Persuasive Communications and the Boomerang Effect: Some Limiting Conditions to the Effectiveness of Positive Influence Attempts

Millard F. Mann, University of Kansas
Thomas Hill, University of Kansas
ABSTRACT - All advertisers attempt to persuade consumers to buy their products. The theory of psychological reactance (Brehm 1966) specifies the conditions under which persuasive attempts are likely to be effective, and when such strategies may actually boomerang. The present investigation is concerned with the special case of litter control. The results demonstrate that the combination of different positive influence strategies can create the "boomerang" effect and decrease the amount of appropriate disposal of waste (a by-product of every purchase a consumer makes).
[ to cite ]:
Millard F. Mann and Thomas Hill (1984) ,"Persuasive Communications and the Boomerang Effect: Some Limiting Conditions to the Effectiveness of Positive Influence Attempts", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 66-70.

Advances in Consumer Research Volume 11, 1984      Pages 66-70

PERSUASIVE COMMUNICATIONS AND THE BOOMERANG EFFECT: SOME LIMITING CONDITIONS TO THE EFFECTIVENESS OF POSITIVE INFLUENCE ATTEMPTS

Millard F. Mann, University of Kansas

Thomas Hill, University of Kansas

ABSTRACT -

All advertisers attempt to persuade consumers to buy their products. The theory of psychological reactance (Brehm 1966) specifies the conditions under which persuasive attempts are likely to be effective, and when such strategies may actually boomerang. The present investigation is concerned with the special case of litter control. The results demonstrate that the combination of different positive influence strategies can create the "boomerang" effect and decrease the amount of appropriate disposal of waste (a by-product of every purchase a consumer makes).

INTRODUCTION

An issue of importance to marketing researchers is the determination of effective means of influencing consumers' behaviors and choices. In recent years the research on this topic has primarily been guided by attribution and self perception theory (e.g., Hansen and Robinson 1980, Mowen and Cialdini 1980, Allen, Schewe, and Wijk 1980, Reingen and Kernan 1979, Tybout 1978).

A different theoretical perspective that has generated a significant amount of research in social psychology, but has received little attention in consumer research, is suggested by Brehm's theory of psychological reactance (Brehm 1966, Brehm and Brehm 1981). In a recent review of the relevance of the theory for consumer research, Clee and Wicklund (1980) state that the "potential contribution to such disciplines as consumer behavior seems imminent (p. 403)." In short, reactance theory is concerned with individuals' responses to coercion or influence attempts. "In general, the theory holds that a threat to or a loss of a freedom motivates the individual to restore that freedom" (Brehm and Brehm 1981, p. 4; detailed discussions of the theory can also be found in Brehm 1966),

Consumers certainly feel free to choose between a large variety of products and any attempt to alter these behaviors may be perceived as a threat to these established freedoms. In general, any persuasive communication can be expected to create two opposite forces: A force leading to positive change (i.e., in the direction advocated by the persuasive communication) and a force leading to resistance (see Brehm and Brehm 1981, p. 140). If a persuasive communication is worded strongly enough (and the freedom not to comply is sufficiently important to the target of the communication), reactance motivation, the force directed at non-compliance. may be aroused. Under these conditions a persuasive communication can cause subsequent changes in behavior or attitudes away from the position advocated in the persuasive communication. This effect which has been labeled the "boomerang effect" has been demonstrated in numerous experimental studies (e.g., Brehm 1966, Brehm and Mann 1975, Snyder and Wicklund 1976, Worchel and Brehm 1970).

Freedom of product selection is not the only relevant behavioral freedom consumers have. Most consumer items are wrapped, bottled, boxed, canned or otherwise sealed for protection. These containers usually serve no further function after the purchase is made and will be thrown away. The elimination of waste is costly and unavoidable. When waste becomes litter, these costs are multiplied. Litter, however, is avoidable. Persuasive communications, used for years to influence buying habits, have been increasingly used in the last ten years in an attempt to modify the ways in which consumers dispose of waste. It has been demonstrated that written prompts urging people not to litter (e.g., Geller 1973, 1975, Geller, Witmer, and Orebaugh 1976), incentive systems (Burgess, Clark, and Bendee 1971, Clark, Burgess, and Hendee 1972, Kohlenburg and Phillips 1973, Powers, Osborne, and Anderson 1973, Chapman and Risley 1974, Hayes, Johnson, and Cone 1976), and obtrusive trash receptacle designs (e.g., Geller, Brasted, and Mann 1979) effectively decrease the amount of littering (see Geller 1980 for a review of the literature).

However, to the degree that people feel free to do what they want with waste, these persuasion attempts may arouse reactance, and boomerang. For example, in a movie theater (Geller 1973), where people may feel that the right to litter is included in the price of a ticket, simply politely specifying a trash can's location appeared to arouse reactance. However, in a grocery store (Geller, Witmer, and Orebaugh 1976) or a public swimming pool (Reich and Robertson 1979), where people may attach less importance to the freedom to litter (or may not even feel free to litter), only much stronger persuasive communications ("You must..." and "Don't you dare litter.") caused the boomerang. A reactance theory interpretation of these unexpected findings suggests that the establishment and importance of the freedom to do what one wants with waste may be affected by the environment.

An understanding of the dynamic relationship between compliance and resistance due to reactance is necessary, if one wishes to create persuasive communications of maximum effectiveness. The purpose of the present study was to explore this relationship in the context of effective litter control. Specifically, the effects of multiple positive influence attempts, their specificity, strength, and the difficulty of complying was systematically investigated in a field experiment.

Eighteen experimental conditions were created by different combinations of three independent variables: (1) Use of a regular or an obtrusive anti-litter trash can designs (2) type (specificity and coerciveness) of prompt and (3) distance required to use a trash receptacle specified by a written prompt. Conceptually, each of the first two variables was designed to create positive influence. Specifically, obtrusive anti-litter trash can designs or specific coercive prompts were expected to increase compliance. Reactance theory suggests that the reactance motive (i.e., the force to resist) is a function of the strength of the influence attempt. The third variable was intended to affect the importance of the freedom not to comply with the anti-litter message. Reactance theory predicts that threats to important freedoms will arouse more reactance motivation than threats to freedoms of less importance (e.g., Brehm and Cole 1966, Brehm and Mann 1975). The following hypotheses were tested:

H1: From previous research (Geller 1973, 1975) it was expected that a polite written prompt specifying the location of a trash receptacle will be more effective (i.e., lead- to more litter being disposed in the specified receptacle) than a general written prompt asking people not to litter.

H2: From previous research (e.g., Geller, Brasted, and Mann 1979) it was expected that an obtrusive (beautified) trash receptacle will generally lead to more appropriate litter disposal than an unobtrusive trash receptacle; however, reactance theory suggests that strongly worded (demand) anti-litter prompts distributed in the presence of an obtrusive trash receptacle will arouse reactance motivation and decrease the effectiveness of the obtrusive receptacle. Thus, a demand request in the presence of the obtrusive receptacle was expected to significantly decrease compliance as compared to a polite request in the presence of the obtrusive receptacle. This effect was not expected when the demand prompt was presented in the presence of an unobtrusive receptacle.

H3: A strongly worded (demand) prompt specifying a distant trash receptacle was expected to arouse more reactance (i.e., induce less compliance) than a polite prompt specifying a distant receptacle. Therefore, it was predicted that the demand prompt specifying a distant receptacle would lead to less litter being deposited in the specified receptacle than a polite prompt specifying the distant receptacle.

The above hypotheses were tested in a field experiment in an indoor shopping mall. Three different prompts (polite/nonspecific, polite/specific, demand/specific) were handed out in three locations (ten feet from the specified trash receptacle, two hundred feet from the specified receptacle, or three hundred feet from the specified receptacle). The specified receptacle was either unobtrusive or was shaped like a bird, attractively painted, and with an additional anti-litter prompt written on it. This procedure resulted in a 3 (distance) x 3 (type of prompt) x 2 (trash can design) factorial design.

METHOD

Setting

Subjects were visitors to University Mall, a T-shaped, indoor shopping center adjacent to the Virginia Polytechnic and State University campus in Blacksburg, Virginia, in the southwest corner of the state. The university community consisted of 10,000 local residents and 18,000 students.

There were 22 retail merchants and a bank in the mall. Figure 1 presents a diagram of the environment. Each store had an entrance to a main, 350 x 30 ft., rectangular section. The bank opened to a smaller, 100 s 30 ft., rectangular section which intersected the larger rectangle perpendicularly at its midpoint to form an interior commons area. In this commons area were fountains and plantings. The main entrance/exit was the only route by which individuals could enter the commons area from the outdoor parking lot without first passing through a store. Three other routes to the central interior area involved passing through the three largest stores: Woolco, Heironimus (department stores), and Peoples (a pharmacy). There were six waste receptacles in the mall.

FIGURE 1

UNIVERSITY MALL, TRASH CAN LOCATIONS 1-6

Procedure

The study was conducted during four consecutive Monday through Thursday periods. The data were collected from 7 PM to 10 PM while all of the stores were open. During these time periods the handouts, containing a list of stores and information about a bus system traveling to and from the mall, were distributed. Stamped across each handout, in large letters, were the behavioral prompts. The percent of handouts recovered from the specified receptacle in front of Woolco (Location 1 in Figure 1) constituted the dependent measure.

Trash can design was manipulated via the use of two types of trash cans arranged two different ways. The first receptacle was the standard, regular can used throughout the commons area. These unobtrusive receptacles were four feet tall and made out of 1 x 4 ft. pine boards arranged vertically and stained to match the wood used as a decorating motif in the commons area. Five of these trash cans were used in the present study (in five of the six locations indicated in Figure 1). In great contrast to them was the beautified anti-litter trash can, shaped like a cardinal (the state bird of Virginia) which was used in the sixth waste receptacle location. This can was approximately 4.5 ft. tall, made from steel sheets and painted in bright colors, making it quite obtrusive while remaining in good taste. "Please be a litter bit thoughtful sings the cardinal" was printed on each side. The configuration of the five regular cans and the obtrusive (beautified) trash receptacle was changed each week in an ABAB design; each arrangement was therefore used for eight days. The first arrangement used the beautified trash can at Location 1 in front of Woolco, while the other five locations were occupied by regular cans. In the second arrangement the beautified can was placed at Location 4 (see Figure 1), in the center of the mall, while a regular can stood in front of Woolco (as well as at the other locations).

Distance of handout location from the specified disposal point was manipulated over three levels. Distribution points were located at the indoor entrance/exits to the three largest stores, Woolco (ten feet from the specified disposal location), Peoples (two hundred feet away), and Heironimus (three hundred feet away).

Type of prompt was manipulated over three levels: (1) A polite/nonspecific prompt stated "Please dispose of properly." (2) A polite/specific prompt stated "Please dispose of in the trash can in front of Woolco." (3) A demand/specific prompt stated "You must dispose of in the trash can in front of Woolco."

Each night a research assistant distributed 50 handouts with each prompt printed across them (150 total) in front of Woolco, just inside the commons area. Simultaneously, just inside the commons area in front of Heironimus and Peoples, research assistants distributed 30 handouts with each prompt printed across them (90 total at each store, fewer than at Woolco due to less traffic). All handouts were inconspicuously marked to identify the location of their distribution. Overall, 110 prompts of each type were handed out each night for a total of 16 days. After the 10 PM closing, the number of handouts found in each trash can was recorded.

RESULTS

Although the point of origin and recovery location of each handout was noted, only the data collected from the trash cans at Locations 1 and 4 were of interest. [No specific predictions were made concerning the other locations. An analysis of the number of handbills recovered from these locations did not reveal any significant differences between experimental conditions.] This information was collected for 16 days and analyzed separately for disposal location. Thus, for each location, a 2 (trash can design: cardinal can or regular can) x 3 (distance: Woolco, Peoples or Heironimus) x 3 (type of prompt: polite/nonspecific, polite/specific or demand/specific) analysis of variance was performed, treating the number of handbills recovered at the respective locations each day as the unit of analysis (n = 16 days). Consequently, trash can design was treated as a between group factor, while distance and type of prompt were treated as repeated measures factors.

FIGURE 2

PERCENTAGE OF PROMPTS RECOVERED AT LOCATION 1

For Location 1 (the specified location in front of Woolco), main effects for trash can design, F (1,14) = 6.64, p < .025, distance, F (2,28) = 3.98, p < .05, and type of prompt, F (2,28) = 31.48, p < .001, were found. These main effects were specified by two reliable interactions, type of prompt x trash can design and distance x trash can design x type of prompt, F (2,28) = 5.81, p < .01 and F (4,56) = 6.36, p < .005, respectively. Figure 2 shows the percentage of each type of prompt recovered at Location 1 (the specified location) for each distribution point and each trash can design. Solid lines denote significant differences between percentages (p < .05) and broken lines indicate nonsignificant differences.

In hypothesis 1 it was predicted that a polite written prompt specifying the trash receptacle in which to deposit the handout would be more effective than a general anti-litter prompt. This hypothesis was supported. Overall, 14.1% of the handouts carrying the polite specific prompt were deposited in the receptacle at Location 1; only 7.6% of the handouts with the general ant;-litter prompt were deposited in the trash receptacle at Location 1.

In Hypothesis 2 it was predicted that an obtrusive trash can design would be more effective in inducing litter disposal than an unobtrusive design. The significant main effect for trash can design supports this hypothesis. Furthermore, it was predicted that the positive effect of an obtrusive trash can design would be canceled if a strongly worded (demand) prompt was provided in the immediate presence (short distance, ten feet) of the obtrusive receptacle. Therefore, a separate 2 (trash can design) x 3 (type of prompt) analysis of variance was performed for the handout point at Woolco, with error estimates derived from the overall analysis. This analysis yielded a significant trash can design x type of prompt interaction, F (2,28) = 3.78, p < .05. An inspection of the panels in Figure 2 shows that the demand/specific prompt in the presence of the obtrusive receptacle significantly decreased compliance. As predicted, compliance with the demand prompt was not significantly decreased when the unobtrusive receptacle was nearby (i.e., at Location 1). These results support hypothesis 2.

In hypothesis 3 it was predicted that the demand/specific prompt would be less effective when a distant receptacle is specified than when a trash can nearby is specified. Figure 2 shows that Hypothesis 3 was only supported when an unobtrusive trash receptacle was in the specified location.

It may be recalled that for the manipulation of trash can design the obtrusive can was either placed in Location 1 (the specified location) or Location 4. All subjects who entered the mall through Peoples or Heironimus had to pass the can at Location 4 before reaching the receptacle at the specified location (Location 1). The attractive cardinal can at Location 4 could have been an obtrusive alternative when the unobtrusive receptacle was at Location 1. More specifically, it could be expected from reactance theory that using a receptacle other than the one specified in the demand/specific prompt is a possible way to restore freedom ("I'll deposit the handout wherever I want"). To further explore this possibility, an analysis of variance was performed on the percentage of handouts recovered from Location 4. This analysis yielded main effects for trash can design, F (1,14) = 28.78, p < .001 and type of prompt, F (2,28) = 10.63, p < .001. A greater percentage of handbills were found in the trash can at Location 4 when the cardinal can was used (i.e., 14.9% vs. 5.2%); and the demand/specific prompt influenced the greatest percentage of handout disposals in the can at Location 4 (i.e., 11.9: for the demand/specific prompt, 9.7% for the polite/specific prompt and 8.6% for the polite/nonspecific prompt). For this overall analysis, two interactions were reliable, type of prompt x trash can design and type of prompt x distance, F (2,28) = 4.42, p < .025 and F (4,56) = 3.37, p < .025, respectively. Figure 3 summarizes the percentage of handouts recovered from the receptacle at location 4 for each experimental condition.

FIGURE 3

PERCENTAGE OF PROMPTS RECOVERED AT LOCATION 4

The overall main effect for trash can design provides further support for Hypothesis 1. The significant interactions between distance and type of prompt, and trash can design and type of prompt support the interesting interpretation that in the condition where reactance motivation was expected to be strongest (i.e., long distance from specified can and demand/specific prompt), subjects made more use of the receptacle in location 4.

DISCUSSION

Before discussing the results, it should be noted that the nature of the study (a field experiment) prohibited an assessment of the successful manipulation of the independent variables. Although the manipulations seemed to have sufficient face validity, conclusions drawn from the results are contingent on the assumption that the conceptual variables have been manipulated as intended.

Hypothesis 1 predicted that a polite anti-litter prompt specifying the disposal location would result in greater use of the specified receptacle than a general prompt. This hypothesis was supported. This finding provides further evidence for the notion that the identification of appropriate disposal locations is more effective than making a polite, though nonspecific, plea.

Hypothesis 2 first predicted another general axiom. It was found that the obtrusive receptacle was generally more effective than an unobtrusive can. From reactance theory it was furthermore predicted that in the presence of the obtrusive receptacle a demand/specific prompt would be less effective than a polite/specific prompt. Both predictions were supported.

Hypothesis 3 examined the relationship between the difficulty to comply with the prompt and the type of prompt. The distance required in order to comply with a request to use a specified disposal point was expected to increase the importance of the freedom not to comply.

Hypothesis 3 invoked reactance theory to predict that the demand/specific prompt would be more likely to decrease compliance (as compared to a polite/specific prompt) when specifying a distant receptacle. This hypothesis was partially supported. Only when the obtrusive can was placed at Location 4 (in the middle of the mall) was the demand/specific prompt less effective than the polite/specific prompt. However, this hypothesis gained further support from the results concerning the average number of prompts recovered from the receptacle at Location 4. As mentioned in the result section, subjects who entered the mall at Heironimus (i.e., furthest away from the specified Location 1) were most likely to use the obtrusive receptacle in Location 4 when handed a demand/specific prompt. This finding is consistent with reactance theory because the greatest amount of reactance motivation was expected in this condition. If subjects saw the obtrusive bird can at Location 4 as a way to restore their freedom not to comply with the demand/specific anti-litter prompt, then one would expect increased disposal of these prompts in the bird can at Location 4. When the regular can was placed at Location 4 subjects were probably less likely to notice the receptacle.

In summary, the hypotheses derived from reactance theory reliably predicted the specific conditions under which common litter control techniques were effective. However, it may often be difficult to exactly specify the conditions under which boomerang effects can be expected. This casts some doubt on the usefulness of the reactance theory perspective in predicting the effectiveness of persuasive communications. Nevertheless, it is important to realize that different methods for persuasion may not additively combine in their effectiveness. Future research should investigate the combined effects of persuasive techniques in different contexts. For example, the combination of incentives and foot-in-the-door techniques may actually reduce compliance.

Another line of research should be directed at determining which behaviors or choices constitute important behavioral freedoms. As discussed briefly in the introduction, Geller (1973) found that a polite prompt specifying a trash can location in a movie theater was less effective than a polite/nonspecific prompt. As pointed out earlier, the freedom to litter in a movie theater day be firmly established and important. Minimal persuasive attempts will be sufficient under these conditions to arouse reactance motivation and produce boomerang effects. This interactive effect of the coerciveness of persuasive attempts (threats) and the importance of the freedoms threatened could be of great relevance when devising a marketing strategy for products of differential significance to the consumer. It would be expected that identical marketing strategies (constant level of coerciveness) may yield positive effects when consumers consider the freedom to choose as unimportant, but may decrease compliance when this freedom is of high importance.

Finally, another interesting aspect of this investigation concerns the issue of how to induce consumers to appropriately dispose of packaging materials such as cans, bottles, or boxes. The present study, as well as previous research on litter control techniques, suggests that the effectiveness of common anti-litter prompts (e.g., "Please recycle," "Keep your environment clean") could be significantly improved by increasing their specificity (e.g., "Please return this can to your grocery store for recycling"). This hypothesis also warrants further investigation.

REFERENCES

Allen, Chris T., Schewe, Charles D., and Wijk, Goesta (1980), "More on Self-Perception Theory's Foot Technique in the Pre-Call/Mail Survey Setting," Journal of Marketing Research, 1 7 ( November ), 498-502.

Brehm, Jack W. (1966), A Theory of Psychological Reactance. New York: Academic Press.

Brehm, Jack W. and Mann, Millard F. (1975), "Effect of Importance of Freedom and Attraction to Group Members on Influence Produced by Group Pressure," Journal of Personality and Social Psychology, 31, 816-824.

Brehm, Jack W. and Cole, Ann H. (1966), "Effect of a Favor Which Reduces Freedom," Journal of Personality and Social Psychology, 3, 420-426.

Brehm, Sharon S. and Brehm, Jack W. (1981), Psychological Reactance: A Theory of Freedom and Control, New York: Academic: Press.

Burgess, Robert L., Clark, Roger N., and Bendee, John C. (1971), "An Experimental Analysis of Anti-Littering Procedures," Journal of Applied Behavior Analysis, 4, 71-75.

Chapman, Clyde and Risley, Todd R. (1974), "Anti-Litter Procedures in an Urban High-Density Area," Journal of APplied Behavior Analysis, 7, 377-384.

Clark, Roger N., Burgess, Robert L., and Hendee, John C. (1972), "The Development of Anti-Litter Behavior in a Forest Campground," Journal of Applied Behavior Analysis, 5, 1-5.

Clee, Mona A. and Wicklund, Robert A. (1980), "Consumer Behavior and Psychological Reactance," Journal of Consumer Research, 6(March), 389-405.

Geller, E. Scott (1973), "Prompting Anti-Littering Behaviors," Proceedings of the 81st Annual Convention of the American Psychological Association, 8, 901-902.

Geller, E. Scott (1975), "Increasing Desired Waste Disposals With Instructions," Man-Environment Systems, 5, 125-128.

Geller, E. Scott (1980), "Applications of Behavioral Analysis to Litter Control," in Behavioral Community Psychology: Progress and Prospects, eds. David Glenwick and Leonard Jason. New York: Praeger Publishers, 254-283.

Geller, E. Scott, Brasted, William S., and Mann, Millard F. (1979), "Waste Receptacle Designs as Interventions For Litter Control," Journal of Environmental Systems, 9(2), 145-160.

Geller, E. Scott, Witmer, Jill F., and Orebaugh, Andra L. (1976), "Instructions as a Determinant of Paper-Disposal Behaviors," Environment and Behavior. 8, 417-438.

Hansen, Robert A. and Robinson, Larry M. (1980), "Testing the Effectiveness of Alternative Foot-in-the-Door Manipulations " Journal of Marketing Research. 17(August), 359-364.

Hayes, Steven C., Johnson, V. Scott, and Cone, John D. (1975), "The Market Item Technique: A Practical Procedure for Litter Control," Journal of APPlied Behavior Analysis, 8, 381-386.

Kohlenburg, Robert and Phillips, Thomas (1973), 'Reinforcement and Rate of Litter Depositing," Journal of Applied Behavior Analysis, 6, 391-396.

Mowen, John C. and Cialdini, Robert B. (1980), "On Implementing the Door-in-the-Face Compliance Technique in a Business Context," Journal of Marketing Research. 17 (May ), 253-258.

Powers, Richard B., Osborne, J. Grayson, and Anderson, Emmett G. (1973), "Positive Reinforcement of Litter Removal in the Natural Environment," Journal of Applied Behavior Analysis, 6, 579-586.

Reich, John W. and Robertson, Jerie L. (1979), "Reactance and Norm Appeal in Anti-Littering Messages," Journal of Applied Social Psychology, 9, 91-101.

Reingen, Peter H. and Kernan, Jerome B. (1979), "More Evidence of Interpersonal Yielding," Journal of Marketing Research, 16(November), 588-593.

Snyder, Mark L. and Wicklund, Robert A. (1976), "Prior Exercise of Freedom and Reactance," Journal of Experimental Social Psychology, 12, 120-130.

Tybout, Alice M. (1978), "Relative Effectiveness of Free Behavioral Influence Strategies as Supplements to Persuasion in a Marketing Context," Journal of Marketing Research 15(May), 229-242.

Worchel, Steven and Brehm, Jack W. (1970), "Effects of Threats to Attitudinal Freedom as a Function of Agreement With the Communicator," Journal of Personality and Social Psychology, 14, 18-22.

----------------------------------------