Factors Influencing Consumer Responses to Product Recalls: a Regression Analysis Approach

John C. Mowen, Oklahoma State University
David Jolly, Oklahoma State University
Gary S. Nickell, Oklahoma State University
ABSTRACT - Two hundred consumers were surveyed to determine their perceptions of four companies which had previously recalled products. The results revealed that impressions of the companies were influenced by: the knowledge that a recall had been made, the perceived danger of the defective product, the perceived corporate responsibility of the company, the knowledge of recalls by other companies, and the perceived responsibility of the company for the defect.
[ to cite ]:
John C. Mowen, David Jolly, and Gary S. Nickell (1981) ,"Factors Influencing Consumer Responses to Product Recalls: a Regression Analysis Approach", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 405-407.

Advances in Consumer Research Volume 8, 1981      Pages 405-407

FACTORS INFLUENCING CONSUMER RESPONSES TO PRODUCT RECALLS: A REGRESSION ANALYSIS APPROACH

John C. Mowen, Oklahoma State University

David Jolly, Oklahoma State University

Gary S. Nickell, Oklahoma State University

[David Jolly is an MBA student, and Gary Nickell is a doctoral student in social psychology.]

ABSTRACT -

Two hundred consumers were surveyed to determine their perceptions of four companies which had previously recalled products. The results revealed that impressions of the companies were influenced by: the knowledge that a recall had been made, the perceived danger of the defective product, the perceived corporate responsibility of the company, the knowledge of recalls by other companies, and the perceived responsibility of the company for the defect.

INTRODUCTION

Negative or unfavorable public information regarding a company and/or its products is of major concern to business. Exposure to unfavorable information results from such sources as reports of defective products, news media presentations of corporate earnings (e.g., the oil industry), and publicity concerning questionable competitive practices. An excellent case in point is the Three Mile Island nuclear incident. The negative information resulting from the incident has seriously jeopardized the future of the nuclear power industry in the United States. In a recent study regarding the effects of negative product information, Weinberger and Dillon (1979) found that: (1) unfavorable product ratings tended to have a greater impact on purchase intention than did favorable ratings (2) unfavorable product information received from an independent testing agency or peer as a source had a relatively stronger effect on purchase intentions than did similar information communicated by the trade and professional association source, and (3) consumers probably place more reliance on outside sources of information when purchasing a service as opposed to purchasing a product. These findings emphasize the necessity of an organization to react effectively when consumers receive negative information concerning that organization.

Product recalls have become a common avenue by which consumers are exposed to negative corporate information. Approximately 25 percent of all consumer goods firms listed in Fortune's 500 were involved in recall campaigns in 1974 (Xerin 1974). The Conference Board estimates that 25 million product units will he recalled every year (McGuire 1975).

Because of the increasing number of product recalls, many articles have been written on the subject from a "corporate management" perspective. Researchers have written articles concerning the physical tracing and recall of the products (Flak and Chandran 1975; Warner 1975), contingency plans for recalling products (Kerin 1974), the impact of recalls on market share (Gray 1978; Wynne and Hoffer (1976), assessing the effectiveness of drug recalls in pharmacies (Gumbhir and Jamison 1975), surviving product recalls (Snyder 1974), and the high costs associated with product recalls (Tamarkin 1978).

For a number of reasons it is likely that the number of product recalls will continue to increase. One is the consumer activist movement. Since the early 1960's, consumers have become more vocal in their criticisms of business organizations in general and those that may have produced a faulty product in particular. The end result is that consumers are becoming much more inclined to report defective products. A second reason is the more activist stance taken by such government agencies as the Consumer Product Safety Commission and the Food and Drug Administration. This increase in intervention is due partly to the increase in voiced demands by the consumers. A final reason for the predicted increase in recalls is the increasing complexity of products. In order for manufacturers to comply with consumer demands and regulatory requirements for safer and better quality products, the products are becoming more complex, thereby, increasing the possibility that defects will occur.

As previously mentioned, a small literature has developed based on the managerial question of "how to" recall a product. It appears, however, that only two empirical studies have been performed to specifically investigate consumer reactions to the negative information resulting from a product recall. Mowen (1979a) manipulated the three independent variables of (1) the extent of injury resulting from the defective product, (2) the length of time taken by the company to make the retell, and (3) the number of previous product recalls made by the company. The results of this study revealed that the extent of injury resulting from a defective product, the number of previous product recalls, and the length of time to recall the defective product each influenced consumer perceptions of the corporation and of a replacement product.

In a subsequent study, Mowen (1979b) manipulated the independent variables of: (1) familiar versus unknown company making the recall, (2) action or inaction by the Consumer Product Safety Commission to force the recall, and (3) whether or not other manufacturers had had similar problems. The results revealed that consumers perceived a familiar company as significantly less responsible for a defect than an unfamiliar company. It was also found, unexpectedly, that consumers perceived the company as more responsible when it acted prior to intervention by the Consumer Product Safety Commission.

In the work by Mowen, studies were conducted in which subjects received fictitious magazine articles describing a product recall. Within the articles, information was varied in order to determine the effects of certain variables on the perceptions of companies. Such as expert-mental approach has the advantage of establishing high internal validity. However, the external validity of the studies are suspect. In the "real world," consumers simply may not attend to product recall information. Thus, while one may find large effects in an experimental setting, in a naturalistic setting the effects may be minimal.

To overcome the above problems in using an experimental methodology, the present study utilized a survey research approach. Two hundred randomly selected subjects completed surveys asking questions about the public's perceptions of Ford Motor Company, Firestone, Inc., Corning Glass Works, Inc., and Conair, Inc. Each of the companies had previously made product recalls. Responses to the questionnaires were then analyzed by regression analysis. The dependent variable was the perception of the company, and the major predictor variables were knowledge of the recall, impressions of the danger of the product, the company's social responsibility in making the recall, the company's responsibility for the defect, the length of tine to make the recall, whether or not other companies had made recalls, and whether or not the company had made previous recalls. (Several additional predictor variables were also included.) If the results of the regression analyses supported the experimental results, one would have impressive collaboration of the previous studies.

METHOD

Seven research assistants were assigned to obtain respondents from seven areas in a Midwestern city (population 42,000). The areas were selected to represent lower middle, middle, and upper middle residential housing areas, Surveyors identified respondents through the following procedure. They would enter the area from any direction they wished. From a random numbers table, they determined which side of the street to sample from and how many houses to count down the street in order to identify the selected house. After completing the survey they would is to the first cross street, randomly select whether to go to the left or the right and then follow the same procedure as before to select the proper house. If the house was vacant, the surveyor would return twice more before selecting a new house. After selecting the house, the surveyor would determine if the respondent answering was a college student or a high school student. Such individuals were excluded from the study.

Respondents cooperated excellently with the surveyors. Only two totally refused to participate, and six others participated reluctantly. Three percent of the respondents were less than 21 years old; 48 percent were 21-25 years, 24 percent were 26 through 39 years old, and 25 percent were over 40 years of age. Ninety-six subjects were male and 105 were female. Forty-one percent of the subjects owned their own homes.

The surveyors were junior and senior students majoring in marketing, who volunteered to work on the research study. Students were carefully selected for their maturity and competency.

On the questionnaire respondents were first asked about their general experience with recalls, their age, and whether or not they owned a home. They were then asked questions concerning Conair, Corning Ware, Firestone, and Ford. Because of the length of the questionnaire, half of the subjects answered questions about Firestone and the other half about Ford. Respondents were-given a copy of the questionnaire, and the interviewer would read the questions and place the answers on a code sheet. All rating scales were composed of six points with appropriate anchors designated.

RESULTS

Twenty point four percent of the subjects reported having owned a recalled product. Of the 41 respondents owning a recalled product, 58.5 percent reported taking an action to correct the problem, The majority of the products reported to have been recalled were autos (53.0 percent), tires (10.2 percent), coffee pots (18.4 percent), and hair dryers (14.3 percent). The low rate of return of known defective products matches reports found in the popular literature (Duns Review 1979).

In the first set of data analyses the independent variables of interest (the consumer's knowledge of a recall, the perceived danger of the product, the company's social responsibility in making the recall, the length of time to recall, the number of previous recalls by the company, the presence of recalls by other companies, and the company's responsibility for the defect) were entered into a stepwise regression and utilized to predict the dependent variable of favorability regarding the company. The analyses were performed separately on data for each of the four companies,

Looking first at the results for Conair, one finds that danger entered first in the stepwise regression (p <.01), and responsibility for the defect entered second (p < .05). The overall R2 was .09. Thus, the greater the perceived danger of the product and the greater the perceived responsibility of the company, the less favorable was the impression of the company.

Following the same procedure as above for Corning resulted in no findings of significant effects. This is, consumer favorability toward Corning was not significantly influenced by the independent variables selected.

The analysis for Ford revealed that one significant predictor variable entered the equation. The perceived social responsibility of Ford in making the recall was significantly related to respondent's favorability towards Ford (R2 = .19, p < .0001,).

The analysis for Firestone showed substantially divergent results from that found above. Here the company's social responsibility in making the recall (p < .0001), previous knowledge of the recall by the respondent (B = .75, p < .0001), recalls by other corporations (p < .05), and danger of the product (p < .10) entered the equation (R2 = .32). Firestone was perceived more favorably when perceived as acting socially responsibly, when respondents were aware that Firestone had recalled the tire, when the danger of the tire was perceived as low, and when other companies were known to have recalled tires. Table 1 summarizes these analyses.

TABLE 1

SIGNIFICANT EFFECTS FOR EACH COMPANY USING STEPWISE REGRESSION ANALYSIS

DISCUSSION

The results of the present research partially support the previous work by Mowen (1979a, 1979b). Mowen previously found, using an experimental methodology, that varying the danger of the defective product, the length of time to recall, and the number of previous recalls by the company influenced consumer favorability regarding the company. Table 1 reveals that for Firestone and Conair the perceived danger of the product was significantly related to the consumers' perceptions of the company. In no cases were knowledge of previous recalls and the length of time to make the recall significantly related to corporate perceptions.

While the perceived length of time to recall and the recognition of previous recalls failed to directly influence the perceptions of the company, these variables may operate through moderator variables. In particular, the length of time taken to make the recall may act by first influencing the perception of the company's social responsibility. Indeed, this was argued by Mowen (1979a). The argument was based on the hypothesis that consumers will tend to expect that companies will delay recalling a product for as long a time as possible in order to maintain corporate profits. On the other hand, if the company acts rapidly and decisively to recall a product, consumers will perceive it as acting against its own short term interest and in the interest of consumers. Consequently, the belief that the action to recall the product was consumeristically oriented would be augmented (Kelley 1971). For both Ford and Firestone social responsibility did enter the regression aqua-tide (both p's < .001). In addition, timeliness was significantly correlated with perceived corporate social responsibility (for Firestone, R - -.36, p < .0001; for Ford, R = -.46. p < .0001). Preliminary work using path analysts supports these suggested relationships. Thus, the length of time taken to make a recall may influence perceptions of the company by first influencing the impression of the company's social responsibility. Mowen (1979a) also argued that the number of previous recalls affected perceptions by first influencing the perception of the corporation's responsibility for the defect. For Conair responsibility for the defect did enter the equation p < .05. However, responsibility was not significantly correlated with the number of previous recalls by the company. Indeed, the preliminary path analytic work tends to reject responsibility for the defect as a moderating variable for the number of previous recalls made by the company.

As a summary statement, evidence from the regression works partially supports the direct impact of variations in the danger of the product and of the indirect impact of the length of time taken to mere the recall on consumer perceptions of a company. No evidence was found indicating that the presence of previous recalls influenced perceptions of the companies.

The results for the Firestone analysis revealed that the mere knowledge of a recall can influence corporate perceptions. A second interesting outcome of the analysis on Firestone was the number of significant predictor variables entering the equation. It is difficult to say just what factors contributed to these differential perceptions of Firestone. In terms of awareness sore people ware aware of the Ford recall (90.0 percent) than Firestone (71.3 percent), Conair (29.8 percent), or Corning (23.3 percent). One possibility is that direct or indirect knowledge of problems with Firestone tires may have been substantially greater than such knowledge for the other products. Of the 22 respondents who reported owning Firestone 500 tires, 13 indicated that they had had problems with the tires. Only one of the seven individuals who reported owning a Pinto indicated that they had a defective gas tank, and only two of 24 Corning percolator owners reported problems. (The problems with the hairdryers was unrelated to performance.) Based on the high percentage of individuals who had problems with the 500 tires and the large number of individuals owning the tires, . it is possible that individuals may have also heard of problems with the tires from friends. Thus, personal knowledge of either a direct or indirect form may have been greater for Firestone than the other companies.

One question arising from the results involves why no significant effects were found in the analysis of the data for Corning Ware. One explanation is that with a relatively small percentage of subjects aware of the recall, it would be difficult to statistically isolate significant effects.

A second explanation is based upon experimental work performed by Mowen (1979b). In this work it was found that the impression of a well-known, highly respected company was not influenced by knowledge of a product recall to the same extent as less favorably viewed companies. Support for the speculation comes from the mean favorability of 1.7 on the six-point scale. Thus, knowledge of the Corning recall may not have influenced the respondent's impressions.

Thus far the published series of investigations performed to determine factors influencing consumer reactions to product recalls includes two experimental studies and the survey research presented in the present article. Future work should involve three separate objectives. The next major step involves the completion of the path analysis work mentioned briefly in this article to test the models developed in the Mowen (1979a, 1979b) studies. The second step will consist of replicating the previous experimental work (Mowen 1979a, Mowen 1979b) across divergent subject populations and product classes. The final stage will involve tracking over time consumer reactions to a series of companies who have made product recalls.

REFERENCES

Dunns Review (1979), "A Record Year for Recalls," 113, 28-34.

Fisk, G. and Chandran, R. (1975), "How To Trace and Recall Products," Harvard Business Review, 53, 90-96.

Gray, R. (1978), "Recalls Haven't Hurt Ford," Advertising Age, 49, 19, 8.

Gumbhir, A. K. and Jamison, A. C. ([975), "Assessing the Effectiveness of Drug Recalls in Pharmacies," Drug and Cosmetic Industry, 117, 40-3, 113-16.

Kelley, Harold H, (1971), Attribution in Social Interaction, New York: General Learning Press.

Kerin, R. A. and Harvey, M. (1975), "Contingency Planning for Product Recall," MSV Business Topics, 23, 5-12.

McGuire, E. P. (1975), "Product Recall and the Factors of Life," The Conference Board Record, 12, 13-15.

Mowen, J. C. (1979a), "Consumer Reactions to Product Recalls: An Empirical and Theoretical Examination," unpublished working paper, Department of Marketing, Oklahoma State University.

Mowen, J. C. (1979b), "Further Information on Consumer Perceptions of Product Recalls," Proceedings of the Association for Consumer Research, Tenth Annual Conference, Advances in Consumer Research: Volume VII, Jerry C. Olson (Ed.). San Francisco, California.

Snyder, J. D. (1974), "How to Survive a Product Recall," Sales Management, 113, 23-27.

Tamarkin, B. (1978), "Recalls--The Costs Soar," Forbes, 122, 79-80.

Warner, H. W. (1977), "Guidelines for Product Recall,' Public Relations Journal, 33, 11-13.

Weinberger, N. C. cud Dillon, W. R. (1979), "Thc Effects of Unfavorable Product Rating Information," Proceedings, Association for Consumer Research, Volume 7, Jerry C. Oleos (Ed.), San Francisco.

Wynne, J., and Hoffer, G. E. (1976), "Auto Recalls: Do They Affect Market Share," Applied Economics, 8, 157-63.

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