Lisa Lyon and Glen T. Cameron, Ph.D.
[WJMCR 1:4 September 1998]
A fully counterbalanced, within-subjects experiment addressed fundamental questions about the value of corporate reputation. The 2 (good vs. bad reputation) x 2 (apologetic vs. defensive) design also compared apologetic and defensive responses to negative news about a company. Reputation profoundly affected memory attitude and behavioral intentions, bearing out platitudes about bottom-line importance of reputation management. Conversely, response style was not particularly robust as a factor affecting cognitive, affective and behavioral measures. Interaction effects of the two factors ran counter to common wisdom abjuring the stonewall response. Corporations with a bad reputation prior to the negative news story in the experiment were further damaged by the apologetic response style. Measures taken after ten days delay supported the differential decay school of thinking about sleeper effects and confirmed the supposition that reputation can be thought of as a prior source cue about a company before subjects process the company’s response to negative news.
Trade Wisdom Prompts Questions
There seems to be strong conventional wisdom in public relations practice that reputation matters. This is evidenced by the frequent trade articles offering advice and input on reputation management.1 It stands to reason that reputation is important, but in what ways and to what extent has less empirical support.
Some suggest that the best time to assess a company’s reputation is during times of difficulty, nothing brings out the true corporate persona like the response to a crisis.2 Conventional wisdom dictates avoiding “no comment,” and discouraging clients from taking a defensive position at all costs.3 In fact, public apology in response to the accusation of corporate misconduct is often touted as one of the most important ways to protect a company’s reputation.4
Given the platitudes about leveling with the media, why does the phenomenon of lying, stonewalling and obfuscating in response to negative information persist? Why might organizations continue to use this approach, when public relations “gurus” and academics discourage this response type? Is response type unimportant, or perhaps simply overshadowed if one’s preexisting reputation is good? And what happens to these attitudes in the long run?
These sorts of questions about the profound impact of reputation as well as the prescriptions for appropriate response to negative news prompted us to wonder just how significant reputation is as a factor and how it interacts with response type to negative news. To explore these issues, we formalized the following research questions.
RQ1: Does a negative news story have a stronger impact on subject evaluations of corporations with an already damaged reputation? Does a negative story about a company with a good reputation cause lower credibility of the negative information?
RQ3: Which is more effective in combating negative information in a news story, an apologetic or a defensive approach?
RQ3: How durable are the effects of response style (apologetic or defensive) in relation to reputation (good or bad) over an extended period of time? What are the long-term attitudes regarding the different variables tested? Are they consistent with the initial attitudes formed after the message response?
Reputation Management Literature
It has been said that by the year 2000, reputation management may be the “dominant force” in public relations.5 In today’s consumer oriented world where everything is seen as a commodity, a company’s reputation can be the key distinction between prices, technologies and product capabilities.6 “Reputation can be a company’s most powerful asset�or its costliest liability,” said Alan Towers, president of the public relations firm Alan Towers Associates. Experts agree, however, that reputations are abstract, subjective personal judgments, the origin of which is hard to pin down.7 On the one hand, the bottom line business implications of a strong reputation have prompted some scholarly work on reputation management. On the other hand, the nebulous nature of “reputation” has caused uncertainty and some disagreement about the effects of reputation.
Within the past decade, business and management scholars have developed a formalizing idea that a firm’s reputation is an asset vital to the future economic success of an organization.8 The notion that bottom line financial success is linked with corporate reputation has been studied, often using the Fortune Most Admired list as the sample for the financial data. These studies indicate variables such as above average financial returns for a firm, quality of job candidates, and initial decisions about pursuing contact with a company all hinge at least partly on corporate reputation.9
However, lack of a widely accepted measure for reputation has created difficulty in replicating the findings of others.10 Additionally, research suggests that simply using dimensions and scales that describe individuals cannot accurately capture our perceptions about organizational reputation. Attributes describing an individual such as expertise, attractiveness and trustworthiness translate to “smart organization,” “role model” and “good organization” when measuring corporate image and prior performance.11
Mass communication experts concur that the elusive concept of corporate image or reputation has caused confusion about how to define and measure it, and that devices traditionally used to measure brand image do not successfully capture corporate image or reputation.12
From a public relations standpoint, a positive reputation allows a company not only to implement its present plans, but also to pursue its goals in the future. This has been described as the “halo effect”�where “a generally positive attitude toward the company lends the company immunity to a certain extent.”13 Goldberg and Hartwick14 found evidence of this in an experiment investigating the combined effects of a company’s reputation and advertisements on product evaluations. Subjects who formed a negative evaluation of the company based on a bad reputation found the advertisements less credible and rated the products less favorably than those who received a positive reputation company.
This would indicate that by defining a method of evaluating reputation, companies could not only help protect the “bottom line,” they could also help ensure a more mutually supportive relationship between an organization and it’s publics.15
Because each day, management teams across America gather to implement reputation damage control due to negative publicity, the effect of unfavorable media coverage plays an important role in establishing corporate strategy to maintain positive reputations.16 In fact, research shows that negative information can have a powerful impact on attitudes and behavioral intentions. It has been shown that a single item of negative information can neutralize five pieces of positive information, and that negative information is more enduring than positive or neutral information.17 In the long term, negative publicity may more strongly influence attitudes and behavioral intentions than positive information.18 Some suggest the reason for this influence is the surprise factor, while others claim it may be that negative information is seen as less ambiguous.19
Most every organization will receive unfavorable coverage at some point. So, it is important that the organization receive knowledgeable counsel on how to address the problems and formulate the most effective response. While the primary focus should be on affecting positive organizational behavior, some type of response will still be required and it is essential for public relations scholars to develop the best response strategy to combat negative publicity and still maintain a positive reputation.
If reputation is well established before negative publicity strikes, the effects of corporate response strategy may decay in the long run. Miller and Campbell20found that when two sides of an argument were presented, the side learned last would be better recalled immediately, but this superiority effect would gradually decrease. This effect was also predicted with attitude change because of the assumed relation between message retention and attitude. This suggests that the power of preexisting reputation may ultimately override the carefully planned public relations response strategies. Individuals may respond more favorably toward a corporation that is willing to “own up” to its mistakes. But, if this favorable reaction is not durable over time, perhaps the apologetic response will only be washed out by corporate reputation. This can be tested by recall of the negative information, attitudinal assessment of the organization and behavior intention in the short and long term.
In rhetoric and speech communication literature, apologia is considered an effective self defense strategy, appropriate when individuals must defend their character.21 Empirical research suggests, however, that an apologetic response strategy can be effective in some cases but not in others.22 Nevertheless, public relations practitioners have since drawn upon this rhetorical strategy as the best way to respond to negative publicity with an internal locus of responsibility.23
Organizational image may be negatively related to perceived crisis responsibility.24 Simply put, past crises color how people may interpret the current crisis. Coombs25 suggests that a “defensive strategy” is most effective when public perception of crisis responsibility is strong. However, repeated past crises will likely intensify perception of crisis responsibility. This reinforces the importance of an accurate understanding of public perception regarding organizational reputation.
Specifically, crisis management specialists have suggested there are five possible response types to negative or crisis situations: denial, distance, ingratiation, mortification, and suffering. Of these, experts suggest the mortification strategy when the negative information results from an internal, intentional action taken by an organization that places publics at risk or harm.26 Mortification strategies must first acknowledge responsibility for the negative information, then the organization seeks to redress the problem in some manner (apology, compensation, etc.).27
Like apologia, the stonewall concept also has roots in rhetoric/speech communication. While silence can be considered a rhetorical strategy, experts claim it likely implies passivity and a relinquished control over the situation. This response type can create or add to existing doubts and uncertainty and has been deemed a “public relations” issue that simply reinforces the public’s negative impressions.28 While some public relations scholars suggest a “transcendence” response, where the company tries to place the negative information in a different, more desirable context, stonewalling has not been empirically tested as a possible response type to negative publicity.29 Instead, anecdotal examples such as Ford Motor Company’s use of delaying and stonewalling facts with its Pinto are cited as examples of the futility of this approach.30
Tyler31 suggests that it is naive to believe that a corporate apology in response to negative publicity would work well in all situations. In fact, the solution could be a combination of rhetorical strategies. Strategic ambiguity, or equivocal communication may be necessary in some public relations scenarios today where the companies need to protect themselves from legal liability, or wish to alleviate doubts about the firms ability to generate future revenue, but also feel the need to apologize. Commonly known as “doubletalk,” this strategy has been recast as “communicative avoidance conflict.” This ambiguous, contradictory or evasive response style is being recommended for situations when corporate executives are trapped because issuing an apology could raise potential liability problems.
The structural guidelines on how to respond to such situations often assume, however, that the media coverage is disseminated upon a tabula rasa public and does not take into account the context or reputation of the organization. In order to fully comprehend the impact negative information has upon an organization, we must not view response as a snapshot or cross section. Taking a reflexive orientation, reputation is an ongoing index of previous responses to situations, making the most immediate response strategy a key element of that index, but also a response that should be made in light of current reputation. The interaction of reputation and response may be such that traditional strategies do not apply in all cases and certainly calls into question pat answers or formulaic principles such as: “Always Apologize” or “Never Say No Comment.”
Durability of Effects of Reputation and Response
Source credibility and memory.
Cameron and colleagues have reviewed and adapted nearly 50 years of sleeper effects literature to current public relations theory and issues in the profession.32Source credibility studies starting with Hovland have included memory for message as a dependent variable, with higher source credibility contributing to better memory which in turn can contribute to attitude change.33
Cook and Flay concluded from a review of over 25 studies of the role of memory in attitude change that it “seems unlikely that retention of message details and delayed attitude are related in any simple way, though they may be related in complex ways.”34 They stated that although “the evidence is equivocal, there are indications of a causal relationship, and it would be premature to suggest that persistence is not at all related to the retention of broader details of a persuasive communication”.35 Other researcher have found just such a relationship between retention of messages and attitude formation and persistence.36
Timing of source notification (before or after message) particularly influences the effect of source on audiences.37 Research has suggested that more than a simple association of a credible source with a persuasive position must occur, otherwise timing of source notification would not be a significant factor. This view gained further support from Pratkanis, Greenwald, Leippe, and Baumgardner in 1988 as well as Homer and Kahle in 1990. Timing of source notification before or after a message affected memory for the message and for the source, as well as affecting persuasive impacts. For the current study, corporation reputation reifies credibility of the source responding to negative news. In sum, the prior source cue here is the corporation’s reputation. The implication of prior notification was summarized by Cameron in 1994:
When a source cue is given after a message, encoding processes are already done, with additional traces laid down in memory for source information found at the end of the message. With prior source attribution, encoding is influenced by the credibility of the source. Prior source attribution leads to integration of source with message. When a source has lower credibility, the message is accorded less credence and less effortful elaboration in memory occurs (i.e. poorer memory for the message).38
Durability of source effects. The persuasion literature includes considerable attention to the durability of source credibility effects. In the current study, durability of source effects are integral to the purported effects of reputation on subsequent processing of information about the source. If reputation is not durable, then it does not affect later processing of messages about the organization.
Most of the work on the interaction of time and source conditions has involved the search for and explication of some variant of the sleeper effect, beginning with the idea that forgetting the source releases deleterious effects of low source credibility, resulting in an absolute increase in persuasive impact for the low credibility message.39
When it was found that source was remembered, the dissociation hypothesis was developed.40 Failure to consistently find an absolute sleeper effect represented by a rise in persuasive impact over time called into question the robustness of the sleeper effect.41 in 1978, Gruder found absolute sleeper effects, but specified certain conditions necessary for obtaining the effect.42
Gillig & Greenwald found more evidence for a relative sleeper effect.43 The relative sleeper effect occurs when a low credibility source has less initial impact than a high credibility source, but experiences less decay of impact over time than does the high credibility source.44 Pratkanis et al45 repeatedly found a relative sleeper effect and offered a replacement for the dissociation hypothesis�the differential decay hypothesis that:
subjects remember the communication events (episodes), but the impact of the communication (on evaluation) decays over time. This assumption is made in the differential decay interpretation. Thus, it is not the dissociation of events in memory but a differential dissipation of the impact of two different persuasive communications. This is consistent with the notion of two separate memory systems46: one for episodes or events (i.e., I heard a message and cue) and one for meaning (i.e., I favor X).47
With the present study, reputation could be considered a prior source cue, therefore a predicted sleeper effect for persuasive impact would probably not be found.
According to the differential decay interpretation, a sleeper effect is not likely to occur when discounting information precedes the message. In such cases, subjects may be more disposed to counterargue with the message as they read. Thus, the persuasive impact of the message is attenuated, and the message and source are more likely to form a unit in memory.48
With prior cuing for source, the message and source cue “can be interpreted only in light of the other.”49 In considering reputation (i.e. source cue) and response style in the face of negative news (persuasive message), reputation and response intermingle as the story is processed.
Cognitive effects and public relations practice. The brief review above indicates that message retention plays an important role in source credibility research, both as the focus of studies and in the relationship of retention to attitude change and persistence. Of equal importance, message retention is recognized as a very important dependent measure relevant to both the theory and practice of public relations.50
Grunig has offered several reasons for the study of cognitive effects in public relations research. In an age recognizing moderate effects51 for mediated communication, cognitive effects take on importance in setting realistic public relations objectives.52 Grunig’s situational theory offers a second rationale for attending to cognitive effects. Causal path analysis of research using the situational theory suggested that active publics may arise through information gain, when information leads to problem recognition and then involvement with an issue.53
Perhaps the most compelling and current reason for examining cognitive effects pertains to the symmetrical model of public relations. In the symmetrical model, persuasion and attitude change are supplanted by an emphasis on negotiation. Imparting of information as part of an exchange leading to accommodation of an organization to its publics makes message retention and other cognitive effects highly pertinent.54
For several specialties in public relations, information gain is usually the paramount objective. Public information officers often seek to impart information without an explicit intent to change attitudes. And some proponents of integrated or mega-marketing take a fairly restricted view of public relations as conditioning the market. This is done by providing information about a product, company, issue or economic philosophy, with more persuasive efforts reserved for other elements of the marketing program.55
Veridicality and public relations. Slater and Rouner56 have suggested that some information is perceived as having a higher truth value, or veridicality than other information. Some conditions, such as a powerful preexisting reputation, may cause an individual to discount or give less credence to a message if it is deemed less veridical. It has been predicted that high source credibility is more veridical and will result in higher truth value toward a message. However, it has not been tested whether this is true with negative information about a company with a strong/positive reputation. When negative information is received about a company with a good reputation, perhaps individuals subject the message to a veridicality test and denigrate the information as less credible. The durability of attitudes formed under these conditions is an important factor in understanding long-term implications for reputation management and response strategies.
In today’s public relations world, many of the messages disseminated by PR practitioners do not pass through a news media filter. Web pages, direct mail, advertorial inserts, and corporate magazines are just a few examples of controlled communication on the rise in the field. If we can test and fine tune the effectiveness of such messages, then public relations professionals have the opportunity to determine whether the conventional wisdom about the importance of reputation is merited. If so, we can then use that information to better shape and counsel management to achieve a better corporate reputation. One of the strengths of experimental design is that it enables us to sort out theoretical issues and isolate the fundamental effects of reputation and response type. The following hypotheses attempt to address both theory and practice.
To assess the effect of a good reputation on memory, attitude, and behavioral intentions over time, the following three hypotheses were developed.
To assess the effect of an apologetic versus defensive response type on memory, attitude, and behavioral intentions over time, the following three hypotheses were developed.
The interplay of reputation and most recent response to negative information was addressed in Hypothesis 7.
Because durability of effect has practical implications and also serves to explore the concept of a sleeper effect, the following two hypotheses were developed.
Seventy three subjects in two administrations (student and non-student adults) participated in both the initial and follow up sessions of the study. In the first group, thirty-six undergraduate students (20 women and 16 men) from a large, public southeastern university volunteered to participate in this study. Each student was offered two points extra credit on a test in a large introductory level mass communication class. This group constitutes the first group to participate in the study. Volunteers were treated in accordance with the Institutional Review Board Code of Conduct.
In the second group, thirty-seven non-student adults (24 women and 13 men) were recruited from the community (a mid-sized southeastern city) using a random digit dialing protocol. Each participant was paid $40 for complete participation in the initial and follow up phases of the study.
A 2 x 2 within-subjects design was employed, meaning that each subject received each of the conditions for each of the two variables. The design controls for individual differences in this way, thereby greatly increasing the sensitivity of the measurements. The first factor was reputation (good vs. bad). The second factor was response style (apologetic vs. defensive). A complete counterbalance was used to randomly distribute variables such as company names, bad news stories, good or bad reputation and apologetic or defensive rsponses to the negative news.
Four fictitious news stories were written by a journalism graduate student with extensive newswriting experience. Then, for each story, two versions were crafted. One, where a principal of the company issued a defensive response to the negative information; and one where a principal of the company issued an apologetic response to the negative information. In all, there were then eight stories. Each story was about a different clothing company responding to negative publicity (e.g. a toxic spill into a river, sexual harassment claims, etc.). Stories were written in news wire style. Each story was independently reviewed to ensure for clarity and consistency of style. (See Appendix for copies of the stories)
Four fictitious reputational paragraphs were also written (one to accompany each news story). Two of the paragraphs were about a company with a good reputation; two were about a company with a poor reputation. (See Appendix for copies of the reputational paragraphs)
Two questionnaires were drafted. The first was a simple one-page questionnaire, asking subjects for unaided recall of the central point and all details about the story that was just read. The questionnaire also asked subjects to rate the various news writing styles of the story they had just completed.
The second questionnaire contained a battery of questions about the (4) stories subjects were asked to read in the form of matching, true/false, and semantic differential scale questions. This questionnaire was five pages long.
Packets were assembled for each participant, containing an Institutional Review Board approved consent form, four reputational paragraphs, wire stories, and one-page reaction questionnaires. At the end of each packet was the five-page, detailed questionnaire. Each subject received a random selection of (1) story topics; (2) company names (3) corporate response to the negative information in the story, and (4) reputational type.
As mentioned, the student participants all volunteered to participate through a sign-up sheet in an introductory mass communication class. The non-student participants were selected using random interval selection from the local telephone book. The University’s Survey Research Center was hired to screen and recruit the participants for the study, after approval from the University’s Institutional Review Board. This group was the second group to participate in the study. The “screening” process involved asking individuals “Have you read a newspaper in the past week?” This ensured that participants were both literate and media consumers. More specific questions were avoided because the researchers felt that a more, rather than less diverse sample would provide better evidence of robust and compelling findings. The internal validity comes from the within-subjects design, while the external validity comes from the diversity of the sample.
Testing for each group occurred in two sessions, one week apart. In the first session, subjects were asked to read the packet assigned to them and answer the questionnaires described in the stimulus materials section.
In the second testing session, subjects were asked to complete the same questionnaires given in the first session, without the news stories or reputational paragraphs to reread. This was done one week after the first session. Because of on-campus availability, the student group was asked to return to the research room where the first session was conducted to fill out the second questionnaire. The non-student group, however, conducted this second session over the telephone, answering the questions orally. Both groups received a debriefing after completion of the second questionnaire.
The questionnaires were written with three main criteria in mind: immediate and delayed memory performance, attitude formation and change, and behavioral intentions in the short and long term. These were measured in the following ways. Memory performance includes total as well as the subset of accurate details recalled about the company and the news story, unaided recall of the central point of the news story, and matching questions where subjects were asked to match the company name with the story topic.
Attitude for the companies was measured using semantic differential scales which asked the subjects to assess how prosocial and typical each company was. True/false style questions also asked whether the subject liked each company’s ethical standards and management styles.
Subjects’ intention to behave was measured using semantic differential scales, which asked the likelihood of investing in each company and the likelihood that the company’s actions would affect purchasing behavior.
To examine possible differences between student and non-student subjects, independent t-test statistics were run for the measures to assess the manipulation of reputation and response, as well as measures of memory, attitude and behavioral intention of the subjects. Significant differences were found in only three instances, in spite of the likelihood of more instances due to listwise error when 30 t-tests are done.
Two of the significant differences as a function of student status occurred with manipulations checks. The tests determined whether subjects recognized the type of response (apologetic or defensive) and the level of reputation (good or bad). Students performed better in the matching task for memory of the news story in immediate test of subjects (students 3.72, non-students 3.24, t=4.52, df 251.76, p=.000). Non students performed better in the true/false, recognition of the apologetic response type in the news stories in the immediate test of subjects. Performance was significantly different for stories employing the apologetic response strategy (students .71, non-students .85; t=-3.01, df 268.58 p=.003.)
The other significant difference pertained to the hypotheses in the experiment. Overall, the non-student adults expressed greater likelihood to purchase across all conditions of reputation and response in immediate test of subjects.
With overall possibility of significant difference in each of the 30 t-tests run to compare students and non-students, the fact that only three were found suggests the comparability of the student and non-student groups. Additionally, each group performed better on one of the many manipulations, with neither group showing consistently better performance in the measures tested. In each of the three cases where a significant difference was noted for the immediate test of subjects, there was no significant difference in the delayed test for these measures. In light of the very limited and balanced differences in performance between the two groups, all subsequent analysis used the combined data set.
To ascertain whether the experimental manipulations were effective, the following tests were run. Responses were manipulated using either apologetic or defensive versions of the news stories. Subjects were asked to rate, using a semantic differential scale, whether each company’s response to the news story was apologetic or defensive (apologetic=1 defensive=7). Using an independent samples t-test, a significant difference was found between the apologetic and defensive responses in both the immediate and delayed tests (Immediate�apologetic 4.48, defensive 3.24, t=5.35, df 286.84, p=.000) (Delayed�apologetic 4.27, defensive 3.48, t=3.54, df 279.33, p=.000).
Reputation was manipulated using either good or bad versions of background information about companies accompanying each news story. Subjects were asked to rate, using a semantic differential scale, whether each company’s reputation prior to the news story was excellent or poor (excellent = 1 poor = 7). Using an independent samples t-test, a significant difference was found between the good and bad reputations in both the immediate and delayed tests (Immediate�good reputation 5.54, bad reputation 2.97, t=12.68, df 286.64, p=.000) (Delayed� good reputation 4.95, bad reputation 3.46, t=7.40, df 283.57, p=.000).
To assure that the particulars of the stimulus materials did not contain reactive elements, a Oneway ANOVA was used for company name and story type. No significant differences were found among any of the dependent measures as a function of company name (the four names of clothing companies used were: International Female, Cricket, Bodywear, and Rustler). For story type, only four of a possible 30 comparisons were significant. Furthermore, no pattern of reaction to a particular story was found; i.e. the significant differences as a function of story type involved different stories in the four instances.
The MANOVA was used for within-subjects design analysis because the subjects received each condition of the independent variables – therefore, this desgin serves to control for individual differences with MANOVA as the best, most powerful statistical procedure to analyze data in this within-subjects design.
Throughout the experiment, MANOVA enabled isolation of the relative effect of reputation on memory performance for subjects by controlling through the experimental design for individual differences of the subjects1 (see Table 1).
Table 1. A summary of each hypothesis and corresponding results.
|H 1: Better memory for companies with bad reputations||Total details�N.S.|
|H2: Better attitude toward companies with good reputations||Prosocial�p=.000|
|H3: Better purchasing and investing behaviors for companies with good reputations||Invest�p=.000 |
|H4: Better memory for companies with a defensive response||Total details�N.S.|
|H5: Better attitude toward companies with an apologetic response||Prosocial�N.S|
|H6: Better purchasing and investing behaviors for companies with an apologetic response||Invest�N.S. |
|H7: Bad reputation and defensive response will produce the most drastic overall effect||Invest�.017|
|H8: Memory will decay after one week||Total details�p=.008 |
|H9: Attitude and behavioral intention will regress toward the mean after one week||Prosocial�p=.017 |
Hypothesis 1 stated that memory for information about companies with bad reputations will be better than memory for companies with good reputations. Because each subject was exposed to both conditions for reputation, the effect of reputation on memory performance was evaluated using repeated measure, MANOVA.
As a function of reputation, memory performance measured by recall of total details from stories, accurate recall of details from stories, identification of the central point of the story, and matching of story topic with company name were all non-significant.
Hypothesis 2 stated that attitudes toward companies with a good reputation will be better than attitudes toward companies with a bad reputation. As a function of reputation, attitude was measured by the semantic differential scales asking subjects if the companies were prosocial or; and if the companies were typical or atypical. A significant difference was found in the measurement of the company’s prosocial stance F (1,59) = 1.33, p=.000. Subjects indicated that companies with a good reputation were more prosocial than were companies with a bad reputation. No significant difference was found as a function of reputation for the measurement of the company’s typical behavior.
True/false questions were used to measure attitude asking whether participants: liked the company’s management styles and liked the company’s ethical standards. MANOVA was used to analyze the true-false questions; in this case, dichotomous data can be treated as interval data, with analyses arrayed from 0 to 1. Significant differences were found in response to both questions as a function of reputation (management styles�F (1, 59) =.35, p=.001; and ethical standards�F (1,58) =.33, p=.000. Subjects were more inclined to like the management styles and ethical standards of companies with good reputations than they were for companies with bad reputations.
Hypothesis 3 stated that intended purchasing and investing behaviors will be more favorable for companies with good reputations than for companies with bad reputations. As a function of reputation, behavioral intent was measured by subjects’ indication on a semantic differential scale of: the likelihood that purchasing behaviors would be affected and the likelihood of investing with the company. Subjects indicated they were significantly more likely to invest in an organization with a good reputation than with a bad reputation F (1,61) =5.55, p=.000. As a function of reputation, likelihood that purchasing behavior would be affected was non-significant.
Hypothesis 4 stated that memory for companies using a defensive response in a news story will be better than memory for companies using an apologetic response. Because each subject was exposed to both response types, the effect of response on memory performance was evaluated using repeated measure, MANOVA. This enabled isolation of the relative effect of response on memory performance for subjects, controlling for individual differences of the subjects.
As a function of reputation, memory performance measured by recall of total details from the stories, accurate recall of details from the stories, identification of the central point of the stories, and matching of the story topic with the company name were all non-significant.
Hypothesis 5 stated that attitudes toward companies using an apologetic response in a news story will be better than attitudes toward companies using a defensive response. As a function of response, attitude measures of the companies’ management styles, prosocial standing, and typical behavior were all non-significant.
Hypothesis 6 stated that intended purchasing and investing behaviors will be more favorable toward companies using an apologetic response in a news story than for companies using a defensive response. As a function of response, behavioral intent was measured by likelihood that purchasing behavior would be affected and likelihood of investment with each company. Differences between response types were non-significant for these measures.
Hypothesis 7 stated that reputation and response type interact such that the combination of bad reputation and defensive response is markedly different from any other combinations of the two variables. The only significant difference found for reputation as a function of response was for the subjects’ likelihood to invest. Subjects said they were least likely to invest in companies with an already poor reputation that responded apologetically. However, subjects indicated they were most inclined to invest in a company with a good reputation that responded apologetically F(1,61) =4.64, p=.017.
Hypothesis 8 stated that memory for information about companies will decay after approximately one week. This hypothesis was supported by memory performance measured by recall of total details from stories, accurate recall of details from stories, identification of the central point of the stories, and matching of the story topic with the company name. One week after subjects completed the first questionnaire, they recalled significantly fewer total details about the stories than immediately after reading the stories F (1,34) =6.71, p=.008. Subjects also recalled significantly fewer accurate details about the stories after a one-week delay F (1,43) =7.60, p=.000. Subjects’ ability to recall the central point of the story was significantly less after the one-week delay F (1,29) =.23, p=.000. After the delay, participants were also significantly less likely to correctly match the company name with the story topic they had been given F (1,62) =.26, p=.000.
Hypothesis 9 stated that attitude and intention to behave will regress toward the mean after approximately one week. As a function of delay, this was measured by subjects’ indication of: whether the companies were prosocial or antisocial, if the companies were typical or atypical, the likelihood the company’s behavior would affect their purchasing behaviors, and the likelihood of investing with this company. Subjects’ ratings of the companies as prosocial or antisocial did significantly regress toward the mean after the one week delay F (1, 59) =1.33, p=.017. No other significant differences were found.
Hypotheses 1 and 4 stated that memory for information about companies would be affected by the company’s reputation and response type. These hypotheses were not supported. Subjects abilities to recall both raw and accurate details about the stories, to identify the central points of the stories, and to match the stories with the topics were not affected by the response type or the reputation of the organizations.
Reputation did, however, affect subjects’ attitudes toward the companies. Not surprisingly, companies with a strong reputation were seen as more prosocial. Subjects also preferred the ethical standards and management styles of good reputation companies. These results support hypothesis 2. It was hypothesized (H5) that the apologetic style of response in the negative news story would also have significant positive effects on attitude. However, subjects did not view companies that gave an apologetic response type as more prosocial. Subjects also did not prefer the management styles or ethical standards of companies that issued an apologetic response.
This is true even though it was shown subjects could recall both the reputation and response type for each organization. Hypothesis 5, stating that attitudes toward a company using an apologetic response would be better than a defensive response, was not supported. This indicates the power of reputation in comparison to response type. Subjects generally favored companies with a good reputation� regardless of whether the organization apologized or reacted defensively to the negative information.
It should be noted that one measure of attitude, whether the company was considered typical or atypical, produced no significant results. This may be because the measure lacked valence. That is, subjects who held favorable attitudes about an organization may mark that company as “atypical.” Conversely, subjects who held unfavorable attitudes about a company could also consider the company “atypical.”
Hypotheses 3 and 6 stated that reputation and response type would affect subjects’ behavioral intent. Lack of valence in the measurement may also have been a factor for behavioral intent questions. As a measure of likelihood that purchasing behavior would be affected, subjects may have interpreted this question idiosyncratically due to a lack of valence. The question asked, “How likely is it that this company’s actions would affect your purchasing behavior?” It appeared that some subjects who felt favorably toward the organization responded “very likely” to this measure�whereas others who felt very unfavorably also responded “very likely” to this measure. The reason there were no significant results produced from this question, may be that the effects were, in fact, “washed out.”
Subjects did indicate, however, that they were more likely to invest in an organization with a good reputation than with a bad reputation. In keeping with the results of the attitudinal measures, response type had no effect on subjects’ likelihood to invest. That subjects were more likely to invest in good reputation companies coincides with the fact that subjects also had more favorable attitudes about these companies�again, this is true regardless of whether the organization issued an apologetic or defensive response. Hypothesis 3, then was partially supported, while hypothesis 6 was not supported.
Hypothesis 7 stated that reputation and response type interact such that the combination of bad reputation and defensive response would be markedly different from any other combinations of the two variables. As we have seen up until this point, prior reputation of a company can have a noticeable effect on attitudes and intended investing behaviors, while response type does not seem to matter.
However, when asked to indicate how defensive an organization was in response to the negative information, significant differences were found both as a function of response and reputation. Conversely, subjects did not link response type to the evaluation of the company’s reputation. That is, in evaluation of the reputation of the company, no significant difference was found as a function of response type. This indicates that, while subjects could correctly link the response type to the story, they seem to confound reputation with response. There was a main effect for reputation, showing up in a measure intended to check for the manipulation of response. Reputation appears to be a powerful force in subsequent judgments about the company�subjects may make unfounded attributions about other aspects of the organization based on reputation. This includes attributing a response type as a function of reputation.
Another powerful implication found from results to hypothesis 7 is that subjects’ likelihood to invest hinges on reputation as a function of response. Subjects were significantly less likely to invest in companies with an already poor reputation that responded apologetically. However, they were most inclined to invest in a company with a good reputation that responded apologetically. The apologetic response to negative information has drastically different results, depending on one’s preexisting reputation. For companies with a bad reputation, it appears that an apologetic response only makes matters worse. However�consistent with conventional wisdom�an apologetic response to negative information appears to reinstate the organization in good graces if their reputation was strong to begin with. This “damned if you do�saved if you do” paradox reemphasizes the importance of a precise understanding of one’s current reputation (see Figure 1).
Figure 1. Subjects’ likelihood to invest depending on reputation as a function of response.
Hypothesis 8, which stated that memory for information about the companies would decay after one week, was supported. As expected, subjects were less able to recall the specifics of the stories and were less able to correctly identify the company name with the story type. Although this finding was not a surprise, it is an important finding when measuring the durability of the effects. In this study, it was important to measure the attitudes and behavioral intentions after subjects were less able to recall the specifics of the stories. In the “real world,” we aren’t normally asked to make evaluative decisions about organizations immediately after reading the newspaper or watching the news�instead these questions may arise a day, week, or months after one is exposed to the negative publicity.
Hypothesis 9 stated that attitude and behavioral intention would regress toward the mean after approximately one week. This hypothesis was partially supported. Over time, subjects’ assessments of how prosocial a company was did significantly regress toward the mean. Also, subjects ratings of the companies’ reputations significantly regressed toward the mean as a function of delay by reputation. After a week, significant differences still existed, showing that subjects could accurately rate the good and bad reputation companies. However, good reputation companies were not rated quite as good (means regressed from 5.77 to 4.93) and bad reputation companies were not seen as quite as bad (means regressed from 2.98 to 3.5). This is likely another indication that, as time passes, the strength and accuracy of our judgments become weaker and more nebulous.
In addition to testing the aforementioned hypotheses, this study sought to develop a useful research tool that would enable public relations practitioners to determine the “reputation” of a company. It also sought to develop a helpful experimental protocol for message testing in public relations. As a result of this study, several strengths and weaknesses for using such a methodology have emerged. Using memory, attitudinal and behavior intent as the dependent measures for this experiment have proven useful in determining a comprehensive understanding of the subjects’ short and long term reactions to the companies. However, as in most social science research, hindsight reveals several areas for improvement should practitioners attempt to employ a “reputation audit” using this experimental design.
As mentioned, valence was not included for two questions in the experiment�one measuring attitude, the other measuring behavioral intent. This should be an important consideration in the structure of each item on future questionnaires, so as not to confound responses.
Also, while the experiment attempts to examine the effects of reputation, the reputational stimulus materials in this study consisted of only one exposure. Some might argue that reputation is something developed over time. Due to the experimental nautre of the study, subjects’ exposure to the corporate “reputation” could only be offered once. However, the reputational materials were crafted to create a “history” of corporate conduct. As shown in the sample reputational paraghaphs attached as appendices, each “reputation” is built on a series of events, occuring over an extended time period.
Finally, results of this study indicate that the appropriate response type to negative information may hinge on the company’s preexisting reputation. This stresses the need for a “reputation audit” measurement tool. Such a service could be an important part of the research program offered to major clients by agencies. However, it also indicates that a few more questions may be necessary to understand why subjects felt less favorable toward bad reputation companies with an apologetic response type. Perhaps subjects find an organization with a poor reputation less credible or sincere when it issues an apology. Likewise, a company that has proven to be a shining star of social responsibility is afforded the benefit of the doubt in a time of crisis�and a public apology might reinforce the view as a good corporate citizen. Further study of why subjects appreciate the apologetic response type from companies with a good reputation but frown upon apologetic companies with a bad reputation is in order. Measures of the companies’ sincerity and credibility should be added to the questionnaire�both for immediate and long-term effects�to sort out a possible “hypocrisy factor” at work.
As controlled messages become more common in the field of public relations, the experimental methodology used in this study will become increasingly important in testing and understanding the effects of messages distributed under the control of the public relations practitioner. The prototype used in this study allows public relations practitioners to adopt copy testing methods with a focus on targeted messages in public relations. The method used in this study shows promise as a tool to evaluate reputation and message testing. Future use of this method should include valence in all individual measurements. Additionally, measurement of message credibility and company sincerity should be included in evaluations of the response type to better determine why the best response type may depend on the existing reputation. The dependent measures of memory, attitude, and behavioral intent provide a solid and comprehensive understanding of short and long term effects of the corporate reputations and responses.
Results of this study show the powerful effect reputation has on both attitude and behavioral intent. Even though subjects were able to accurately recall both the reputation and response type in the short and long term, subjects were more inclined to use reputation as the predictor of attitude and behavioral intent. This may provide some research-based insight to the conventional wisdom addressed in this study: Why might organizations continue to react defensively in response to negative information when it is traditionally discouraged as a response type? It appears that if an organization is operating with a good reputation, response type to negative information may not have the powerful damaging effect traditionally thought. In fact, this study indicates that there are some situations when one’s reputation is such that a defensive response type might even be more effective in the long-term.
Of course this is a cyclic dilemma, and it is possible that continued lying, obfuscating, and stonewalling could lead to an even worse reputation over time. Notably, this study shows that we hold less favorable attitudes and are less likely to invest in organizations with a bad reputation, linking reputation to bottom line considerations for all companies.
If so much of our attitude and behavior toward an organization depends on reputation, then an accurate reputational measurement is essential to conduct a reputational audit. It is likely that organizations all too often operate on the assumption of a strong reputation, without any reliable proof. This might metaphorically resemble a slow boiling crisis, where retroactively executing the wrong response type could be too little too late. A reputational audit measurement would provide companies with a practical and useful tool, that could be used to plan the most effective response type for that individual organization before a crisis occurs.
This study also shows that, using naturalistic stimulus materials, individuals are able to accurately discern: a good from a bad reputation, an apologetic from a defensive response. This manipulation check reinforces the existing belief that our efforts to shape reputation, and craft individual response types do matter in the short and long term.
In order to most effectively counsel clients, public relations practitioners must have a solid grasp of reputation evaluation and management. In the field of public relations, an accurate measurement tool such as this is essential in order to provide counsel that won’t backfire.
Sample of a good reputation paragraph
International Female has recently been hailed by Forbes magazine for its progressive “Vision 2000.” This organizational plan for the future involves every employee in the company and comes as a result of a year-long progress of meetings and surveys involving all 2,000 employees.
Over the past five years, International Female has implemented extensive internal communication audits, focusing on how it could improve new corporate health, safety and environmental policies. It has used this formal communication to make decisions impacting the company’s relationships with communities, customers and employees.
One result, stemming from a growing concern for environmental safety was to restore a wetland, building a sanctuary for local wildlife�including crocodiles, fish, turtles, birds and various flora�with a lake at its center.
This company is considered a leading force for corporate good�committed to social responsibility and ethical trade as an integral part of business.
Sample of an apologetic response to negative publicity
CHICAGO (AP)�Officials of a women’s underwear company admitted Monday that a toxic chemical spilled from their factory into the Chicago River.
Landon J. “Lanny” Isham, chairman and CEO of Cricket Women’s Wear, 155 Canal St., said the chemical, apparently leaked into the river for 30 hours. Isham said officials are still trying to find out how many gallons of the chemical flowed into the river, but that their estimate as of late Monday morning was about 2,000 gallons.
Todd Foster of the Illinois state Environmental Protection Agency said that the chemical is toxic to humans and animals, who typically become ill from inhaling its fumes and may die from ingesting even small quantities through water, other liquids, or food. Foster speculated that the spill will eventually kill hundreds or even thousands of fish. He added that the spill will be carefully monitored for whether the chemical drifts toward Chicago water treatment plants or beaches.
Foster said that water treatment plants typically can remove such chemicals from drinking water and therefore that the spill posed little hazard for Chicago residents unless it hits beaches.
He noted that Cricket was using the chemical legally and that it is fairly commonly used in clothing manufacturing to set various shades of purple dyes so that they don’t fade in sunlight or wash out during cleaning.
“I think I speak for the entire family of Cricket managers, employees and even our chemicals supplier, Gas Products Inc., when I say that we are horrified by this event,” Isham said. He said that he is active himself in various environmental causes and organizations, and that the chemical spill is “like a nightmare to me.”
He explained that the company had no operational need to be located next to a river, and that the factory located to Canal Street only because executives wanted to help revitalize the area by rehabilitating the abandoned warehouse that became its factory.
Isham said that company policy and procedures had been specifically designed to prevent spills both inside and outside the factory, especially because of its location next to the river. In a prepared statement, Isham said his company has been “socially responsible in both employment and environmental matters,” and he hoped the public will “see this incident as the tragic, isolated instance that it is.”
Although the exact cause for the spill is not certain, it appears to have resulted from a combination of accidents. Troy Montoya, the factory’s operations director, said that a pipe from a vat containing the chemical burst nearby a broken window, and that the chemical spurted out the window onto the sidewalk below. The sidewalk had been designed to drain rainwater into the river, a common practice for old sidewalks near the river, Montoya explained.
Montoya explained that the leak was in a little used area of the factory and that employees discovered it Monday morning about 10 a.m. when they went outside to smoke.
Cricket, based in Grand Tetons, Wyo., was founded in 1988 by Isham’s late wife, Lori, and is still closely-held by the family. In 1989, it was briefly in the news when employees of Kricket, a Hanover, Pa.-based men’s clothing company, urged their executives to file suit against Cricket for trademark infringement. However, Kricket executives said they had no plans to ever manufacture women’s clothing, and the Cricket name was probably helpful to Kricket’s marketing.
Sample of a bad reputation paragraph
Bodywear, known as a leading clothing manufacturer, has recently come under fire once again. The company, already known for a poor environmental record and repeated union disputes, is once again facing public scrutiny. The Federal Trade Commission has issued a cease and desist order for a series of false and misleading advertisements.
The television advertisements claim that Bodywear has been and continues to be a leading philanthropic force in AIDS research, when in fact, there are no financial records proving this true.
Upon public release of this news, 250 AIDS activists hanged an effigy of Bodywear President and CEO, while demonstrators marched on the company’s headquarters, according to a recent article in Business Ethics.
Sample of a defensive response to negative publicity
CHARLESTON, W.V. (AP)�Although sooty, black smoke belched from the smokestack of Rustler Jean Co.’s factory for hours Wednesday night and was visible for more than two miles, company executives insist all factory operations have remained normal.
Area residents, including various city and county officials, began calling state and federal officials, as well as Rustler Jean Co. offices, Thursday morning. They said the smoke seemed significantly more and thicker than usual; estimates put the smoke’s quantity and density at five to 20 times normal levels. One executive of a nearby company, speaking on condition of anonymity, said that the smoke was typical of that from smokestacks with malfunctioning or non-functioning smoke scrubbers.
But Rustler’s President Glorianne Osborne said that smoke emitting from the smokestacks, located in east central downtown near the Byrd Freeway contains only steam, carbon dioxide, low levels of carbon monoxide, and “extremely low” levels of other gases. She said that she has no indication that emissions changed significantly this week.
Osborne admitted that the entire factory was shut down by late Thursday morning, but “only until we resolve these baseless complaints.” She added that her company wishes to “apologize to our employees, our employees’ families, and our retailers” for being forced to shut down the factory without just cause. About 300 people work at the site.
State air pollution control officials are investigating the situation and had no details as yet. Jimmy Mathias, director of the state Environmental Protection Agency, said that shutting the plant down relatively quickly limited the amount of risk to citizens in Charleston and to its east, if there was any risk.
To be safe, Mathias said, persons with asthma, bronchitis, emphysema, and other lung and respiratory ailments should stay indoors. He also urged Charleston residents to take immediate steps to decrease the number of other pollutants in the air for one or two days. Downtown workers should try to carpool or take public buses, Mathias explained, and residents in the eastern part of the metropolitan area should limit burning in woodstoves, backyards and farm fields.
Osborne explained her factory’s smokestack scrubbers have been in use since the late 1970s, and had undergone extensive modification since then to comply with increasingly strict local, state and federal air quality standards. She rebuffed reporters’ questions about scrubber inspections and whether they should already have been replaced with new ones.
Matthew McGovern, spokesman for Our Lady of Mercy Hospital, said that his hospital has received no admissions of Rustler employees or others since the accident, but said that he wouldn’t be surprised if they were. “Persons with respiratory ailments exacerbated by inhaling unscrubbed smoke may not seek emergency care unless breathing problems or pain persist,” he explained.
Charleston Mayor Dana Fields said that he had heard many complaints and comments about the Rustler plant but was unyet aware of any significant injuries or other damage. “We have confidence that Rustler not only works to prevent environmental hazards, but that in the event of problems, it will quickly fix them so that its employees can return to work. We also believe that Rustler is the kind of company that will take steps to make sure that no specific accident ever happens a second time,” he said.
“Rustler is an important Charleston employer and has always acted responsibly before,” Fields said. “If something has gone wrong, we will consider this to be an isolated incident for which Rustler is taking appropriate steps.” He confirmed that he had been in touch with Osborne.
The mayor added that advisors told him that state and federal laws are clear about required performance levels of air pollution scrubbers, but are vague about the circumstances under which scrubbers must be replaced.
About the Authors:
Lisa Lyon is a doctoral student at the University of Georgia. Glen T. Cameron, Ph.D., is a professor at the University of Missouri.
Manuscript produced under the auspices of The C. Richard Yarbrough Public Relations Laboratory Henry W. Grady College of Journalism and Mass Communication The University of Georgia, Athens, Georgia 30602-3018 (706) 542-4791.
Funding for this study was provided by The Institute of Public Relations Research & Education. Correspondence concerning this article should be addressed to Lisa Lyon at the above address, office phone: (706) 542-4475, home phone: (706) 613-6915, Internet: Llyon@arches.uga.edu