J BUSN RES 1989:18:89-l%
89
Temporal Instability as a Moderating Factor on Advertising Effectiveness Jacob Hornik Tel Aviv
University
Hypotheses that diurnal variation in advertising exposure will influence consumers’ immediate and delayed memory of emotional and rational television commercials were tested. The experiment employed a forced-exposure approach using 287 subjects recruited randomly from mall shoppers. Subjects were exposed to different formats of television commercials at three different times of the day. The experiment also varied levels of ad repetition. Analysis of recall and recognition revealed time-of-day X memory task interactions, such that both recall and recognition were greater following early-in-the-day exposures for immediate tests, especially for emotional ads. Scores for delayed tests were greater for the recognition measures following late-in-the-day exposures, particularly for rational appeals. The arousal theory in long-term memory is offered to explain the results. Implications for advertising research and testing are discussed. Introduction Marketers are frequently concerned with the effect of their advertisements. The effect of advertising depends on several factors: the quality of the ad and the media, the characteristics of the potential buyers, situational factors such as the sequential order of the ads, the amount of time between ads, and momentary consumer moods and other circumstances that are not under the direct control of the advertiser. The study of the effects of advertisements has generated a large and still-growing body of literature. Much of this research is concerned with the influence of ad format, media factors, and individual characteristics on consumer’s response (Bagozzi and Silk, 1983). But most of the important testing and diagnostic procedures employed by advertising researchers ignore possible situational moderating factors (e.g., Green and Schaffer, 1983; Burke and Edell, 1986). Exceptions are more recent attempts to examine advertising in situational and temporal contexts. For example, Aaker et al. (1986) examined sequence effects-how feelings created by prior commercials influence perception of subsequent ads. Batra and Ray (1986) explored moods effects on postexposure brand attitudes. Temporal and other sit-
Address correspondence to Jacob Hornik, The Leon Recanati istration, Tel Aviv University, Tel Aviv, Israel. Journal of BusinessResearch 18.89-106 (1989) 8 1989 Elsetier Science PublishingCo., Inc. 655 Avenue of the Americas, New York, NY 10010
Graduate
School of Business
Admin-
0148-2%3/89/$3.50
90
J BUSN RES 1989:18:89-106
J. Hornik
uational issues are, in the view of Schmalensee (1983), “ . . . among today’s top priority advertising research questions by top advertisers and agencies” (p. 49). In part, this emerging interest has come with the growing recognition that static models are too far removed from reality to explain advertising influences adequately. For copy-testing purposes, measures of ad effectiveness, such as attitude change, recall, and recognition all require an explicit consideration of the impact of momentary and temporal conditions. There are many examples in the social and behavioral sciences in which the predictive validity of some measures varies systematically in accordance with various situational independent variables. For example, the amount of time between stimuli and the time of the day that they are received may be important factors in determining individual response (Mehrabian and Russell, 1974). If such factors are significant and uncontrolled, they may reduce the reliability and validity of test results.
Objective To incorporate situational variables in advertising models and measures, one needs two types of research: first, the development of a situational inventory (Kakkar and Lutz, 1981); and, second, the testing of specific situational cues, such as timeof-day effect on consumer response and behavior (Schlinger, 1982). This paper is intended to focus on one situational variable-time of day, and to determine its influence on immediate and delayed consumer response to different advertising formats (factual versus emotional appeals), and different levels of advertising repetition. These will expand Hornik’s (1988) recent note that variance in consumer response is associated with temporal situations. The paper is organized as follows: First, we present a theoretical framework of diurnal variation in response to various stimuli. Special attention is devoted to the conceptual difference between immediate and delayed response and their association to short-term memory (STM) and long-term memory (LTM). This will be followed by a description of an empirical study and results. In the concluding sections, findings are summarized and recommendations are discussed.
Theoretical
Framework
to a time and Belk (1975) defined situations as “ . . . all those factors particular place of observation which do not follow from a knowledge of personal (intraindividual) and a stimulus (choice alternative) attributes and which have a demonstrable and systematic effect on current behavior” (p. 175). Since behavior occurs within the context of varied sets of circumstances, one may assume that these circumstances are an important source of variations in consumer response. The role of situational variable has been acknowledged in a growing number of studies in consumer behavior (e.g., Hornik, 1982). These studies rely on the tradition of environmental psychology (Mehrabian and Russell, 1974), which is concerned with such stimulus issues as the effects of noise, temperature, mood, time of day, and other specific variables on individual response and behavior.
Temporal
Instability
as a Moderating
Factor
.I BUSN RES 1989:18:X9-106
Time of Day and Response Behavior Various pieces of evidence bearing on the relationship between diurnal fluctuation in stimuli and individual response and learning can be found in the behavioral science literature. Interest in the effects of time of day on learning dates back to the beginning of this century when practical considerations concerning arrangements of school timetables led educational researchers to investigate the best times to teach different subjects (e.g., Winch, 1913). Recent studies of STM and LTM tasks involving immediate and delayed recall of educational material have been the main source of knowledge available to educations and psychologists regarding how time of day of instruction is likely to influence achievement (e.g., Folkard et al., 1977). Generally, these studies have found that STM performance is better in the morning, while LTM performance is better in the late afternoon or early evening. Basing their arguments on studies involving recall of information presented in various forms, some researchers have stated that afternoon presentation results in better LTM performance at any time of day (Jarrett and Furnham, 1983). Theories that support the findings of time-of-day and recall studies posit that physiological processes involving factors such as basal arousal and diurnal rhythms in plasma hormone level are basic determinants (Eysenck, 1976). It was shown that arousal rises from a low level in the early morning to a peak in the evening, usually operationalized in terms of oral temperature. In addition, arousal during initial learning has consistently been found to result in superior retention when testing is delayed for at least 30 minutes (Craik and Blankstein, 1975). Blake (1971) found responses on a range of stimuli not involving STM to increase over the day and to parallel fairly closely the circadian rhythm in body temperature. He also found digit span, a “classic” test of STM, to show virtually the opposite time-of-day effect and to be highest in the morning. Similarly, Folkard et al. (1977) found that immediate memory for either a passage of prose or for a short film sequence was better among subjects who received the learning material at 8 A.M. than for those who read or saw it at 8 P.M. On tests of delayed memory, however, performance is typically found to improve rather than to deteriorate later in the day. They also found better recall for a story after a delay of a few hours among children who heard it at 3 P.M. than among those who heard it at 9 A.M. Finally, Craik and Blankstein (1975), In . an extensive review of the memory and arousal literature, concluded that the beneficial effect of arousal on LTM has been found rather more consistently than its detrimental effect on STM. In view of this, it might be expected that LTM should be better following afternoon, rather than morning, presentation.
Message Format From the perspective of cognitive theory, a message format might affect immediate and delayed processing of information. There is growing evidence, e.g., that emotional ads may involve a considerable reliance on STM (Choi and Thorson, 1983), particularly when semantic coding of information can be avoided. Furthermore, based on the demonstration of Pallak et al. (1985), it can be argued that responses to rational persuasive messages are based on systematic processing and tend to be more enduring, whereas responses to emotional messages are based on heuristic
92
J BUSN RES 1989:18:89-106
processing, which are less enduring. On this subject, Krugman (1986) also raised the concern that delayed-recall test may penalize “emotional” (“feeling”) ads as opposed to “thinking” ads. This, according to Krugman, is because current techniques that require verbalization favor thinking or “left-brain advertising.” Recent research has made an attempt to evaluate the relative effects of ad format, especially “rational” (or straight sell) versus “emotional” (mood or warm) commercials. “Warmth” is defined by Aaker et al. (1986) as “a positive, mild, volatile emotion involving psychological arousal and precipitated by experiencing directly or vicariously a love, family, or friendship relationship” (p. 366). Aaker and Bruzzone (1981) found that the warmth factor was associated with commercials that utilized sentimental feelings in the creative approaches. Along these lines, several advertising researchers, particularly the ones involved in emotional research, argue for multiple exposure testing in consumer behavior (e.g., Aaker et al., 1986). This argument is often based on the notion that single exposure is systematically biased against “emotion” commercials, and that multiple exposure will allow their more emotional appeal to sink in. It is important to realize that the emotion lies in the viewer and not in the commercial. In other words, an emotional commercial is one that is capable of creating an emotional response in the viewer. Hypotheses Based on a review of the literature,
the following hypotheses
are proposed:
H,: Immediate response to television commercials will produce significantly higher scores in morning exposures relative to early-afternoon and evening exposures. H,: Emotional ads will produce significantly higher immediate scores than rational ads. H,: Delayed response to late (in-the-day) television commercials will be significantly higher than delayed scores to early (in-the-day) exposures. H,: Rational ads will produce significantly higher delayed scores than emotional ads. H,: Delayed response score differences between rational and emotional commercials will be smaller as the number of exposures increases (repetition effect). H,: Repetition will lead to continuing gains in immediate response scores for emotional ads over rational ads.
Method Overview In disguised form, the experimental study involved a forced-exposure analysis of six different television commercials. The forced-exposure approach has shown greater reliability than on-air procedures (Clancy and Ostlund, 1976). A total 287 subjects (118 men and 169 women) was recruited randomly from mall shoppers
Temporal Instability
as a Moderating
J BUSN RES
Factor
93
1989:18:89-106
Table 1. Experimental Design Time of day Day 1 9 A.M. 1 P.M. 5 P.M. Day 2 9 A.M. 1 P.M. 5 P.M. Day 3 9 A.M. 1 P.M. 5 P.M. Total
Immediate Test
Delayed Test
Total
17 19 18
14 15 13
31 34 31
15 18 20
15 15 15
30 33 35
17 18 17
15 14 12
32 32 29
159
128
287
and screened for age (adults). The study was performed in conjunction with a study of people’s reactions to different television formats in a suburban Chicago shopping mall. The sample size exceeded the minimum requirement under Feldt-Mahmond (1958) specifications. Subjects were each offered a gift for participating in the session. Experiments were conducted by professional experimenters three times a day for three consecutive days, producing nine exposure groups ranging in sizes from 29 to 35 (See Table 1 for experimental design.) The forced-exposure approach involved having respondents watch a 50-minute videotape of four complete television segments, two from “60 Minutes” and two from “20/20,” each segment being about 10 minutes long. In an attempt to preclude any effects of prior exposure, six 30-second television commercials selected from local New York City television commercials were interpersed in the clips. These were new advertisements for the following different East Coast products or brands: beer, car dealer, ice cream, airline, house paint, and bank.
Design A 3 x 2 x 2 x 3 between-subjects design was used with time of day (9 A.M., 1 P.M., 5 P.M.), memory task (immediate or delayed), advertising format (rational or emotional), and repetition (two, three, or five exposures) as factors. Time of Day. Subjects were exposed to the television commercials on one of three occasions: in the morning (between 9 and 10 A.M.), in the afternoon (between 1 and 2 P.M.), and in the early evening (between 5 and 6 P.M.). Some subjects were tested immediately following the videotape presentation and some 2 hours after presentation. Repetition. Commercials appeared two, three, or five times. In a method that followed Krugman’s (1972) argument that the first exposure to an ad results in an orienting response and that the acquisition of information from an ad occurs primarily on the second exposure, the commercials were shown to subjects at least twice. Three videotapes were produced. All tapes were identical except for the
94
J BUSN RES 1989:18:89-106
sequence of ad exposures. and recency effects.
J. Homik This was done in order to control
for possible
sequence
Advertising Format. A panel of seven creative managers and senior account executives followed the procedure of Golden and Johnson (1983) and reviewed 36 commercials, identified the appeals, and ranked the ones they judged to be wellexecuted, rational ads and the ones they judged to be well-executed, emotional commercials. The three highest-ranking emotional and the three highest-ranking rational commercials were used as the experimental stimuli. The alpha coefficients for the ranking ranged from 0.83 to 0.92, indicating a high level of interjudge reliability.
Procedure At the beginning of each session, subjects were told that they would be shown excerpts from two television programs, which they would be required to evaluate. The experiment was described as one in which university researchers are concerned with the measurement of program appreciation and how it relates to content. The testing room was equipped with standard 314” videotape playback equipment, so that the communications could be viewed on standard TV monitors. Respondents were told that the videotaped programs they would be watching were actual programs already broadcast, and, since the researchers were interested in finding out how people felt about the programs in conditions as similar as possible to watching them at home, the tape would be shown to them complete with the announcements and commercials that would have been seen in home viewing. Before they turned to the paper and pencil tests, subjects were asked the following question: “Who is going to spend the next 2 hours shopping in the mall?” Of those who responded, the experimenter selected (in each session) about 15 subjects and asked them kindly to leave the room and return in exactly 1 hour and 45 minutes for a lo-minute session. They were led to believe that this was done because the “room is too crowded.”
Measures Subjects viewed the 50-minute videotape and then answered two short questionnaires on the cover story and commercials. The main dependent measure was subjects’ recall. Investigation of recall of television advertisements dates back to the very beginning of advertising research and remains the basic and most widely used source of information about consumer response to advertising in electronic media (Batra and Ray, 1986). Recall measures, however, have recently come under criticism (Gibson, 1983). The practical result has been a move to augment recall measures with other measures of consumer response, such as recognition, purchase intent, or brand use (Krugman, 1986; Petty et al., 1983). The first questionnaire was a free (unaided) recall test in which subjects were told to recall spontaneously as many as they could of the products and brand names advertised in the session. Three minutes were allowed for completion of this task. On the second form, subjects were given descriptions of the six product cate-
Temporal
Instability
as a Moderating
Factor
J BUSN RES 1989:18:89-106
95
gories and asked to select the correct brand names and selling points from a list of choices provided. Half of the items were correct, and half were incorrect. Recognition was coded for analysis as items checked correctly and adjusted for items checked incorrectly. Scores were corrected for guessing by using the Leigh and Menon (1986) procedure. Two minutes were allowed for this recognition test. Purchase intentions were operationalized by using a six-point scale anchored by the phrases “not at all likely to buy” and “very likely to buy.” As a check for possible effects of prior exposure to the commercials, a question was added to determine whether subjects recalled having seen the commercial elsewhere.
Results Of the 287 respondents who participated in the experiment, two subjects (in different sessions) indicated that they might have seen some of the ads before. Of the 128 subjects who were expected to return to the second session, 122 did so. The analysis, then, was based on the remaining 279 respondents. Since preliminary analysis indicated that there were no significant sex differences across treatments, this factor was ignored in the subsequent analysis. The mean immediate and delayed response (undaided recall, recognition, and purchase intention) scores for each experimental condition are presented in Table 2 (for rational ads) and Table 3 (for emotional ads).
Unaided Recall The ANOVA for the response variables are presented in Table 4, graphs of these data are presented in Figures l-4. The ANOVA results provide the necessary information to test the hypotheses of this study. H, states that immediate recall following ad exposures deteriorates across the day. Analysis of nonrepeated exposures shows a significant effect of time of day (F(2,58) = 5.86, p < O.Ol), on immediate recall. But there was no significant difference between 9 A.M. and 5 P.M. recall on five exposures (F(1,58) = 2.17, ns). Independent F tests indicate that late-day exposures led to higher but marginally significant delayed recall scores (F(2,117) = 3.22, p < O.l), thus directionally confirming H,. The ANOVA also confirms the presence of a significant interaction between time of exposure and recall test (F(2,246) = 6.28, p < 0.01). The general form of H, and H, would be confirmed by a significant two-way interaction effect of advertising format and recall (M x A). Variations in this effect for high and low repetition, H5 and He, would be indicated by a significant threeway interaction involving the repetition factor (M x R x A). As can be observed from Table 4, only recall and ad format interaction was statistically significant. Finally, as exposure frequency increases, immediate and delayed recall increases [F(2,246) = 4.85, p < 0.051; and [F(2,246) = 7.31, p < O.OOl] respectively. These findings replicate the results of previous studies and support the hypothesized this effect was not influenced by repetition effect on receiver recall. However, either time of day (T x R) or recall (M x R).
2.9 2.3 4.6’ 2.5 2.5 1.7
(n = 54) (n = 41)
(n = 54) (n = 37)
(1389)
(L93)
(L91)
3.0 2.2 6.1b
(n = 49) (n = 44)
2
bpi 0.05.
5
3.8 3.7 1.0 3.6 3.9 2.2
3.1 2.9 1.5 2.9 3.0 2.0
3.4 2.9 4.7b
3
3.9 3.6 2.7
Response
Unaided Recall (No. of Exposures)
and Delayed
‘Degrees of freedom are given in parentheses.
9 A.M. Immediate Delayed F-Value 1 P.M. Immediate Delayed F-Value 5 P.M. Immediate Delayed F-Value
Time of Day”
Table 2. Mean Scores for Immediate Repetition
Measures
3.9 4.9 5.2’
4.2 4.3 1.0
4.3 3.9 4.9b
2
4.3 4.9 4.4b
4.4 4.4 1.0
4.4 4.2 1.6
3
4.4 5.1 4.7h
4.6 4.7 2.6
4.6 4.4 1.6
5
Ads as a Function
Recognition (No. of Exposures)
on Rational
2.5 2.4 1.5
2.8 2.5 1.7
2.9 2.6 1.2
2
Exposure
2.7 2.5 1.0
2.9 2.6 2.2
3.0 2.7 2.4
3
PI (No. of Exposures)
of Time-of-Day
2.9 2.9 1.0
3.1 2.9 1.4
3.1 2.9 1.0
5
and
5
1
9
2.7 2.3 4.1b
‘p < 0.01.
3.0 2.8 2.1
3.2 2.6 5.36
3.7 2.4 6.7b
3.9 3.9 1.0
4.1 3.8 1.7
4.1 3.8 2.7
4.2 4.6 2.6
4.9 4.3 5.0b
5.1 3.8 10.3‘
4.3 4.8 2.5
4.9 4.4 2.8
5.1 4.1 7.9’
3
2
3 5
Recognition (No. of Exposures)
4.6 4.9 2.2
5.1 4.7 1.6
5.3 4.3 5.2’
5
2.8 2.8 1.4
2.9 2.8 1.0
3.0 2.8 1.5
2
2.9 2.9 1.0
3.0 3.1 1.0
3.1 2.8 1.3
3
3.0 3.0 1.0
3.1 3.1 1.0
3.2 3.1 1.0
5
Exposure and
PI (No. of Exposures)
Ads as a Function of Time-of-Day
Unaided Recall (No. of Exposures)
3.0 2.2 4.7b
3.4 1.9 8.1’
2
and Delayed Response Measures on Emotional
“Degrees of freedom are given in parentheses. “p i 0.05.
Immediate (n = 54) Delayed (n = 37) F-Value (1989)
P.M.
Immediate (n = 54) Delayed (n = 41) F-Value (L93)
P.M.
Immediate (n = 49) Delayed (n = 44) F-Value (L91)
A.M.
Time of Day”
Repetition
Table 3. Mean Scores for Immediate
98
.I BUSN RES
J. Hornik
1989:18:89-106
Table 4. Analysis of Variance for Response Measures” RCGill
Source Time of Day Memory Task Repetition Ad Format TXM TXR TXA MXR MXA RXA TXMXR TXMX MXRXA TXMXRXA
(T) (M) (R) (A)
Recognition
df.
ss
F
ss
F
2 1 2 1 2 2 2 1 1 2 2 2 2 2
6.4 3.1 4.8 2.5 9.4 3.3 2.x 4. I 2.x 2.4 4.3 4.0 4.9 5.1
4.iih 3. I 4.3” 1.7 6.3’ 2.4 1.9 2.4 3.9” 1.7 3.3 2.3 2.x 2.1
5.8
4.‘!lh 4.2h 4.0” 3.2 8.8’ 2.4 2.6 2.3 4.1* 1.5 2.9 2.3 2.9 2.8
3.3 3.0 2.6 10.2 5.1 6.1 4.s 7.2 3.s 5,s 3.6 4.1 6.4
“All PI where not significant and therefore omitted from the table. “p < 0.05. ‘p -c 0.01.
Recognition Overall, recognition measures revealed similar patterns to those found in the timeof-day effect on unaided recall scores. Hence, the hypothesized time-of-day effect on recognition was supported [F(2,246) = 4.37,p < 0.051.The apparent tendency for the delayed recognition scores to be higher at 5 P.M. were also significant [F(1,117) = 8.41, p < 0.011. The main effect for time of day supported H, and H,, that the ads would receive better scores in the immediate recognition test early in the day, where the opposite measures would reveal a late-day effect on delayed recognition (See Figs. 3 and 4 for a graphic representation). The data of Table 4 demonstrate a significant two-way interaction involving advertising format and recognition. Furthermore, the nature of this interaction conforms precisely with the pattern predicted in H, and H,. Under immediate recognition, emotional ads outperformed rational ads (p < 0.01). For a delayed recognition test, a difference exists between the two: rational ads received higher delayed recognition scores 07 < .OS). The three-way interaction relevant to the impact of repetition (M x R x A) approached significance [F(2,246) = 2.93, p = 0.0861.
Purchase Intentions As can be observed in Tables 2-4, subjects’ purchase intention was not affected by time of day, memory task, and ad format. The only notable results from Table 2 and 3 are that the mean immediate and delayed intention scores increase with advertising repetition. Although these were not statistically significant, they are in the predictive direction and reflect some repetition effect on attitudes.
J BUSN RES 1989:18:89-106
99
--
____-_----m--e
-_--w--------e--
c--
g .d
I
1
2 Exposures
9
5
I9 I I
I
I
I
I
I
I I
3 Exposures
1
5
I I I
I I
I
I
I I I I I
1
X
5 Time day rate of exposure.
5 Exposures
9
Figure 2. Time-of-day effect on immediate and delayed recall for rational ads
Mean Recall
of
J BUSN RES 1989:18:89-106
D--_---------m--
\
\
\
\ \ \ \ \
\
2
-3’ %\
2
z z2 z
%
/
P
d
m
rr;
0
In
0
VI
Y;
;
;
m’
0
VI
c; c;
0
h;
101
2.0
2.5
3.0
3.5
sss 1
Figure 4. Time-of-day
2 Exposures
9
I9 I I
effect on immediate
5
I I I I I I I I
and delayed
5
recognition
3 Exposures
1
1
ads
x
day
5 Time of
rate of exposure.
5 Exposures
9
for rational
I I I
I I I I I I I I
Temporal Instability
as a Moderating
Factor
J BUSN RES 1989:18:89-106
103
Discu .ssion This research provides a contribution to the growing empirical evidence on situational influences on advertising effectiveness. The meanings of the results are clear: a consumer’s response is highly dependent on time of exposure. It appears, therefore, that time-of-day effects on advertising performance are far more complex and task specific than was recognized previously. The results of the immediate recall tests appear to support the findings of earlier studies in education and psychology that immediate recall is superior in the morning. They can also be considered to extend the findings of at least the more recent studies (e.g., Hornik, 1988) that this superiority was found when emotional rather than rational advertisements were used. The presence of an interaction between time of day of exposure and delay of recall is entirely consistent with previous research findings with nonadvertising materials (e.g., Folkard et al., 1976, 1977). So, too, is the superiority of delayed recognition following afternoon exposure. These variations in memory performance have been explained in terms of changes in arousal across the day, with a higher level of arousal known to be beneficial particularly to delayed retrieval of factual information (e.g., Craik and Blankstein, 1975). Thus, the findings provide additional evidence that delayed recall and delayed recognition measures penalize emotional ads relative to rational ads. Results indicate that presentation of emotional commercials during the morning will result in better immediate retention, but in worse delayed retention, than presentation in the late afternoon. Since purchases are rarely made just after advertisement is seen, most commercials operate on the premise of delayed recall. Given that rational ads are more temporally stable than emotional ads, it is suggested that commercials containing evaluative material should be aired in the late afternoon and early evening. Finally, the results also support other empirical findings showing advertising repetition effect on recall and recognition. Although the repetition effect was not statistically significant, it was directionally consistent with intention to buy.
Limitations As with any controlled experiment, the current study necessarily implies some compromise of external validity for the sake of increased measurement control. It should be emphasized, therefore, that these findings must be viewed in proper perspective. Only six commercials and two advertising formats were used as experimental stimuli. There may be other appeals that might interact with temporal factors. Also, certain key variables relevant to time-of-day behavior may have been omitted from the design. For example, there is some evidence that key personality variables might explain diurnal fluctuation in human performance (e.g., Gale et al., 1972). Third, our main response measures included only three of the many aspects of memory and behavior involved in the experience of consumer perception. Moreover, these verbal measures failed to account adequately for such important nonverbal components as psychobiological arousal of brain activity, physical movement, or subliminal perception (Golden and Johnson, 1983).
104
JBUSN
RES 1989:18:89-106
J. Hornik
A fourth possible limitation of this study is that no attempt was made to investigate whether delayed response is affected by time-of-day test regardless of time of exposure. However, studies in other disciplines (e.g., Folkard et al., 1977) provide no evidence for such an interaction. Nevertheless, future studies should attempt to further validate these results within the advertising context. Recommendations As the above-mentioned limitations indicate, the study had no pretention to achieve a complete description of time-of-day effect on advertising measurements. Nevertheless, these results have both theoretical and practical values. First, it seems inevitable that advertising effects should be conceptualized as a process occuring over time, in which direct advertising effects interact with effects of other elements of the marketing mix and also with the impact of situational factors such as time of exposure. Second, there is growing evidence that advertising effect measures have been shown to depend heavily on the research design (Clancy and Ostlund, 1976). The variability in results are partially due to situational factors during the experimental procedure. Moreover, there might be some danger in using reliability tests, such as test-retest correlations, in advertising experiments. Silk (1977) has argued that test-retest conditions must be equivalent. If, e.g., consumers have been exposed to the advertising in different contexts between the two testing occasions or at different times during the day, the testing conditions may not be equivalent, and a low test-retest correlation may not signify low reliability. Testing and validating procedures that cannot control for these error sources as a part of the research design or, ex post facto, through covariance or related statistical analysis are doomed to produce quasi-random results. Practically, the present data argue against using measures of brand name recall as the sole indicator of advertising effectiveness. More important, these finding show that temporal exposure circumstances serve as a generative cue for consumer’s response. The knowledge of these circumstances will enable advertiers to schedule advertisements at optimum response times. Furthermore, it implies that in gathering and analyzing advertising data, it is important to control for such temporal variables. References Aaker, David A., Stayman, Douglas M., and Hagerty, Michael R., Warmth in Advertising: Measurement, Impact, and Sequence Effects, Journal of Consumer Research 12 (1986): 368-381.
Aaker, David A., and Bruzzone, Donald E., Viewer Perceptions of Prime-Time Television Advertising, Journal of Advertising Research 21 (October 1981): 15-23. Bagozzi, Richard P., and J. Silk, Alvin, Recall, Recognition, and the Measurement of Memory for Print Advertising, Marketing SCience 2 (1983): 95-134. Batra, Rajeev, and Ray, Michael L., Situational Effects of Advertising Repetition: The Moderating Influence of Motivation, Ability and Opportunity to Respond, Journal of Consumer Research 12 (1986): 432-445.
Belk, Russell, W., Situational Research 2 (1975): 157-164.
Variables
and Consumer
Behavior, Journal of Consumer
Temporal Instability
as a Moderating
J BUSN RES 1989:18:89-106
Factor
Blake, M. J., Temperament and Time of Day, in Biological Rhythms and Human Behavior. W. P. Colguhoun, ed., Academic Press, New York, 1971. Burke, Marian C., and Edell, Julie A., Ad Reactions Over Time: Capturing Changes in the Real World, Journal of Consumer Research 13 (1986): 114-118. Choi, Y. and Thorson, E., Memory for Factual, Emotional and Balanced Ads Under Two Instructional Sets, in Proceedings of the 1983 Conference of the American Academy of Advertising. A. D. Fletcher, ed., University of Tennessee, Knoxville, Tenn, 1983. Clancy, Kevin, J., and Ostlund, Lyman E., Commercials of Advertising
Effectiveness
Measures, Journal
Research 16 (1976): 29-34.
Craik F. I., and Blankstein, K. R., Psychophysiology and Human Memory, in Research in Psychophysiology. P. H. Venables and M. J. Christie, eds., John Wiley, New York, 1975. Eysenck, M. W., Arousal, Learning and Memory, Psychological Bulletin 83 (1976): 389404.
Feldt, L. S., and Mahmond, M. W., Power Function Charts for Specification of Sample Size, Psychometrika 23 (1958): 201-210. Folkard, S., Monk, T. H., Bradbury, R., and Rosenthal, J., Time of Day Effects in Schoolchildren’s Immediate and Delayed Recall of Meaningful Material, British Journal of Psychology
68 (1977): 45-50.
Folkard, S., Knauth, P, Monk, T. H. and J. Rutenfranz, The Effect of Memory Load on the Circadium Variation in Performance Efficiency Under a Rapidly Rotating Shift System, Ergonomics 19 (1976): 479-488. Gale, A., Bull, R., Penfold, V., Coles, M., and R. Barraclough, Extroversion, Time of Day, Vigilance Performance, and Physiological Arousal, Psychonomic Science, 29(1972):
l-5.
Gibson, Lawrence D., Not Recall, Journal of Advertising
Research 23(l)
(1983): 39-46.
Golden, Linda L., and Keren, A. Johnson, The Impact of Sensory References and Thinking Versus Feeling Appeals on Advetising Effectiveness, Advances in Consumer Research 10 (1983): 203-208.
Green,
Paul E., and Schaffer, Catherine
M., Ad Copy Testing, Journal of Advertising
Research 23 (1983): 73-80.
Hornik, Jacob, Diurnal Variations in Consumer Response, Journal of Consumer Research (March 1988): 588-591. Hornik, Jacob, Situational Effects on the Consumption of Time, Journal of Marketing 46 (1982): 44-55.
Jarrett, Joanna, and Furnham, Adrian, Time of the Day Effects on Immediate Memory for Television News, Human Learning 2 (1983): 261-268. Kakkar, Pradeep, and Lutz, Richard J., Situational Influences on Cosumer Behavior: A Review, in Perspectives in Consumer Behavior. Third Ed. Herald H. Kassurjian and Thomas S. Robertson, eds., Scott, Foresman and Co., Glenview, Ill., 1981, pp. 204214. Krugman, Herbert E., Low Recall and High Recognitions of Advertising, Journal of Advertising
Krugman,
Research 26 (1986): 79-88.
Herbert
E., Why Three Exposures
May be Enough,
Journal
of Advertising
Research 12 (1972): 11-14.
Leigh, James H., and Menon, Anil, A Comparison of Alternative Recognition Measures of Advertising Effectiveness, Journal of Advertising 15 (1986): 4-12. Mehrabian, A., and Russel, J. A., An Approach to Environmental Psychology, MIT Press, Cambridge, Mass., 1974. Pallak, Suzanne R., Murroni Enid, and Koch, Julian, Communicator Attractiveness and
106
J
BUSN RES 1989:18:89-106
Expertise, Systematic
J. Hornik Emotional Processing
Versus Rational Appeals, and Persuasion: A Heuristic Interpretation, Social Cognition 2 (1985): 122-141.
Petty, Richard E., Caciopo, John T., and Schumann, to Advertising Effectiveness: The Moderating Consumer Research 10 (1983): 135-146.
David, Central and Peripheral Routes Role of Involvement, Journal of
Schlinger, Mary J., Respondent Characteristics that Journal of Advertising Research 22 (1982): 29-36. Schmalensee, Diane H., Today’s Top Priority Advertising Research 23 (1983): 49-60.
Versus
Affect
Advertising
Silk, Alvin J., Test-Retest Correlations and the Reliability Marketing Research 14 (1977): 476-486.
Copy-Test
Attitude
Scales,
Questions,
Journal of
of Copy Testing,
Journal of
Research
Winch, W. H., Mental Adaption During the School Day as Measured Reasoning, Journal of Educational Psychology 4 (1913): 17-28.
by Arithmetical