 1 , 1-16. MALTZMAN,I., & WOLFF, C. Preference for immediate versus delayed noxious stimulation and the concomitant GSR. Journal of Experimental Psychology, 1970, 83, 76-79. S~MON, E. B. Over-all list facilitation and the von Restorff effect with nonnoxious isolation stimuli. Psychological Reports, 1972, 31, 631-637. SOKOLOV,E. N. Perception and the conditioned reflex. New York: Macmillan, 1963. VENABLES,P. H., & MARTIN, I. Skin resistance and skin potential. In P. H. Venables & I. Martin (Eds.), A manual of psychophysiological methods. New York: Wiley, 1967. Pp. 53-102. YAREMKO,R. M., GLANWLLB, B. B., & LECKART,B. T. Imagery-mediated habituation of the orienting reflex. Psychonomic Science, 1972, 27, 204-206.
Accepted February 20, 1976.
PRIMACY EFFECT I N ORIENTING RESPONSES TO AUDITORY STIMULI OF TONES AND MUSIC ELEANOR B. SIMOW California State College, Dominguez H i l h Summary.-This 26 factorial experiment investigated the primacy effect in the orienting response. The rype of stimuli (tone or "music"), stimulus intensities (loud or soft), length of adaptation period (same, 5 or 30 sec.; or different, 5 min.), interstimulus intervals ( 5 or 30 scc.), and sex were studied. College students, 32 males and 32 females were randomly assigned to each group. In the same condition, the tone (or music) was soft (or loud) for 5 sec. (or 30 sec.) in adaptation and was the0 changed alternately without interruption to loud, soft, etc. (or soft, loud, etc.) for 5 sec. (or 30 sec.). The different condition was identical except for the length of the adaptation period in which the stimuli sounded continuously for 5 min. Analyses of the GSR manifestation of the orienting responses indicated: ( a ) an over-all primacy effect with the auditory stimuli and ( b ) the primacy effect occurred in the 5-sec.-same but not in the 30-sec.-same condition as predicted.
Although much of the research concerning the orienting response both in the Soviet Union and in the U. S. has been concerned with its facilitating effects (e.g., Kintsch, 1965; Maltzman & Raskin, 1965; Simon, 1972; Sokolov, 1963) there have been indications (Maltzman, Harris, Ingram, & Wolff, 1971; Maltzman & Wolff, 1970) that it may also have inhibitory effects under certain conditions. This may occur when one orienting response follows another within a short interval (e.g., 5 sec.) producing a larger orienting response to the first stimulus change, the source of the first orienting response, than to the subsequent (second) change. This phenomenon, obtained by Maltzman, Harris, Ingram, and Wolff (1971), was termed the primacy effect. In the Maltzman, et ai. (1971) study, alternating increases and decreases (or the reverse) of light intensity presented continuously for 5-, lo-, or 30-sec. periods to separate groups, found that the mean magnitude of the first orienting response (in terms of the galvanic skin response, GSR) of each pair in the series (onset-offset, onset-offset, etc.) was reliably larger than that of the second. This result was obtained in all groups including the group given 30 sec. between presentations of stimulus-change, a relatively long interval. While a primacy effect was expected in the shorter periods, its appearance in the longer 30-sec. interval mas interesting. In this condition, the GSR, i.e., the orienting response to the first stimulus change, which is known to have a relatively long response latency ranging from approximately 1 to 3.5 sec. (Edel'This research was supported by Research Grant 1R03-MH22037-01 from the U. S. Public Health Service and an award from the Society of Sigma Xi. The author would like to thank Robert Gray for collecting the data and b o t h ~ o b e r rGray and John Douglas for their assisrance in serting up the equipment.
920
E. B. SIMON
berg, 1967; Venables & Martin, 1967), would require continuous GSR activation of an additional 27 sec. at least for it to have an inhibitory effect on the next GSR, the orienting response to the second change. According to Sokolov (1963), the GSR, once activated, returns to baseline level even while the stimulus is still in operation (see p. 116). The records presented in his book (e.g., pp. 58, 117, and 124) show that the return to normal takes place sooner than 30 sec. after the response reaches its lowest resistance level. The question arises then as to what to attribute the occurrence of the primacy effect with a 30-sec. interstimulus interval. One possible explanation may lie in the inadvertent confounding of the two variables, time and light intensity. A compounding of stimulus changes in temporal interval (a stimuluschange) and light intensity (another stimulus-change) could have produced an unexpectedly large GSR (Sokolov, 1963). A careful examination of the data shows that the change away from the initial level of illumination occurred first after a 10-min. adaptation period and again after only 30 sec. This temporal change combined with the simultaneous change in light intensity on the second trial when the direction of illumination change was away from the level experienced in adaptation may have produced a significantly larger GSR than the previous response when illumination was changed back to its initial level. Differences in the GSR for change and change-back conditions on subsequent trials in the 30-sec. group were small. In order to obviate any effect due to temporal changes, and, hence, determine whether the primacy effect in the 30sec. group was in fact due to changes in illumination, the experimental design should include a control condition wherein the duration of the adaptation period is the same as the intervals between presentations of stimulus-change. This experiment was designed to include such a control condition. This experiment also attempted to demonstrate the generality of the primacy effect by employing shifts in auditory stimulation instead of changes in illumination. According to Sokolov (1963) the modality of stimulation is irrelevant. It is any change in stimulation which produces an orienting response. Therefore, results obtained by shifts in auditory stimulation should be similar to those found for shifts in visual stimulation. In summary, the purposes of this experiment were: ( a ) to determine whether the primacy effect in the orienting response occurs when the length of the adaptation period is the same as the duration of stimulus presentations, and ( b ) to extend the generality of the primacy phenomenon from visual to auditory stimuli both simple (pure tones) and complex (music). It was hypothesized that: ( a ) there will be an over-all primacy effect; ( b ) that the primacy effect will occur in the short ( 5-sec.) same condition when the adaptation period and interstimulus interval are both 5 sec., and not i n the long (30-sec.) same condition; and ( c ) the primacy effect will occur in both the 5- and 30-sec. dif-
PRIMACY EFFECT IN ORIENTING RESPONSES
92 1
ferent conditions when the adaptation period is 5 min. and the interstimulus intervals are 5 or 30 sec. The latter finding would constitute a replication of the earlier results with illumination. METHOD Expen'nzental Design The experiment was a 2 X 2 X 2 X 2 X 2 factorial design with the following variables: ( a ) type of stimulus, tone or music; ( b ) length of adaptation period, same ( 5 or 30 sec.) or different ( 5 min.); ( c ) the interstimulus intervals, 5 or 30 sec.; ( d ) initial stimulus intensicy (in adaptation period), loud or soft; and ( e ) sex. The GSR was recorded throughout the experiment. It was measured in terms of the largest resistance change occurring between 0.5 sec. and 5 sec. following a stimulus-change. Sixty-four students from California State College, Dominguez Hills, 32 males and 32 females, served as subjects. These students were in the introductory psychology classes and volunteered to serve to fulfill a course requirement. They were assigned to each group on a predetermined random schedule.
Apprwatw by a Beltone Model 12c The tones, 1000 Hz at 45 d b and 80 db, Audiometer, were prerecorded on magnetic tape using a Wallensak T-3500 tape recorder. Similarly, the music, specifically prepared for this experiment by an electronic synthesizer, was first played and recorded simultaneously for 10 min. continuously at 45 db. Later, this tape was re-recorded such that the intensity of the music was varied at 45 db and 80 db according to the particular experimental treatment. For example, one tape played the music for 5 min. at 45 db then shifted without interruption to 80 db for 5 sec., then to 45 db for 5 sec., etc. for the required number of stimulus-changes. A separate tape was prepared for each of the experimental conditions yielding a total of 16 tapes. The tapes played on the recorder were presented to the subjects via earphones. The recorder was also hooked u p to one of the four channels on the Narco Biosysterns Model 4 Physiograph to mark the time when each stimulus was presented. This made for accurate recording of the GSR intervals during which the responses were read. Palmar GSRs were recorded with circular zinc discs 11 mm in diameter which were embedded 2.5 mm in plastic cups. The cups were filled with Beckman electrode pasre. The output from a Narco Biosystems Preamplifier using - a constant dc current of 20 pa was transmitted to one channel of a fourchannel physiograph recorder.
E. B. SIMON
The experimental room contained a comfortable lounge chair, the arms padded with foam rubber. All of the electronic equipment except for the leads which were attached to the subject were hidden behind a large wooden screen. After the subject entered the experimental room and was seated in the chair, one electrode was attached to each palm of his/her hands. Then, the following instructions were read: These pickups measure changes in the skin which occur to certain stimuli. Since the pickups are sensitive to body movements, it is important that you remain as still as possible, especially your hands. So, please try to get as comfortable as possible before the experiment begins. Are there any questions before we start?
The experimenter then placed the earphones on rhe subject, and went behind the screen to monitor the equipment. First, the polygraph was turned on to obtain the subject's baseline resistance level, then, the adaptation period and subsequent stimulus manipulations followed. The soft (45 d b ) and loud ( 8 0 d b ) 1000-Hz tones and music pre-recorded on tape were presented through the earphones. The first sound with appropriate intensity and duration constituted the adaptation period. The subsequent auditory changes (loud to soft, soft to loud, etc. and vice versa) followed immediately without: interruption. The subjects in the condition in which both lengths of adaptation period and the interstimulus interval were the same heard the sounds for 5 or 30 sec. throughout the experiment. For example, he or she listened to the soft tone for 5 sec. (adaptation period), then to the loud tone for 5 sec., then to the soft one for 5 sec., and so on. The subjects in the "different" condition listened to the loud (or soft) auditory stimulus for 5 min. then to the soft (or loud) auditory stimulus for 5 or 30 sec., and so on. Half the subjects in each group received the soft sound and half the loud sound as the initial stimulus, and half the subjects in each group were males and half were females. All subjects in the 5-sec. interstimulus interval condition received 60 trials of onset and offset, and those in the 30-sec. groups were given 10 such trials in order to equate the amount of time spent in the laboratory.
Before statistical analyses, the GSRs for the first 10 trials after the adaptation period for all groups were subjected to a log conductance change transformation (Haggard, 1949). A mixed analysis of variance with five berweengroups variables (cype of stimulus, adaptation period, stimulus intensity, interstimulus interval, and sex) and one within-groups variable (trials) was performed on these data. The results indicated that the type of stimulus used was reliable ( F 1,3? = 24.3 1, # < .001) . The over-all mean for tones (0.7982 ) was significantly larger than that for music (0.2391). Sex was statistically sig-
PRIMACY EFFECT IN ORIENTING RESPONSES
923
nificant = 8.64, p < .01), with males having a larger over-all mean GSR than females (0.6853 and 0.3520, respectively). The effect of trials was also significant ( F O ,=~ 6.47, ~ ~ p < .001), indicating that habituation over trials had occurred as expected. Fig. l a shows the trend of habituation for the group means over the 10 trials and the trends for tones and music separately over the 10 trials. The rate of habituation for tones was greater than that for music. The interaction between type of stimuli and habituation was statistically significant (p9,2s8 = 4.89, P < .01). It should also be noted that the levels of the GSR for music are far below those for tones.
Primacy Effect TO determine whether the primacy effect occurred, the analysis involved comparison of the mean GSR for Trials 1, 3, 5, 7, and 9, the first of each pair of stimuli in the series, with those for Trials 2, 4, 6, 8, and 10, the second of each pair. This constituted the within-groups variable in a mixed analysis of variance including the five between-groups variables. Again the main effects for type of stimulus = 23.8, p < .001) and sex = 8.6, p < .01) were statistically significant. The mean for tones was larger than for music ( 3.9910 vs 1.21 l l ) , and the mean GSR for males was larger than that for females (3.4420 vs 1.7601).
0.1
5
TRIALS
FIG. la Mean GSRs over 10 trials for music and tones combined, tones only, and music only
4
3
TRIALS
FIG. lb. Mean GSRs for stimuluschange and change-back for all trials. (Means for stimulus-change resulted from Trials 1, 3, 5, 7, and 7, and for change-back, Trials 2, 4, 6, 8, and 10.)
924
E. B. SIMON
The over-all means for the primacy effect were 3.0064 and 2.1956, and the analysis of difference between these means indicated the difference was reliable = 8.1, p < .01). The significance of this result lies in the fact that the GSR to the first stimulus-change in each pair of stimuli is larger than the GSR to the second change, that is, the change-back to the initial level of intensity. Fig. l b shows that mean GSRs for stimulus changes are clearly larger than the GSRs for change-back over all trials. A most interesting result of the analyses was the statistically significant triple interaction among the variables of the adaptation period (same or different), length of the interstimulus interval ( 5 or 30 sec.), and the primacy effect = 6.2, p < .05). Considering the condition in which the length of the adaptation period was the same as the interstimulus interval, the primacy effect occurred when the interval was 5 sec. in length but did not occur in the 30-sec. interstimulus interval. The mean GSRs for these conditions appear in Table 1. The analysis of the difference between the means for the 5-set.-same condition was reliable (FIq8= 13.58, p < .01). When the length of the adaptation period is different ( 5 min.) from the interstimulus interval, the means for both the 5- and 30-sec. conditions indicate the appearance of the primacy effect as expected (Maltzman, et d., 1971). However, only the result for the 30-sec. group was statistically significant (Fl,s = 7.40, p < .05). TABLE 1
Adaptation Period Different: Intentimulus Interval Same: Interstimulus Interval Primacy Effect Change Change-back
Figs. lc, d, e, and f show the primacy effect separately for the 5-sec.-same, 30-set.-same, 5-sec.-different, and 30-sec.-different conditions. The relationship is clearly evident in the 5-sec.-same condition compared with the 30-sec.-same condition. The reverse occurs in the different conditions, with the primacy effect shown in the 30-sec. but not in the 5-sec. condition. The results of the analysis of variance also indicated that the interaction between the initial stimulus intensity and the primacy effect war reliable (F1,32 = 24.2, p < .01) as was the triple interaction among type of stimulus and these variables (F1.32 = 10.2, p < .01). For tones, when the first stimuluschange after the adaptation period was from soft to loud and the change-back was from loud to soft, there is a significant primacy effect. The means for stimulus-change and change-back are 5.91 50 and 2.582 3, respectively. When
PRIMACY EFFECT I N ORIENTING RESPONSES
0.1 1
2
3
4
6
TRIALS
FIG. lc. Mean GSRs for stimuluschange and change-back for all trials i n the 5-sec-same groups
TRIALS
FIG. l d . Mean GSRs for stimuluschange and change-back for all trials in the 30-sec.-same groups
1
TRIALS
FIG. le. Mean GSRs for stirnuluschange and change-back for all trials in the 5-sec.-different groups
2
3
4
5
TRIALS
FIG. l f . Mean GSRs for stimuluschange and change-back for all trials in the 30-sec.-different groups
926
E. B. SIMON
the intensity changes were in the reverse directions, from loud to soft and soft to loud, the primacy effect did not occur. The means for stimulus-change and change-back were 3.0922 and 4.3742, respectively. For music, the means for stimulus-change are larger than the means for change-back in both situations: soft to loud, loud to soft, and vice versa. The differences, however, were too small to be significant.
DISCUSSION Habituation The habituation curves for the two types of stimuli differed in two respects: ( a ) the magnitude of the GSRs for the 1000-Hz tones was significantly larger than that for music even though the intensities and durations of these stimuli were the same for both, and ( b ) the rate of habituation for tones was more rapid than that for music. According to Sokolov (1963) and Lynn (1966), the magnitude of the GSR varies with the intensity of the stimulus; for moderate intensities, the greater the intensity the larger the magnitude of GSR. In this study, the size of the GSR should have been the same for tones and music since their intensities were the same. If there was a difference, musical stimuli being more complex than tones should have elicited larger GSRs (Berlyne, 1960; Berlyne, Craw, Salapatek, & Lewis, 1963). However, since pure tones are seldom encountered in real life experiences, whereas music (even the metallic sounding music produced on the electronic synthesizer) is heard quite frequently, it may be that the familiarity of the music resulted in reducing the magnitude of the GSRs. During the experiment, many of the subjects in the music condition sat back comfortably in the chair, closed their eyes and seemed to enjoy the music. Actually, the music was an original composition composed extemporaneously by a graduate student in music for the sole purpose of this study. It had no regular rhythm nor apparent melody. Nevertheless, the college students accepted it as real music (perhaps rock music?). Lynn (1966) has summarized the results of studies indicating that the rate of habituation differs as a function of the stimulus intensities, stimulus durations, and other variables. The results of habituation in this experiment indicate that the type of stimulus is another variable affecting the habituation rate. Some support for this finding comes from a study (Yaremko, Glanville, & Leckhart, 1972) in which one group of subjects was instructed to imagine they had heard 1000-Hz tones at equal intervals and another group actually heard them before both groups received 10 trials of real tones. The results though not statistically reliable, indicated a difference in the rate of habituation, and the real-tone group tended to have the faster rate.
Primacy Effect The finding of a significant over-all primacy effect in this study supports
PRIMACY EFFECT IN ORIENTING RESPONSES
927
that obtained by Maltzman and his colleagues (1971). Of particular importance and interest, however, are the conditions under which it occurs. When the effect of the length of the adaptation period is eliminated or at least reduced as a variable affecting the primacy effect obtained during s~~bsequent stimulus presentations, it appears that the primacy effect occurs in the brief interval of 5 sec. but not in the extended interval, i.e., 30 sec. Thus, the expectation that the primacy effect would occur in the 5-sec.-same condition and not in the 30-sec.same condition was confirmed. The results of the different conditions in this experiment should have replicated those obtained by Maltzrnan, et a!. (1971). Although there was a primacy effect in both the 5-sec. and 30-sec. groups, only the primacy effect in the 30-sec. group was statistically significant. However, these results may not be in complete agreement because of several methodological differences in the two experiments. Maltman, et al. (1971) used a 10-min. adaptation period compared to 5 min. in this study, and they used lights where this study employed tones and music. The longer adaptation period in the Maltzman, et al. ( 1971) study may have been the more critical variable since the longer temporal interval plus the change in light intensity constituted a greater stimulus-change in the 5-sec. group in that study than the 5-min. temporal interval plus the change in tone (music) in the comparable group in this study. Hence, in that study, larger initial GSRs may have occurred and produced a larger primacy effecc. It should be noted that although the over-all primacy effecc was reliable, the direction of change in stimulus intensity was not statistically significant. This finding indicates that for the over-all primacy effect, at least, the GSR evoked by the first type of stimulus-change, whether the change was from loud to soft or soft to loud, was larger than the GSR evoked by the second change, i.e., change-back from soft to loud and from loud to soft, respectively. This finding agrees with that obtained previously (Maltzman, et al., 1971). However, a careful examination of the results of the triple interaction among the variables of initial intensity, type of stimulus and primacy effect shows that the primacy effecc did not occur regardless of direction of stimulus-change. The only statistically reliable primacy effecc occurred for tones when the direction of change was from soft to loud and then reversed. For tones, there was no primacy effect in the reverse direction, from loud to soft and back again. For music, the means for stimulus-change and change-back were in the direction indicating a primacy effect, but the difference becween these means was not reliable. The interpretation based on these results is that the primacy effect occurs with auditory stimuli. REFERENCES BERLYNB,D. E Conflict, moural, and curiosity. New York: McGraw-Hill, 1960. BERLYNB,D. E.. CRAW, M. A,, SALAPATEK,P. H., & LEWIS,J. Novelry, complexity, incongruity, extrinsic motivation, and the GSR. Iousnal of Experimental Psychology, 1963, 66, 560-567.
928
E. B. SIMON
EDELBERG,R. Electrical properties of the skin. In C. C. Brown (Ed.), Methods in psychophysiology. Baltimore, Md: Williams & Wilkins, 1967. Pp. 1-53. HAGGARD,E A. On the application of analysis of variance to GSR data: the selection of an appropriate measure. JournaL of Experimental Psychology, 1949, 39, 378-392. KINTSCH, W. Habituation of the GSR component of the orienting reflex during pairedassociate learning before and after learning has taken place. JownaI of Mathematical Psychology, 1965, 2, 330-341. LYNN, R . Attention, arousal and the orientajion reaction. New York: Pergamon, 1966. MALTZMAN, I., HARRIS,L., INGRAM,E., & WOLPF, C. A primacy effect in the orientin reflex to stimulus change. Journal of Experimental Psychology, 1971, 87, 202-20t MALTZMAN,I., & RASKIN,D. C. Effects of individual differences in the orienting reflex on conditioning and complex processes. Journal of Experimental Research in Personality, 1965, 1 , 1-16. MALTZMAN,I., & WOLFF, C. Preference for immediate versus delayed noxious stimulation and the concomitant GSR. Journal of Experimental Psychology, 1970, 83, 76-79. S~MON, E. B. Over-all list facilitation and the von Restorff effect with nonnoxious isolation stimuli. Psychological Reports, 1972, 31, 631-637. SOKOLOV,E. N. Perception and the conditioned reflex. New York: Macmillan, 1963. VENABLES,P. H., & MARTIN, I. Skin resistance and skin potential. In P. H. Venables & I. Martin (Eds.), A manual of psychophysiological methods. New York: Wiley, 1967. Pp. 53-102. YAREMKO,R. M., GLANWLLB, B. B., & LECKART,B. T. Imagery-mediated habituation of the orienting reflex. Psychonomic Science, 1972, 27, 204-206.
Accepted February 20, 1976.
