The Effects of Time Compressed Advertising on Brand Attitude Judgments

Douglas R. Hausknecht, University of Akron
Danny L. Moore, Burke Marketing Services Inc.
ABSTRACT - Three models of time compression effects on brand attitudes are presented and contrasted. A previous study which competitively evaluated the three theories is replicated and potential shortcomings in the design are corrected. The results of the present study support the earlier conclusion that time compression affects advertising by interfering with message processing. However, the increased attention to non-message cues reported in the prior investigation was not replicated.
[ to cite ]:
Douglas R. Hausknecht and Danny L. Moore (1986) ,"The Effects of Time Compressed Advertising on Brand Attitude Judgments", in NA - Advances in Consumer Research Volume 13, eds. Richard J. Lutz, Provo, UT : Association for Consumer Research, Pages: 105-110.

Advances in Consumer Research Volume 13, 1986      Pages 105-110


Douglas R. Hausknecht, University of Akron

Danny L. Moore, Burke Marketing Services Inc.

[The authors wish to thank kanchana kalaiarasi for her assistance in the early phases of this research. We are also grateful for the support received from the University of Florida Center for Consumer Research.]


Three models of time compression effects on brand attitudes are presented and contrasted. A previous study which competitively evaluated the three theories is replicated and potential shortcomings in the design are corrected. The results of the present study support the earlier conclusion that time compression affects advertising by interfering with message processing. However, the increased attention to non-message cues reported in the prior investigation was not replicated.


Recently, time compression has gained increasing acceptance among advertisers as a method of reducing media costs by placing more copy into promotional spots (Advertising Age November 19, 1979; USA Today November 19, 1984). The principal difference between compressed and normal speed ads is that the speaking rate in compressed ads is faster; typically, an ad is recorded at a normal speaking rate (i.e., 140 to 160 words per minute) and electronically processed to achieve a faster presentation without distorting sound quality. While most advertisers only compress ads between 105% and 130% of normal speed to minimize the likelihood that consumers will detect the fast speaking rate, some commercials have been compressed by as much as 160% (e.g., from 48 seconds to 30 seconds in length).



The purpose of this paper is to report one experiment in a program of research designed to evaluate the effect of compressed ads on consumers' attitude toward the advertised brand. Two classes of models are tested: 1) perceptual/evaluative models and 2) a cognitive elaboration model. The mediating processes postulated by these models are summarized in Figure A to facilitate discussion.


Perceptual/Evaluative Models

Perceptual/Evaluative models assume that the effect of compressed ads on brand attitude is mediated by consumers' perceptions of the advertisement or source. Specifically, these models predict that time compression enhances the favorability of brand attitude judgments because consumers react more favorably to compressed presentations or because fast speakers are perceived to be more knowledgeable about the subject matter.

The Preference Hypothesis. MacLachlan and his associates (LaBarbera and MacLachlan 1978; MacLachlan 1982: MacLachlan and Siegel 1980) have proposed that consumers prefer ads compressed between 120% and 130% of normal-speed. They postulate an optimal information transmission rate and find evidence from paired comparison tests that consumers rate moderately compressed ads more favorably than normal speed and highly compressed ads. This suggests that compression effects on brand attitude may be mediated by affective reactions to the ad. Furthermore, brand attitude judgments should be curvilinearly related to time compression; compression rates between 120% and 130% are predicted to generate the most favorable brand attitude judgments. In Figure A, the processes hypothesized by MacLachlan's preference model are depicted by lines 1 and 5.

Source Credibility Hypothesis. A second evaluative mechanism that may enhance the effectiveness of compressed presentations has been postulated by MacLachlan (1982) and Miller, Maruyama, Beaber, and Valone (1976). MacLachlan hypothesized that consumers ray use speaking rate as an index of the source's confidence because confident speakers tend to talk faster and pause less frequently. This conjecture is supported by person perception studies that show that fast speakers are judged as more competent, truthful, fluent, and persuasive than slow speakers (Apple, Streeter, and Krauss 1979; Smith et al. 1975). In addition, Miller et al. (1976) found that fast speakers were more persuasive when delivering messages about "the dangers of caffeine" and "hydroponic gardening". A simple interpretation of these findings is that the effects of compressed ads on brand attitude judgments are mediated by evaluative reactions to the source (links 3 and 6 in Figure A).

A Cognitive Elaboration Motel of Time Compression Effects

Moore, Hausknecht, and Kalaiarasi (forthcoming) noted that time compression is likely to affect the way consumers process advertising information. According to their motel, fast presentation rates disrupt cognitive elaboration; that is, consumers generate fewer associations between advertising information and information contained in memory. If so, then time compression may either enhance or reduce persuasion by reducing the number of favorable or unfavorable associations to the brand. Moore et al.'s formulation is consistent with a variant of Petty and Cacioppo's (1981) Elaboration Likelihood Motel (ELM) and Kisielius and Sternthal's (1984) availability-valence hypothesis.



The ELS interpretation predicts that consumers' attitude judgments are determined by cognitive responses to both message-related (or "central") and nonmessage (or "peripheral") factors. In advertising contexts, message-related thoughts represent inferences about message claims or product features. Nonmessage thoughts, in contrast, include evaluations of the source and ad, and are assumed to require fewer cognitive resources. Moore et al. proposed that fast presentation rates draw attention to nonmessage elements of the ad and distract attention from the central, message cues by inhibiting subjects' ability to process the ad fully (Petty and Cacioppo 1981 - Consumer behavior researchers may be more familiar with motivational operationalizations of the ELM, such as found in Petty, Cacioppo and Schumann 1983). Thus, for compressed ads: 1) the relative impact of nonmessage cues on subsequent attitudinal judgments should increase, 2) the relative persuasive impact of messages evoking counterarguments should increase, and 3) messages that generate favorable reactions should be less persuasive.

In comparison with the perceptual/evaluative models, the ELM interpretation does not predict changes in the valence of affective reactions to the ad or source; instead, the relative weight that consumers give to the ad or source when evaluating the brand should increase for compressed ads because message and product information is processed less elaborately. In Figure A links 2, 4, and 7 depict the processes implied by the cognitive elaboration model.


To conserve space, an overview of time compression research is given in Table I. The discussion in this section is limited to a brief description of the major findings related to the processes depicted in Figure A. The link between time compression and affective reactions (#1 in Figure A) to the ad has received the most attention in previous research. MacLachlan and LaBarbera (1978) and LaBarbera and MacLachlan (1979) found that ads compressed to 120%-130% of normal speed are rated as more interesting than normal-speed presentations. However, Schlinger et al. (1983) reported that compressed ads are rated as more patronizing and less personal. Finally, Moore et al. (forthcoming) found no difference between compressed and normal speed ads on a set of semantic differential scales designed to measure attitude toward the ad. Thus, the prediction that time compression affects consumers' attitude toward the ad has received only marginal support.

Only one study has examined the relationship between time compression, attention processes and cognitive response (82 and #4 in Figure A). Moore et al. (forthcoming) found that compressed ads capture less attention. They also reported that consumers generate fewer cognitive responses to advertising claims and the product and more cognitive responses to ad execution factors when ads are compressed beyond 130% of normal-speed.

The evidence that time compression influences consumers' judgments of source credibility in advertising settings (23 in Figure A) is mixed. MacLachlan (1982) found that spokespersons are rated as more enthusiastic and energetic when ads are compressed to 125% of normal. However, no significant compression effects on source attractiveness and expertise scales were reported in the Moore et al. study. Thus, even though person perception research suggests that fast speakers are rated as more confident, fluent, and persuasive, this does not appear to occur in advertising contexts.

With respect to the prediction that time compression influences brand attitude judgments, Wheeless (1971) and Lautman and Dean (1983) found no significant compression effects. However, other studies have shown that product ratings are more favorable for compressed versions of some ads, but normal-speed versions of other ads evoke more favorable reactions to the product (Riter et al. 1982; Schlinger et al. 1983). Moore et al. (forthcoming) demonstrated that time compression generates more favorable brand reactions for ads containing weak, unpersuasive arguments; they also found a small decline in the favorability of brand attitude when ads containing strong, persuasive arguments were compressed.

In short, the pattern of evidence from previous research provides little support for the predictions of the preference and source credibility hypotheses. Time compression does not appear to exert strong effects on affective reactions to the ad nor does it appear to enhance source credibility. Furthermore, time compression does not simply enhance brand attitude Judgments. Instead, the evidence suggests that compression interacts with advertising characteristics to produce an effect on brand attitude, favoring the cognitive elaboration model.


The experiment reported in this paper was designed to test further the predictions of the three models. At present, the only explicit competitive test of these motels was conducted by Moore et al. (forthcoming). In their research, consumers listened to a series of ads with a critical test ad occupying the middle position. Following each at cognitive responses were collected. This procedure has two potential shortcomings. First, the experimental setting bears little resemblance to natural communication settings. When ads are embedded in a program and when consumers are engaged in distracting activities, compressed ads may be preferred because consumers view advertising as an unpleasant interruption in the program. Second, collecting cognitive responses to the at forces consumers to focus their attention on message arguments, spokespersons, and ad execution. Thus, the observed effects in Moore et al.'s study may not replicate when cognitive responses are not collected.

To overcome these problems, we attempted to replicate Moore et al.'s findings in a more natural listening setting. Several versions of a radio ad were created by manipulating source credibility, argument strength, and time compression - factors identical to those used in the Moore et al. study. The message reception environment was modified by: 1) embedding commercials within a radio program, 2) instructing consumers that their task was to evaluate the radio program rather than the ads, 3) providing a concurrent distractor task, and 4) eliminating the cognitive response measures. Thus, the critical problem addressed in our study is whether the cognitive elaboration model provides an adequate account of time compression effects when attention is not focused on advertising material. More important, the experimental design permitted a further competitive test of hypotheses examined in the Moore et al. study.

H1: The Preference Hypothesis: Moderate increases in presentation rate (i.e., 130%) produce more favorable reactions to the at and brand than slow or fast presentation rates.

H2: The Source Credibility Hypothesis: Compressed ads enhance the perceived credibility of the spokesperson an produce more favorable reactions to the brand.

H3. The Cognitive Elaboration Hypothesis: Time compression reduces consumers' ability to process message assertions. This reduces the impact of message arguments and enhances the impact of source credibility on brand attitude judgments.



Students in an introductory Marketing class participated in the experiment for course credit. Of the 254 participants, data from 11 were discarded due to item nonresponse leaving results from 101 women and 142 men for subsequent analysis.


Subjects listened to one of 24 radio commercials generated from a 2 x 2 x 2 x 3 between-subJects factorial design. The factors in the experiment were Product Class (Calculators or Razors), Argument Strength (Strong versus Weak), Source Credibility (High versus Low), and Message Exposure Rate (100%, 130%, and 160%). The between-subjects manipulation was chosen to control for demand effects which may arise from within-subjects manipulations of presentation speed.

Materials and Stimuli

The stimulus commercials used were identical to those used by Moore et al. (forthcoming), as were the noncritical commercials used to simulate clutter. A thirty minute segment of a syndicated rock and roll program provided by a local ratio station constituted the wain tape in which commercials were embedded at logical breaks. Five "puzzles" (maze, wordsearch, etc.) were photocopied and provided the distractor task.


The experiment was conducted in a language laboratory that seated 24 subjects at individual booths, each containing a tape recorder and a set of headphones. Subjects participated in groups ranging in size from 11 to 22. At the beginning of the experiment, instructions were read that explained the use of the tape recorder and indicated that the purpose of the experiment was to evaluate the appropriateness of a radio program for the local, predominantly student market. Subjects were also instructed to work on distractor tasks while listening to the radio program. The use of the distractor tasks was explained as necessary since, "[people] usually read or go about their normal routine when listening to the radio..we'd like everyone to work on these puzzles so that all of you are doing similar activities during the program".

The critical ad appeared midway through the program (third in a total of five ads presented during the thirty minutes). The calculator commercials contained a testimonial communicated by either a "Princeton Professor" (High Credibility) or a "high school student" (Low Credibility). The "strong arguments" version listed such reasons to buy as, "provides an ideal combination of financial and statistical capabilities needed in college courses" and "performed computations with the speed and accuracy of a computer". The "weak arguments" version relied on reasons like, "comes with a case that can be attached to your belt" or "does not provide a warranty, but I don't believe that you'll need one".

The razor commercials provided endorsements from either professional athletes (High Credibility) or citizens of a southern city (Low Credibility). Sample strong arguments are, "designed with a new advanced honing method that creates unsurpassed sharpness" and "special chemical coating eliminates nicks and cuts". Weak arguments included, "designed for beauty" and "can only be used once". The razor commercials were based on ads used in previous investigations of the elaboration likelihood motel (Petty et al. 1983).

At the end of the program segment, subjects were instructed to complete a questionnaire that could be found in an envelope in the booth. The questionnaire first asked for ratings of the program (both to provide an intervening task and to justify the stated purpose of the study) and then probed for recall of products and brands advertised. Finally, copy point recall, brand attitude, attitude toward the at and various manipulation checks relevant to the critical ad were collected.

Dependent Measures

The primary dependent measures were: 1) unaided recall of the brand name; 2) copy point recall; 3) brand attitude ratings; 4) ratings of the critical ad; 5) ratings of the product endorsers; and 6) ratings of argument strength. Recall measures were dummy-coded mentioned/not mentioned and ratings were collected on seven point semantic differential scales.


Ratings of Argument Strength

Three semantic differential scales were used to assess perceptions of the message arguments: 1) strong-weak, 2) persuasive-unpersuasive, and 3) important-unimportant. These items loaded on the same factor and were summed for subsequent analyses (Cronbach's 1 = .89).

Three significant effects were revealed in the analysis: 1) a main effect for Argument Strength, F(1,217) = 6.58, p<.02, 2) an Argument Strength x Exposure Rate Interaction, F(2, 217) = 3.03, p<.03, and 3) a Source Credibility x Argument Strength interaction, F(1, 217) = 4.77, p<.03.

Although the Argument Strength effect indicated that subjects rated strong arguments as more persuasive (M=11.67) than weak arguments (H=10.33), this difference was not constant across Exposure Rate and Source Credibility conditions. Figure 3 shows the pattern of means for the strong and weak argument conditions within each exposure rate condition. As this Figure shows, subjects did not discriminate between the strong and weak arguments in the 160% condition, but argument strength significantly affected ratings in the 100% and 130% conditions. This result is consistent with the Cognitive Elaboration Model (H3) and suggests that message cues were not considered carefully in the fast exposure rate condition.

Simple main effect tests on the Source Credibility x Argument Strength interaction revealed that strong arguments were rated as more persuasive than weak arguments in the high credibility condition (p<.05) but not in the low credibility condition (F< 1). This pattern is not unusual in persuasion research and suggests that subjects' interpretation of argument strength was influenced by source cues (cf. Norman 1976).

Attitude Toward the Ad

The Preference Hypothesis (H1) predicts that consumers react more favorably to ads played at 130% of normal speed. Responses to a set of eight semantic differential scales about the ad were used to test this hypothesis. Preliminary factor analyses of these scales revealed three underlying dimensions. To conserve space, these factors will be referred to as "at affect", "at credibility", and "speed of speech".

Ad Affect. Analyses of five summed ad affect scales (Cronbach's a = .87) revealed a significant effect for Product Class, F(1, 218) = 5.26, p<.03. Subjects reacted more favorably to the razor ads (M-18.54) than the calculator ads (M=16.69). No additional effects were significant.

The failure to find a significant effect of Exposure Rate on ad affect is damaging to the Preference Hypothesis. Inspection of the means for the Exposure Rate conditions indicated a pattern consistent with the Preference Hypothesis. The 130: condition evoked the most favorable reactions (M-18.46) followed by the 100% (M-17.95) and the 160% (M-16.47) conditions. However, the difference among these means did not reach a conventional level of statistical significance, F(2, 218) = 2.22, p<.12.

Ad Credibility. The analysis of summed ratings for the two ad credibility scales (Cronbach's w - .74) indicated a main effect for Product Class, F(1, 216) w 5.60, p<.02. The razor ads were rated as more believable and truthful (M=8.93) than the calculator ads (M-8.16). All remaining main effects and interactions were not statistically significant.

Speed of Speech. An interesting issue for practitioners concerns whether or not consumers notice that ads are compressed. In previous research this issue has been examined with open-ended questions that probe for awareness of time compression (e.g., MacLachlan x Siegel 1979). In the present study awareness of the time compression manipulation was examined by comparing ratings of the commercials on a semantic differential scale ranging from slow to fast. A significant Exposure Rate effect, F(2, 217) = 28.32, p<.0001 and interactions between Product Class and Argument Strength, F(1, 217) 5 6.54, p<.02, and Product Class and Exposure Rate, F(2, 217) = 4.35, p<.02, were uncovered in the analysis. The most important of these effects was the Product Class x Exposure Rate effect. Table 2 shows the mean ratings for each Exposure Rate condition by Product Class. Comparisons of the speed ratings within each Product Class indicated similar trends. No significant differences were detected for comparisons between the 100% and 130% condition; however, the 160% condition was rated as significantly faster for both the razor and calculator ads. Thus, subjects did not report significant increments in speed until the compression rate surpassed 130%.





Perceptions of the Product Endorsers

According to the Source Credibility Hypothesis (H2), time compression should enhance the perceived credibility of the product endorsers. Nine semantic differential scales were included to test this hypothesis. We expected these scales to reflect an attractiveness dimension and expertise dimension (cf. Moore et al. forthcoming). However, a factor analysis yielded only one factor. Therefore, in our analyses all nine scales were summed (Cronbach's X .95).

Three significant effects emerged from the analysis. First, high credibility endorsers were perceived to be more credible (M"38.38) than low credibility endorsers (M-29.03), F(1, 217) s 49.27, p<.0001. This effect was qualified by a Product Class x Source Credibility interaction, F(1, 217) - 5.73, p<.02. Simple main effect tests of this interaction indicated a significant product effect for the high credibility conditions (M's= 41.13 and 35.74 for the calculator and razor ads respectively), p<.05; the product effect was not significant in the low credibility conditions (M's= 28.52 and 29.52 for the calculator and razor ads respectively). Thus, subjects generally rated the "Princeton Professor" in the calculator ads as a more credible source than the "professional athletes" used in the razor ads, but the other sources were perceived as equally credible.

The third significant effect in the analysis of source ratings was an Argument Strength x Exposure Rate interaction, F(2, 217) = 4.32, p<.02. Table 3 shows the relevant mean ratings for this interaction. This Table indicates that product endorsers were not simply given higher ratings as exposure rate increased (Hypothesis H2). Instead, when strong persuasive arguments were given, the endorsers were rated lower in the 160% condition than in the 100% and 130% conditions. However, consistent with Hypothesis H2, fast speakers were perceived as more credible when weak message arguments were given. One plausible interpretation of this finding is that subjects used argument strength as a cue for their ratings of source credibility. When the exposure rate reached a point where subjects did not respond to message quality, the suppressing effect of weak arguments was reduced. This interpretation is clearly at odds with the simple process postulated by the Source Credibility Hypothesis (H2).



Brand Attitude Judgments

Analysis of three summed brand attitude scales (Cronbach's a - .90) provided the critical test of the hypotheses. Both the Preference Hypothesis (H1) and the Source Credibility Hypothesis (H2) predict a main effect for Exposure Rate. The Cognitive Elaboration Hypothesis (H3), in contrast, predicts that message arguments should exert less impact and source credibility should exert more impact on brand attitude judgments as exposure rate increases; if so, then an interaction among Argument Strength, Exposure Rate, and Source Credibility should occur (cf. Moore et al. forthcoming).

The analysis provided little support for H1 and H2; Exposure Rate did not exert a significant main effect on brand attitude judgments, F<1. However, a significant Argument Strength x Exposure Rate interaction emerged, F(2, 218) - 3.20, p<.05. Figure C shows the pattern of brand attitude judgments for this interaction. Follow-up tests revealed that strong arguments evoked more favorable judgments than weak arguments in both the 100% and 130% conditions (p<.05). For the 160% condition, the Argument Strength effect tit not approach significance. This pattern is consistent with the Cognitive Elaboration Motel and suggests that fast exposure rates reduce the impact of message quality on attitudinal judgments. However, the expected Argument Strength x Source Credibility x Argument Strength interaction was not significant; this suggests that a decrease in the utilization of message cues is not necessarily accompanied by an increase in the use of source clues.



The Argument Strength main effect and the Source Credibility x Argument Strength interaction were also significant. These effects were in the expected direction (e.g., strong arguments delivered by a high credibility source elicited the highest brand evaluations), and since they are not critical to the hypotheses, they will not be discussed further.

Brand Name and Copy Point Recall

The attitude results suggest that subjects may have ignored test ads when the compression rate surpassed 160%. The recall data support this interpretation. As Table 4 indicates, recall of the brand name and copy points was significantly lower in the 160% condition.




The data collected in the present study consistently indicated that message quality effects diminish when compression rates surpass 130%. This finding appeared in the analyses of argument strength ratings, source credibility ratings. and brand attitude judgments.

There are two plausible accounts of the observed data pattern. First, high levels of time compression may cause consumers to ignore advertising information. If so, then one would expect lowered recall and reduced message quality effects. A second process that could account for our data is that the fast speaking rate distracted attention from processing message claims and product information and focused attention on executional components of the at. Although this latter interpretation is consistent with the conclusions of Moore et al. (forthcoming), there are some discrepancies between our findings and the results reported in their research. Moore et al. found that high levels of compression are associated with reductions in the impact of argument strength cues and increased effects of source credibility cues on brand attitude judgments. In the present study we found the reduction in argument strength effects on brand attitude judgments, but source credibility effects did not increase in the high compression rate conditions. This difference could be attributed to procedural differences between our study and the Moore et al. study. That is, when consumers were not specifically instructed to generate cognitive responses to the ads, source credibility did not play a major role in brand attitude judgment. Since this could be due to a number of factors we feel that it will be more constructive to focus the discussion on consistencies; that is, why time compression reduced the effects of message quality cues in both the present study and Moore et al.'s study.

Neither the Preference Hypothesis nor the Source Credibility Hypothesis predicts an interaction between time compression and advertising characteristics. Thus, it is difficult to account for the observed data with perceptual/evaluative mechanisms. This points to cognitive elaboration motels as the most promising account of time compression effects. The diminished effect of message quality is predicted by both the Elaboration Likelihood Motel and Kisielius and Sternthal's (1984) availability-valence hypothesis. The key distinguishing characteristic between these models is that the ELM predicts that peripheral cues exert more impact on attitude change when cognitive elaboration is limited whereas the availability valence model is mute on this point. Both models can account for our findings.

Perhaps the major conclusion that can be drawn from both our findings and Moore et al.'s data is that there are clear limits to the use of time compression in advertising settings; rates exceeding 130% seem to interfere with message reception. Note that the 130% rate is not offered as a maximum level for advertisers, or even a theoretically critical level. The present data reveal that time compression affects learning and attitude change processes at relatively low compression levels even though comprehension may be unaffected (up to rates of 200% see, for example, Fairbanks, Guttman and Miron 1957). Thus optimal speeds are dependent on message and nonmessage aspects of each specific ad. More important, since we have ignored more subtle effects such as alterations in the consumers' interpretations of advertising claims, future applied research is likely to provide additional insights by incorporating measures of comprehension and acceptance as well as brand attitude.

Finally, the cognitive elaboration investigations of time compression have been limited to radio advertising, whereas much of the previous work dealt with television (Table 1). It is conceivable that compression may affect visual information processing differently than verbal. Relatively modest compressions can result in unintentionally humorous "Keystone Cop" movements that can detract from the impact of a commercial by creating new execution related cues.


"Do Commercials Run Faster?" Advertising Age (February 18. 1980).

"New Compression System Trims Seconds Off Ads," Advertising Age (November 19, 1979), 58-59.

"Is it Xmas if There's no Ronco?" USA TODAY (November 19, 1984), 2B.

Apple, William, Lynn A. Streeter and Robert M. Krauss (1979), "Effects of Pitch and Speech Rate on Personal Attributions," Journal of Personality and Social Psychology, 37(May), 715-727.

Fairbanks, Grant, Newman Guttman and Murray S. Miron (1957), "The Effects of Time Compression upon the Comprehension of Connected Speech," Journal of Speech and Hearing Disorders, 22(March), 10-19.

Kisielius, Jolita and Brian Sternthal (1984), "Detecting and Explaining Vividness Effects in Attitudinal Judgments," Journal of Marketing Research, 21(February), 54-64.

LaBarbera, Priscilla and James MacLachlan (1979), "Time Compressed Speech in Ratio Advertising," Journal of Marketing, 43(January), 30-36.

Lautman, Martin R. and K. Jeffrey Dean (1983), "Time Compression of Television Advertising," in Larry Percy and Arch G. Woodside (eds.), Advertising and Consumer Psychology, Lexington, MA: Lexington Books, 219-375.

MacLachlan, James (1982), "Listener Perception of Time Compressed Spokespersons," Journal of Advertising Research, 2(April/May), 47-51.

MacLachlan, James and Priscilla LaBarbera (1978), "Time Compressed Television Commercials," Journal of Advertising Research, 18(August), 11-15.

MacLachlan, James and Michael H. Siegel (1980), "Reducing the Costs of Television Commercials by Use of Time Compressions," Journal of Marketing Research, 17(February), 52-57.

Miller, Norman, Geoffrey Maruyama, Rex Julian Beaber and Keith Valone (1976), "Speed of Speech and Persuasion," Journal of Personality and Social Psychology, 34 (October). 615-624.

Moore, Danny L., Douglas R. Hausknecht and Kanchana Kalaiarasi (forthcoming), "Time Compression, Response Opportunity, and Persuasion," Journal of Consumer Research.

Norman, Ross (1976), "When What is Said is Important: A Comparison of Expert and Attractive Sources," Journal of Experimental Social Psychology, 12(May), 294-300.

Petty, Richard E. and John T. Cacioppo (1981), Attitudes and Persuasion: Classic and Contemporary Approaches, Dubuque, IA: William C. Brown Co.

Petty, Richard E., John T. Cacioppo and Davit Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10(September), 135-146.

Riter, Charles B., Phillip J. Balducci and Donald McCollum (1982), "Time Compression: New Evidence," Journal of Advertising Research, 22(December), 39-43

Schlinger, Mary Jane Rawlins, Linda F. Alwitt, Kathleen E. McCarthy and Leila Green (1983), "Effects of Time Compression on Attitudes and Information Processing," Journal of Marketing, 47(Winter), 79-85.

Smith, Bruce L., Bruce L. Brown, William J. Strong and Alvin C. Rencher (1975), "Effects of Speech Rate on Personality Perception," Language and Speech, 18(April-June), 145-152.

Stephens, Nancy (1982), "The Effectiveness of Time Compressed Television Advertisements with Older Adults," Journal of Advertising, 11(Number 4), 48-55.

Wheeless, Lawrence R. (1971), "Some Effects of Time Compressed Speech on Persuasion," Journal of Broad- casting, 15(Fall), 415-420.